review article benchtop technologies for circulating tumor

Review ArticleBenchtop Technologies for Circulating Tumor Cells SeparationBased on Biophysical Properties

Wan Shi Low and Wan Abu Bakar Wan Abas

Department of Biomedical Engineering University of Malaya 50603 Kuala Lumpur Malaysia

Correspondence should be addressed to Wan Shi Low low89shigmailcom

Received 26 December 2014 Revised 26 February 2015 Accepted 26 February 2015

Academic Editor Jianyu Y Rao

Copyright copy 2015 W S Low and W A B Wan Abas This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Circulating tumor cells (CTCs) are tumor cells that have detached from primary tumor site and are transported via the circulationsystem The importance of CTCs as prognostic biomarker is leveraged when multiple studies found that patient with cutoff of 5CTCs per 75mL blood has poor survival rate Despite its clinical relevance the isolation and characterization of CTCs can bequite challenging due to their large morphological variability and the rare presence of CTCs within the blood Numerous methodshave been employed and discussed in the literature for CTCs separation In this paper we will focus on label free CTCs isolationmethods in which the biophysical and biomechanical properties of cells (eg size deformability and electricity) are exploited forCTCs detection To assess the present state of various isolation methods key performance metrics such as capture efficiency cellviability and throughput will be reported Finally we discuss the challenges and future perspectives of CTC isolation technologies

1 Introduction

Cancer is one of the leading causes of death worldwideAccording to the International Agency for Research on Can-cer (IARC) there are an estimated 82 cancer-related deathsin 2012 where 90 of them are caused by metastasis [1] Asa result metastasis has become the prime prognosis factor incarcinoma patients Generally cancer metastasis involves thespread of cancer cells whereby the tumor cells detach fromprimary tumor site and be transported via the circulationsystem to a distant organ to form secondary tumors Thesecells which shed into vasculature are referred to as circulat-ing tumor cells (CTCs) The presence of CTCs was first dis-covered by Thomas Ashworth in 1869 after comparing theirmorphology to tumor cells from different lesions Despite hisdiscovery its impact on cancer detectionmethodwas lesswellestablished in the early stage due to the lack of detailed insightinto the mechanisms of tumor

In clinical practice the cancer diagnostics are commonlyperformed through radiological imaging modalities such astraditional radiography (X-ray) magnetic resonance imag-ing (MRI) computed tomography (CT) positron emissiontomography (PET) or ultrasound These techniques allow

visualization of internal body structure Thus it enables phy-sicians to delineate the group of tumor cell colonizationHowever there are some pitfalls in these techniques Forinstance the deficiencies of resolution in imaging modalitieshave precluded them to image small numbers of cancer cellsbefore angiogenic switch which in turn limit the detectionsensitivity [2 3] Furthermore most of the cases are normallydiagnosed at advanced stages where patients often relapsedwithin 24 months of therapeutic intervention [4 5]

In recent years the emerging data have challenged thetraditional theory of metastasis sequential development Infact study carried out by Husemann et al highlights thatCTCs can be found in patients even before a primary tumoris detected with conventional clinical screening methods [4]The importance of CTCs is further augmented when thereare increasing evidences about the presence of significancecorrelation between the number of circulating tumor cellsand patients survival times It has been scientifically validatedby prospective multicenter studies that patient with cutoff of5 ormore CTCs per 75mL of blood would have poor survivalrate [4 6 7] A similar analysis of prognostic value of CTCsamong colorectal cancer patient was performed by Allen andEl-Deiry Their study points out that the median progression

Hindawi Publishing CorporationBioMed Research InternationalVolume 2015 Article ID 239362 22 pageshttpdxdoiorg1011552015239362

2 BioMed Research International

free survival (PFS) and overall survival rates were twice ashigh for patient with less than 3 CTCs per 75mL of bloodthus it has confirmed the previous findings Additionallythis group also presented significant data which showed thatpatients with elevated CTCs density after therapy wouldhave poor survival rate [8] Nevertheless simple enumerationof CTCs is inadequate because cancer is a constellation ofdiseases with various pathologic alterations that might causeprognosis Since the ability in analyzing proliferation of viableCTCs has still been lacking it is difficult to assess CTC infor-mation which is the representative of cellular informationavailable in primary tumor In fact the dimension of CTCbiological feature is especially significant for basic researchpertaining to metastasis as well as drug development Tofurther complicate matters the recent appreciation of geneticalterations and biomarker expression for instance KRASwithin tumors means a single biopsy sample is no longersufficient [9] Henceforth detecting and analyzing these cellson a sample of blood may shed new light on circumventsclinical need to improve therapeutic efficacy as well as theoverall patient survival rate Despite its high potential incancer treatments the detection of CTCs from whole bloodsample is particularly challenging due to their extremely rarepresence which is only 1 to 10 CTCs per billion normalblood cells in patients with advanced cancer Besides thelarge morphological variability among CTCs has imposedtechnical challenge to isolate the whole population wheresophisticated algorithms are required to identify the CTCsAforesaid CTC counts are associated with patient prognosisThus a highly sensitive detectionmethod is vital to accuratelycharacterize and enumerate the CTC

Numerous methods have been employed and discussedin the literature for CTCs separation To date the benchtopdevice developed by Veridex is the sole system approvedby United States Food and Drug Administration for theclinical monitoring of CTC counts in patients withmetastaticbreast colon and prostate cancerThrough this approach theCTCs are distinguished from other blood cells based on theirimmunoaffinity properties The selection is carried out byusing antibodies coupled with magnetic beads targeting thetumor-associated antigen onCTCs surface (EpCAMCK andCD45)This system is proved to have sensitivity of 877withthe capability to detect of approximately 5 CTCs per 75mL ofwhole blood Despite the proven clinical ability of this systema few data have been published concerning the inconsistentlymarkers expressed by all tumor cells Noteworthy EpCAM isnot expressed in all histological tumor types and thus mightresult in false-negativepositive result and limit the purity ofthe enriched CTC fraction Using affinity capture methodin macroscale analytical system will also lead to permanentattachment of target cells to marker protein on CTC Con-sequently the downstream options for the extraction andsubsequent characterization of CTC will be narrowed down

In addition to immune-affinity based separation in fact alarge panel of approaches for CTC isolation that are indepen-dent of cell surface antigens have been developed on the basisof its physical properties The majority of the physical meth-ods for CTCs isolation exploit the differences in mechanicaland physical properties including cell size deformability

electrical polarizability and magnetic susceptibility In con-trast to the immune-affinitymethodwhere epithelial antigensare needed to mediate the intercellular adhesion this tech-nique is label-free Therefore the interference such as samplecontamination due to the tagging molecules can be avoidedSince CTCs are unmodified by physical separation cells iso-lated using these methods are compatible with wider range ofanalyses including those requiring viable cells Consequentlyit allows large range of molecular biological analysis of CTCsto be implemented and thus provides clinician with furtherinsight into CTCs role in tumor formation

In this paper we will present an overview about the pot-ential application of benchtopCTCdetection device based ontheir physical properties in clinical oncology Both detaileddescription of various separation principle and comparisonsbased on separation metrics such as efficiency viability andthroughput are included in this review Additionally thechallenges faced by each technique will too be succinctly dis-cussedThe readers are encouraged to read the original papersfor additional details

2 Comparison of Biophysical andBiomechanical Properties betweenCTCs and Normal Blood Cells

According to most CTCs histological study researchers haveobserved changes in cell cytoskeleton contents and theircytoskeletal structures as it progresses toward a cancerousstate For instance CTCs are found to have greater nuclear tocytoplasmic ratio larger size and distinct nuclear morphol-ogy in contrast to the normal cells In fact these cytoskeletalchanges have resulted in changes in the overall mechanicalproperties of cells Computational mathematical model con-ducted by Rejniak has demonstrated the insight of CTCrsquosdeformation trajectories within blood vessel [36] His find-ing highlighted that CTCsrsquo cellular biomechanics such asdeformability played an important role in metastasis As anexample in order to invade distal sites the stiffness of CTCscytoskeleton is modified subsequently in a very dynamic wayso that it could squeeze across small spaces in extracellularmatrix and endothelial cell-cell junctions as well as circulat-ing through the small capillaries In order to successfully per-form the metastatic cascade the CTCs cell membrane mustwithstand hemodynamic forces and overcome the effects offluid shearswithin the blood vessel Consequently it alters theconservation of membrane structure which in turn affectsthe surface charge (electrical charge) in contrast to normalblood cells [37] These distinct physical differences in sizedeformability and electrical properties between tumor cellsand blood cells have allowed researchers to employ themas physical-based CTC detection method The physical andbiomechanical properties of CTCs will be briefly discussedin the following section

21 Size Inmorphological studies themeasurementmethodsuch as flow cytometry [38] blood smear or detailed exami-nation with microscopic study are broadly employed to ana-lyze the cell size of CTCs in comparison to normal cells

BioMed Research International 3

The size measurement of cell is normally interpreted in cellarea length width and shape features Despite the differ-ences in these measurement methods they provide similarquantitative value for both CTCs and normal blood cellmeasurement respectively For example normal humanerythrocyte also called red blood cells (RBC) is found tohave biconcave disks with the diameter of approximately 6to 8120583m [38ndash40] White blood cells (WBC) are derived intotwo groups which are granulocytes and agranulocytes Thegranulocytes such as neutrophils and eosinophils are typicallyof 12 to 15 120583m in diameter [41] Meanwhile agranulocytessuch as lymphocytes vary widely in size where the smalllymphocytes are 7ndash10 120583mwhile large lymphocytes are appro-ximately 14ndash20 120583m in diameter Monocytes are among thelargest leukocytes with a diameter of 15 to 25 120583m in bloodsmears [38 41ndash43] In contrast to the mentioned blood cellsthe size of CTCs reported in the literature is generally largerthan normal blood cells ranging from 17120583m to 52 120583m Theclinical relevance of this taxonomy has been investigated andconfirmed by many research groups [44ndash47] For instancefew groups of study done on breast cancer patients had foundthat CTCs showed significant variability with consistentelongated shape and most of them were larger compared tothe leukocytes surrounding it [45 48 49] Similarly in astudy of patient with prostate cancer Park et al had reportedthat highly ruffled surface membrane of CTCs had createdan excess of membrane surface area As a result majority ofCTCs have larger size in contrast to normal blood cells [50]The observation of differences in size of CTCs in comparisonto other blood cells has motivated several devices to exploitit as their primary label-free separation criteria Despite thefact that CTCs have larger cells compared to other blood cellsthere is a significant overlap in size of CTCs and leukocytesthat might hinder size-based separation process The calcu-lation conducted by Marrinucci et al had concluded thatCTCs are highly heterogeneous including both high and lownuclear to cytoplasmic ratios whichmight cause them to varyin size [51] Therefore proper consideration in the proposeddevice design and their operational parameters are needed forhigh efficiency isolation-based-size method

22 Deformability The measurement of tumor cells defor-mabilitymainly refers to viscoelasticity a property ofmaterialthat falls into a category between that of an elastic solid and afluid Cellular viscoelasticity arises from the combination ofhighwater content conflatedwith a polymerized cytoskeletonstructural matrix In microrheological studies this mechan-ical property is governed by cell surface interactions as wellas normal force exerted by the cell on the capillary wallunder physiological conditions and in response to externalsignals Various approaches were employed in vitro to studythesemechanical changes with techniques ranging fromcon-ventional atomic force microscopy (AFM) [52ndash55] micro-pipette aspiration (MPA) [56 57] and magnetic twistingcytometry to more recently tailored automated systems suchas microfluidic resonator [58] and inertial focusing method[59] The advantages and drawbacks of these techniqueshave been discussed in detail [56 60] In fact the tradeoff

between experimental automation and complexity ofmeasur-able properties method (such as AFM andMPA) has resultedin most measurement lacking throughput and precision

Despite of their shortcomings several consistent cellmechanical behaviors are observed across multiple studieswhere cancer cell lines are employed Measurements of thecellular mechanical behavior are frequently lumped into asingle universal parameter Youngrsquos modulus It quantifiesthe measurement of cancer cell stiffness which is intimatelylinked with the distribution of actin network within cellcytoskeleton Various studies have compared the viscoelasticproperties of cultured cancer cells and blood cells usingAFM techniqueThe stiffness ofmetastatic cells was generallyfound to be lesser (Youngrsquosmodulus of 37 kPandash150 kPa) com-pared to normal blood cells (Youngrsquosmodulus of 02 kPa) [53ndash55 61] This statement is concurred with a recent study con-ducted by Byun et al [58] whereby a suspended microfluidicresonator is used to track the cellrsquos velocity as the blood sam-ple spikedwith human lung cancer cells traverses through theconstriction region of the device Their result has indicatedthat lung cancer cells took a shorter time to deform and trav-eled through the integrated constriction in the microfluidicchannel compared to normal blood cells thus suggesting thatcancer cells exhibit high cytoskeletal deformability [58]

While researching the viscoelasticity properties of tumorcell in the context of metastasis various studies have exp-ressed the strong correlation between cell deformability andcell malignancy For instance Mak and Erickson [56] andMohammadalipou et al [57] have measured the mechanicalproperties of both metastatic and benign breast cancer celllines withMPA whereby the metastatic cancer cell lines werefound to have longer aspiration length which in turn resultedin higher Youngrsquos modulus measurement value (calculatedfrom the pressure slope versus aspiration length graph)Likewise the AFM-based analysis performed by Chen et alacross prostate cancer cells lines (LNCap-AD LNCap-AIand PC-3) indicated that the noncancerous BPH-1 cells werefound to have the least elasticity with Youngrsquos modulus of37 kPa whereas the highly metastatic PC-3 cells (Youngrsquosmodulus of 013 kPa)were almost 30more elastic in contrastto BPH-1 Noteworthy Youngrsquos modulus values described inthese experiments were formulated based on the cell elasticproperties using Hertz model which was used to describethe physical relationship between the applied force and thecantilever deflection on indentation Apart of cultured celllines Chen et al also havemeasured Youngrsquos moduli of CTCsisolated from blood of patient with castrate-resistant prostatecancer and bone metastasis When a comparison was madebetween CTCs and prostate cancer cell line the values ofYoungrsquos moduli obtained for CTCs (ranged from 023 to11 kPa) were similar to that of PC-3 thus inferring they aremetastatic In accordance with the relationship between cellmalignancy and deformability investigation of ex vivomech-anical changes of cancer cells was conducted by Cross et al[62] Their study has reported the stiffness of live metastaticcancer cells taken from pleural fluids of breast cancer patients(Youngrsquos modulus of 053 plusmn 010 kPa) that was 70 lower incontrast to the benign cells reactive mesothelial cell (197 plusmn007 kPa) within the same fluid samples This statement is in


agreementwith study demonstrated byChen et al on prostatecancer cell line as mentioned previously

The brief finding presented here indicates that celldeformability can be considered useful parameters to reflecta histological background of a cell and distinguish betweennonmetastatic and metastatic cells

23 Electrical Properties Studies of the cell morphology haveprovided fundamental statement that the cell membraneselective permeability is controlled by electrical charges [3763] The different concentrations of molecules on innerand outer sides of the membrane have created an electricalpotential across themembraneThework by Becker indicatedthat the major charge carriers of biological cell membranewere negative at physiological pH such that the healthy livingcells were found to have a membrane potential within therange of minus60 to minus100mV [64ndash68] The negative sign of themembrane potential indicated that the inner surface of thecell membrane is relatively lower than the immediate exteriorsurface of cell membrane [69] However the cell membranecharge was found to be altered during tumorigenesis due toits abnormalmetabolic transformation [37]Therefore inves-tigation of cell electrical behavior when they are subjected toan electric field can furnish information for the purpose ofcancer cell isolation

In literature two technique are typically used to measureelectrical properties of cells such as impedance spectroscopy[70] and electrorotation (ROT) [71] Both methods providequantitative information of inherent electrical and dielectricproperties of cells such asmembrane capacitancemembraneresistance cytoplasmic conductivity and permittivity Forimpedance spectroscopy it is employed by researches tomeasure cell suspension dielectric properties as a functionof frequency giving a population averaged value for thecell properties Meanwhile measurement for electrorotationis performed on single cell level which is located at thecenter of four electrodes A 90-degree phase excitation signalis applied to the electrode and the cell rotation speed isrecorded Depending on the frequency of ROT excitationsignal the cells will exhibit vary rotation speed based on theircytoplasmic conductivity and permittivity

Becker et al has investigated the differences in dielec-tric properties of metastatic human breast cancer cellline MDA231 erythrocytes and T lymphocytes with ROTmethod The result showed that the metastatic human breastcancer cell line MDA231 had higher membrane capacitance(26 plusmn 42mFm2) than T lymphocytes (11 plusmn 42mFm2) anderythrocytes (9 plusmn 08mFm2) [71] Meanwhile Qiao et alpresented an investigation of impedance for cell suspensionsfor four breast cell line namely MCF-7 (early stage cancercell) MDA-MB-231 (invasive cancer cells) and MDA-MB-435S (late stage cancer cells) with impedance spectroscopymeasurement technique [70] In their studyMaxwell-Wagnertheory was employed to analyze the electrical parameters of asingle cell based on the average result obtained from the exp-eriment The result uncovered that different stages of cancerbreast cells can be distinguish by the conductivity presentedby each cell For instance early stage cancer cell MCF-7

had higher whole cell conductivity and cytoplasmmembranecapacitance (value of 558mScm and 394 120583Fcm2) in con-trast to late stage of cancer cell MDA-MN-435s (value of397mScm and 110 120583Fcm2) Based on the outcome of theseresearches we are clearly informed that each cell line had aspecific electrical signature which could be utilized for iden-tification of cancer cells and differentiation of the pathologystages of malignant cells The technique which employedthe principle of cellrsquos membrane electrical differences forseparation is named as electrokinetic (eg dielectrophoresis)Its details will be discussed in the next section

3 Benchtop Technologies for CTCs Isolation

In general the physical cell-based isolation tools that resear-chers developed for the purpose of benchtop CTCs detectioncan be classified into two types the conventional macroscaleanalytical system and microfluidic devices For macroscalesystem it mainly relies on large laboratory equipment withthe usage of a few milliliters (mL) of cell suspension (egcentrifuge) Since these devices often require a large numberof cells from human and animal models it reduces thelikelihood for the patient and physician to receive the resultsquicker which causes them rarely to be used in point-of-care application To overcome the limitations imposed bymacroscale system vigorous efforts have been undertakento develop robust laboratory test over the past decades Asa result gained from the advances of micro- and nanofabri-cation approaches there is a growing trend towards carryingout microscale laboratory work on a scale of one-tenth toone-thousandth of that macroscale system These miniatur-ization reaction platforms are termed as microfluidic devicesIn contrast to macroscale system the use of microfluidicdevice has offered various advantages such as scalable shorteranalysis time and smaller sample size Furthermore thistechnology enables the actuation of fluid and manipulationof bioparticles at microscale Such feature is especially impor-tant for CTCs research in considering the rare presenceof CTCs within the blood as mentioned previously Bothmacroscale and microfluidic techniques employed for phys-ical based CTCs detection will be discussed in detail in nextsection

31 Macroscale CTCs Detection System

311 Density Gradient Centrifugation Separation of cellularconstituents within blood is widely achieved by densitygradient centrifugationThis technique uses centrifugal forceto separate the cells based on their sedimentation coefficientdifferences According to Stokes law of sedimentation therate of a particlersquos sedimentation is directly proportional toits size and density and relative to the density of suspensionfluid As the mixture of diverse cell sample is subjected tocentrifugation the different types of cells will pass throughthe density gradient at different rates depending on theirdensity resulting in distinct zones appearance (Figure 1)The heavier particles such as RBC and neutrophils (densityof gt1077 gmL) will appear at the bottom while the CTC


Blood sample

Ficoll

Plasma CTCs

WBC

Ficoll

RBC neutrophils

Centrifuge

Figure 1 Principle of density centrifugation separation method Sample is layered on top of a density gradient Ficoll Under centrifugalforce particles move through the medium and density gradient and be suspended at a point in which the density of the particles equals thesurrounding medium

plasma and mononuclear (density of lt1077 gmL) willremain at the top as buffy coat [72]

In fact centrifugation has been employed as early as 1950by Fawcett et al to separate cancer cells from peritoneal fluid[73] A floatation medium is developed in their study tooptimize the cell separation into its distinct layerwithout sub-jecting the particle to high osmotic or ionic stress Althoughthis research provided a favorable result such that four layers(which consist of saline malignant cells albumin solutionerythrocytes and leukocytes) are formed in accordance withthe particles density and the use of albumin as floatationmediumwas costly and uneasy to be prepared [73] Followingthis research another density gradient centrifugationmethodis set forth by Seal for CTCs isolation in which siliconblending oil was used as floatation solution In his studycancer cell is successfully detected in 53 gastrointestinaltract cancer patient samples and 33 breast cancer patientsamples [74] Although both studies showed low separationefficiency they have uncovered the importance of the use ofgradient medium in centrifugation process Consequently ithas led to escalating efforts in search of gradient media mate-rial which provided high separation efficiency Nowadaysdensity centrifugationmedia such as Percoll Ficoll-Hypaqueand OncoQuick are more widely employed in preclinical andclinical cell centrifugation researches

Ficoll-Hypaque is a type of sucrose polymer with highsynthetic molecular weight Its gradient media is mainly usedin the isolation of human mesenchymal stem cells (hMSCs)[75] Despite its popularity various studies reported that thismethod has low CTCs separation efficiency Considerablenumbers of tumor cells and hMSCs are found to accumulatein the lower fraction instead of preferable upper fraction afterdensity gradient separation For instance study conducted byLara et al indicated that the use of Ficoll-Hypaque gradientmedium had mononuclear cells recovery of 57 [26] Sim-ilarly when this gradient medium is employed by both vanBeem et al and Aktas et al for bone marrow mononuclearcell enrichment process the investigated cell recovery rate isbetween 15 and 30 [76 77] They also highlighted thatboth cell load and individual harvesting techniques havegreat impact on the centrifugation performance with FicollFurther study done by Ahmadbeigi et al and Posel et al hassuggested that such an excess cell loss during Ficoll density

gradient centrifugation is a consequence of density medium-related cytotoxicity [75 78]

Meanwhile Percoll density gradient media is made ofcolloidal silica particle suspension Unlike Ficoll-Hypaquemediumwhich is in ready-to-use form the osmolality of Per-coll medium must be adjusted with saline medium to makeit isotonic with physiological salt solution Such a premadegradient allows it to cover a wide density range for isopycnicbanding of all biological particles of interest such as variouscells andmicroorganism In the literature Ellis et al reportedthe use of Percoll has obtained more than 90 purity forthe gradient that yielded populations of mononuclear cellsneutrophils and platelets [79] However this research is incontradiction with recent study done by Chang et al onrelative isolation efficiencies of both Percoll medium andFicoll medium For instance Ficoll density centrifugationshows high number of isolated mononuclear cells (253 plusmn89times 107 cells) compared to thosewith Percoll (136plusmn 66times 107cells) [80] Yet both Percoll and Ficoll share the same pitfallsuch that the blood sample tends to mix with the gradientmedia if the centrifugation did not perform immediately afterapplying the sample to the gradient media Such a conditionis undesirable as it will cause a reduction of therapeuticallyrelevant CTCs cell populations

The ongoing study to increase the CTCs recovery ratefocusing on reducing the cell loss and contamination has ledto the OncoQuick centrifugation system It is a novel tech-nology in which a porous barrier is nestled within the 50mLcentrifuge tube to prevent the lower compartment (separa-tion medium) from mixing with the blood sample prior tocentrifugation Following buoyant density gradient centrifu-gation the cells will be separated and pass through the barrieraccording to their different buoyant densities As previouslymentioned RBCs and the granulocytes have higher buoyantdensities in contrast to other blood cells Thus they arepartitioned below the porous barrierMeanwhile CTCs alongwith the mononuclear lymphocytes will remain above theporous layer which allow them to be easily accessible forsubsequent collection and analysis of CTCs In the literatureOncoQuick has showed a significant improvement in CTCsisolation over centrifugation based on Ficoll and Percoll Forinstance study conducted by Gertler et al indicated thatOncoQuick has higher relative tumor cell enrichment than


Ficoll density gradient centrifugation when the separated cellfractions were evaluated with flow cytometry Despite havinglow cell fraction (95 times 104 mononuclear cells) in contrast to18times 107 cells by Ficoll the tumor cell recovery rate is between70 and 90 for both systems [81] Similarly Konigsberget al have applied the same technique to isolate CTCs fromblood sample of 26 metastatic breast cancer patients TheCTCs were spotted in 692 of blood samples [27]

32 Microfluidic Device

321Microfiltration Microfiltration is a technique of flowingcell sample through an array of microscale constrictions inorder to capture target cells based on size or a combinationof size and cell deformability Several microfilter designs aredeveloped in the literature for benchtop CTCs separationvarying in terms of their blood passing capability and trap-ping efficiency The microfilter developed for the CTCs sepa-ration can be categorized based on their geometrical designAmong themmembrane weir pillar and packed bead-basedmicrofilters are frequently discussed in the literature

For membrane microfilter it consists of a semiperme-able membrane perforated with a 2D array of small holes(Figure 2(a)) Such filters are commercially available in differ-ent pore size and most of the reported membranes for CTCsseparation have pore sizes of 6ndash11 120583mdiameters Noteworthya pore size around 8 120583m in diameter is proved to be optimalfor CTCs retention [82] Since the pore size can be preciselyselected at dimensions commensurate with blood cell exclu-sion membrane based filtration is suited for microfluidicblood enrichment application In early research stage thedead-end flow configuration is commonly employed formicrofiltration sample advection such that the blood flowis perpendicular with membrane surface Particles smallerthan the pore sizes will pass through the membrane andvice versa for the larger particle Despite its simplicity inimplementation process study conducted by Shiau et alshowed that the deposited layer of trapped cells on themembrane in the late filtration period has governed thebuildup of filtration resistance As a result the efficiency ofdevice to isolate cultured cancer cells from the whole bloodsample is reduced [83] To overcome these issues a 3Dmem-brane microfilter which consists of two-layer membrane wasproposed by Zheng et al [11] Both the top and bottom layershave pores defined by microfabrication and the cell captureis realized by the gap between both layers The importantfeature of this design is that the larger pores (9120583m diameter)located on the top layer of microfilter patch are aligned withthe center of corresponding hexagon pattern of the smallerpores (8 120583m diameter) on the bottom later When the CTCcultures from human blood sample flow through the 3Dmembranemicrofilter the smaller cell such as RBC andWBCwill be easily traversed through the gap meanwhile thetumor cells are trapped in the pores of top membrane Thebottom membrane will provide direct force on the trappedtumor cells to counter hydrodynamic force in oppositedirection Subsequently the concentrated tension stress oncell plasma membrane can be greatly reduced in contrast to

the conventional membrane microfilter This device hasdemonstrated capture efficiency of 86 with a throughputof 375mLmin [11] Very recently Zhou et al have reporteda new design of 3D microfilter membrane as illustratedin Figure 2(b) [13] This device has fundamentally differentstructure and filtration principle in contrast to Zheng et alFor instance the pores on the top membranes are five timeslarger than the pores (8 120583m diameter) on the bottom mem-branes Consequently the captured tumor cells will wedgeinto the gap between the top and bottomofmembranes Sincethis device features separable 3D membrane the capturedtumor cells from the healthy donor blood sample spiked withmultiple cancer cell line can be accessed by separating thetwo layers membrane A capture efficiency ranged from 78to 83 and cell viability of 71 to 74 is reported Leveragingthe advantage of two-layer membrane microfilter Yusa etal have developed a palladium filter unit in which a filtercassette (consists of two-layer membrane with 8 120583m sizedpores in bottom layer and 30120583m-sized pores in upper layer) issandwiched in between the upper and lower rings as shownin Figure 2(c) [12] In contrast to the previously discussed 3Dmembrane microfilter in this paper this device allows only arelatively low flow rate (24mLmin) A further analysis withcomputational modeling software has indicated an inversecorrelation between the numbers of pores with the filterrsquos flowrate The recovery rate of spike breast and gastrointestinalcancer cells by this 3D palladium device is sufficiently highwhich isgt85Although these devices are reportedwith highcapture efficiency of CTCs low enrichment factor is showedacross multiple studies To enhance the enrichment factor aswell as eliminate the resistance buildup around membranemicrofilter Lu et al have proposed 2D membrane slot filteras depicted in Figure 2(d) [28] Compared to the commonlyemployed circular pores slots allow easier deformation ofblood cells in their longitudinal direction which facilitateseasier passage of normal blood cells Furthermore the largefill factor of cell also reduces the buildup of flow resistanceduring filtration and thus minimizes the forces exerted oncells Consequently a high viability (gt90) andhigh recoveryrate (gt90) of isolated cancer cells are reported Apart ofchanging the architectural design of membrane microfiltera crossflow configuration is introduced to resolve the prob-lems of pressure buildup during membrane filtration (seeFigure 2(a)) Unlike dead-end flow configuration its filteredflow is parallel to the filtration surface Membrane micro-filter developed by Adams et al using this principle hasdemonstrated the ability to capture more than 98 of MCF-7 cancer cells from a diluted 75mL blood sample given thatthe membrane consists of uniform patterned distribution ofgt160000 pores with diameter of 7 120583m [10]

Apart of membrane basedmicrofiltration weir-type stru-ctures are typically employed as filter element across thewidth of the microfiltration chip It involves an individualbarrier obstructing the flow path to trap most of the CTCsfrom the blood sample while WBC and RBC will passthrough the narrow slit located on the top of barrier as shownin Figure 2(e) In the literature such a device is commonlyused to isolate white blood cells from the whole blood sample[84ndash86] Up to themoment Chung et al are the first and only


Cross-flow filtrationDead-end filtration

(a)

Gap distance

350120583m

25120583m 10120583m 10120583m

0bottom 8120583m

0top 40120583m

Captured tumor cells

Red blood cells

Leukocytes

Parylene C-membrane

(b)

(c)

Slot-shaped filter

(d)

Pump Drain

(e)

Large cells

Small cells

Outlet 2(small cells)

Outlet 1(large cells)

Inlet 1(samples)

Inlet 2(buffer)A

Physicalbarrier

A

Cross-sectional view

GapA998400

A998400

(f)

Figure 2 Continued


Microchannel

Inlet of microchannel

Bead-packed inlet

1mm

(g)

Figure 2 Schematics of various microfiltration mechanisms (a) membrane microfilter (reproduced with permission from [10] copyright2014 The Royal Society of Chemistry) Its fluid flow configuration can be further categorized into two types which is dead-end filtrationand crossflow filtration (b) 3D membrane microfilter with key geometrical parameters labelled The smaller cells can easily traverse throughthe gap while the large cells (eg tumor cells) will be trapped Two types of force are exerted in the trapped cell such that force is caused byhydrodynamic pressure from top and supporting force from bottom membrane (reprinted by permission from Macmillan Publishers LtdScientificReports [11] copyright 2015) (c) 3Dpalladiummembranemicrofilter cassette and its SEM images of filter (reprintedwith permissionfrom [12] copyright 2014 PloS One) The cross-sectional view showing tumor cells will be trapped within the gap of the membranes [13] (d)Membrane slot filter design (e)Weir-type filter (adapted with permission from [14] copyright 2001 American Chemical Society) A silt-typestructure is fabricated within the flow channel to improve the target cells retentionThe smaller weir gap is designed to allow human RBC andplasma to pass through while retaining CTCs (f) Cross-sectional view of diagonally weir-type filtration (reproduced with permission from[15] copyright 2012 John Wiley and Sons) (g) Bead-packed based filtration The microchannel entrance is blocked by packing large sizedbeads Different bead sizes were used to implement a bloodplasma separator at the inlet of the microchannel Subsequently when wholeblood was dropped into the inlet of the microchannel the structure was allowed for the capillary flow of blood through the hetero-packedbeads During this movement of blood the RBC will pass through small pores while big sizes cells such as CTCs will be blocked from flowinginto the channel (reproduced with permission from [16] copyright 2012 The Royal Society of Chemistry)

groupwhich employedweir-basedmicrofiltration to performCTCs isolation from the unprocessedwhole blood cellsTheirproposed weir-type device featured a barrier with heightof 10 120583m and 80 120583m in width across the main channel ata small angle of 10∘ (Figure 2(f)) which effectively allowedthe passage of 99 of blood cells (erythrocytes 6ndash8120583mleukocytes 8ndash10120583m) that passed through the barrier whileimpeding the CTCs (gt10 120583m) This strategy is found toachieve high enrichment ratios (gt2 times 104) and recovery rates(gt95) at the flow rate of 20mLsdothrminus1 [15] In contrast to weir-type structure the pillar structures are much more widelyused in CTCs isolation [29 87ndash89] Similarly to the weir-based microfilter the layout of this device consists of arraysof pillar structures within themain flow channel to allow cellssmaller than the slit to pass through it Using this technologyLin et al (2010) have demonstrated recovery rate of gt90 in57 blood samples from cancer patients [29]

Besides the performances of bead-packed based filtrationfor CTCs isolation process too are discussed in the literatureSuch a method is typically incorporated with the capillary-driven analysis systems whereby a batch of uniform (diam-eter of 45 120583m) [30] and nonuniform beads (diameter rangedfrom 100 120583m to 15 120583m) [16] are packed into a chamber andact as the filter element (Figure 2(g)) Subsequently when theblood is channeled into the filtration chamber inlet blood

cells such as RBC and WBC are allowed to flow through thepacked-bead whereas CTCs will be immobilized within thepacked bed To prevent the cells from binding to the beadsprotein-blocking solution is sequentially introduced beforethe experiment to develop hydrophobicity on the channel andbead surface According to study conducted by Arya et althe overall CTCsrsquo capture efficiency for packed-bed based onpolystyrene and chitosan beads was lower (varied between21 and 40) in contrast to filtration generated by mem-brane pillar and weir-based structure

In general themain advantages ofmicrofiltrationmethodlie in its simplicity and its capability to obtain the fractiona-tion of whole blood in a single pass Additionally it allowsfor the counting of CTC per milliliter of blood andmaintainscell integrity for CTC detection and further characterizationTo achieve high purity and recovery rate with this devicetwo parameters such as the flow rate of the fluid and thecross section of the pores or the gap dimensions requireproper considerationThe fluid flow rate will impact the forceapplied to each cell as it is deformed through a constrictionwhereas the pores or gap dimensions determine the size anddeformability of target cells that can be captured by thefilter The significance of their role in filtration is furtherhighlighted whenmathematical study conducted byMa et aldiscovered the direct correlation between cell lysing and


Flow

Separated particles

Sample inlet

Ultrasonic transducer

Clean fluid inlet

Sample inlet

(a)

Flow

(b)

Figure 3 Illustration of (a) an acoustophoresis device and (b) the particle gradient within the microchannel cross section after passing overthe transducer (adapted with permission from [17] copyright 2007 American Chemical Society)

the applied mechanical trauma (eg shear force) such thatmembrane area with increment exceeding 3 will result inpermanent damage on the cells [90] Since the effectivenessof the isolation mechanism is controlled by the tailoredgap and pores dimensions a few problems have arousedFirstly the mechanical filtration system which separates cellsbased onmicrofilterrsquos geometrical differences is not universalVariation in microfilter architecture is required when the sizeof target cells is changed Besides the heterogeneity of CTCshas caused the smaller cells which exhibit carcinoma char-acteristic from being detected by this device Consequentlya loss in the target cells for further downstream analysisis feasible Moreover the high concentration of blood cellscan easily cause clogging of microscale constriction or filterstructure designed for cell sorting in microfluidic devicescompromising their performance for applications involvingwhole blood samples Noteworthy CTCs are needed tobe removed from membrane microfilter to perform down-stream analysis (such as PCR-assays and cytomorphology)However CTC handling such as aspiration and ejection ofsingle cells into the PCR tube by micromanipulator is timeconsuming Furthermore the nontransparent material ofcertain membrane microfilter has resulted in the need touse upright fluorescence microscopy rather than the usualinverted fluorescence microscopy for manipulation of CTC

322 Acoustophoresis Acoustophoresis is the separation ofparticle using high intensity acoustic waves Its device con-sists of interdigitated piezoelectric transducers which areemployed to generate acoustic standing wave When thecell suspension is subjected to the acoustic field cells willexperience acoustic force Depending on the cell densityand the difference of compressibility properties between celland the surrounding fluid this force can vary by orders ofmagnitude which in turn translates cells toward the pressurenodes (point where the periodic pressure variations are zero)or antipressure node (point with maximum acoustic pres-sure) If the flow rate acoustic force and particle mixture arecorrectly balanced particle gradient will be developed acrossthe channel at its end as illustrated in Figure 3 By introducing

multiple outlet branches into themicrofluidic design the cellscan be separated and guided toward their respective outlet

Following the acoustophoresis success in RBC sorting byPetersson et al [91] the differing of acoustophysical prop-erties of CTCs is investigated as the principle for separationfrom other blood cells Up to the moment there are only twogroups that have published their work on the performanceof acoustophoresis in preclinically CTCs separation A firsteffort to capitalize on this aspect was reported by Augusts-sonrsquos group They have presented an acoustophoresis devicewith trifurcation inlet and outlet to separate three differentprostate cancer cell lines (DU145 PC3 and LNCaP) fromno-cancer control subjectsrsquo blood sample spiked with thementioned cancer cells [32] As previouslymentioned cancercells have higher density and compressibility in contrast tonormal cells Consequently they are found to move towardand align themselves in the center of the channel (pressurenode) while normal blood cells can be detected near thechannel wall (antipressure node) To minimize the influenceof parabolic flow profile which may otherwise affect theefficacy of tumor cells separation from WBC method of cellprealignment on paraformaldehyde (PFA) fixed cells beforeacoustophoretic separation is introduced into their devicearchitecture In contrast to typical nonprealignment method[17] an improvement in cell separation is observed yieldingtumor cells recoveries of 87 and 83 of DU145 and PC3 cellsrespectively with regard toWBC suppression of 993Whenthe device performance on nonfixed sample is evaluatedchanges in intrinsic acoustic properties are experienced bythe WBC such that the WBCs are detected to have highdepletion rate at low acoustic energy However the tumorcapture efficiency is virtually unchanged in contrast to PFAfixation sample that is the device shows an increase inseparation efficiency for both samples in accordance withacoustic energy For instance the average central outlet recov-ery of DU145 cells (nonfixed sample) was 854 and 966at 119864ac of 120 Jm

3 and 188 Jm3 respectively Meanwhile fornonfixed method the average DU145 cell recovery was 361at 119864ac of 66 Jm

3 and 837 at 119864ac of 103 Jm3 Additionally

this research also highlights that no significant difference in


cell viability between acoustophoresis treated and untreatedcells was found for both types of sample This statement isconcordant with study done by Burguillos et al [33] focusingon the effect of acoustophoretic force on prostate tumor cellrsquosviability and proliferation Changes in mitochondrial poten-tial aswell as cell function via their inflammatory response areselected as themeasurement parameter for cell viabilityTheirexperimental result depicted no changes in mitochondrialpotential due to acoustophoresis Also the cell propertiesare unaltered after acoustophoretic processing as measuredby cell turnover assays as well as inflammatory cytokineresponse up to 48 hours following acoustophoresis [33]

Although both studies suggested that the microchannelacoustophoresis can be used for effective CTCrsquos detectionthis technology is still under preclinical development stageThe main challenge for acoustophoretic microfluidic deviceto be clinically relevant is that the current system is unable togenerate data for blood samples containing much lower con-centrations of cancer cells Furthermore in accordance withmost microfluidic systems that handle particles acousto-phoresis systems are also challenged by clogged systems andoccluded channels when the cell concentration is increasedSuch a condition can be countered by extending the length ofchip the optimization of flow rate within the microchanneland the increase of the acoustic force Although the increaseof acoustic force can result in significant improvement oftumor cell recovery which is low in concentrations Peterssonet al [17] have reported that the acoustic force cannot beincreased without limit as high acoustic pressure can causebiological cells lysis Henceforth there is a need for acousticradiation forcemodel which can be used to calculate the forcemagnitude to be modified to accurately represent the settingin a microfluidic chamber

323 Dielectrophoresis Dielectrophoresis (DEP) is one of thephenomena commonly grouped as part of AC electrokineticIt refers to the particles motion under nonuniform electricfield Depending on the conductivity and permeability of thecells as well as its suspending medium the cells can exhibitattractive and repulsive response at a given electric field fre-quency For instance if the particle is more polarizable thanits surrounding medium the particle will experience a netforce toward the high electrical field gradient Such a condi-tion is termed as positive-DEP (pDEP)Meanwhile negative-DEP (nDEP) happenswhen the surroundingmedium ismorepolarizable than the particle which causes the particle tomove toward the low electrical field region As alluded inprevious section CTCs feature high surface areas which inturn give them larger capacitance per unit area in contrast tothe normal cells These factors have contributed to the dif-ferences in cell dielectric phenotypes thus directly affectingtheir motion under electric field in contrast to the normalcells The carcinoma cells will exhibit nDEP and vice versa tothe normal cells In the literature the benchtop DEP devicehas been successfully used to isolate oral cancer [92 93]colon cancer [94] breast cancer [31] lung cancer and prostatecancer cells [95] with a recovery rate of 70ndash90 Followingthese successful separation and subsequent preliminary DEP

analyses of cancer cells a multitude of DEP-based clinicalCTCs isolation studies are ensued

In order to convey the proper control of DEP force onblood sample it is necessary for researchers to select ordevelop appropriate microelectrode designs which fix theirstudies requirements For the existing DEP devices the 2Dmetallic microelectrodes with various geometries such asinterdigitated [96 97] castellated [98] curved [31 99] spiral[100] and ring shape [101] are normally patterned within thechamber of microfluidic device to generate nonuniform elec-tric field that is of desirable strength for the CTCs isolationThese microelectrode designs are easily to be fabricated withcommon lithograph technique Nevertheless they are com-patible with on-chip analysis for continuous manipulation ofblood samples However there is a pitfall in this techniquesuch that electrical fields are found to decay exponentiallywith the distance away from the electrode It has causedcancer cells away from the electrode less controlled by DEPforce To circumvent these issues Lewpiriyawong et al havesuggested the use of 3D electrode deposited at the channelwall to generate electrical field covering the whole volume ofchannel as illustrated in Figure 4(a) [18] This configurationhas been successfully demonstrated for cell sorting in studydone by Wang et al (2009) [102] Their results show animprovement of 15 in cell recovery rate in contrast to theconventional microelectrode However due to its fabricationcomplexity such a configuration is yet to be tested withCTCs In the latest research Huang et al developed aDEP-based technology known as contactless DEP (cDEP)which replaced the metallic electrodes by fluidic electrodechannels (see Figure 4(b)) [19]The cDEP is capitalized on thesensitivity of traditional DEP while eliminating challengessuch as bubble formation electrode delamination expensivefabrication and electrode sample contaminationThe experi-mental results have indicated that the cervical carcinoma cellsare successfully isolated from the concentrated RBC with arecovery rate of 645

The major advantage of DEP compared to other separa-tion schemes is that the variability in the frequency responseof cells is selective enough for DEP microsystems to monitortherapeutic efficacy and to account for constantly evolvingtumor phenotypes However some considerationmight needto be taken into account during the design of DEP devicefor blood sample In practice most of the reported on-chipDEP separation microfluidic devices require the use of lowconductivity medium in order to generate pDEP force to trapcell of interest Noteworthy blood is a very high conductivitymedium As a result it might cause cells to experience nDEPmost of the time and thus influence separation performancethus influencing the cell separation performance as well asthe purity outputThough Gascoyne et al [96] suggested thatCTCs enrichment through blood dilution allowsDEPdevicesto have optimal recovery Leu and Liao [103] have reportedthat the actual extraction efficiency drops for 20 if the diluteratio 1 3 of whole blood sample was conducted Despite theuse of nDEP that is able to levitate particles above the elec-trodes and thus protects vulnerable biological particles fromhigh electric fields RBC could also be irreversibly damaged


Outlet C Outlet DOutlet B

Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)

PDMSCells introduced

using pipette

ITO

GlassDirection of cell movementDirection of relay switching

Red blood cell (RBC)HeLa cell

Inertial

14

8

Voltage applied

(b)

Figure 4 Schematic of DEP cell isolation devices (a) DEP microchip with 3D side wall microelectrode By imposing an AC voltage onthe side wall microelectrodes cells will experience repulsive (nDEP) or attractive forces (pDEP) depending on their relative polarizabilitybetween cells and fluid (adapted with permission from [18] Copyright 2011 American Chemical Society) (b) DEP system with contactlessmicroelectrode This method is capable of manipulating cells without direct contact between electrodes and sample The schematic of cDEPplatform design is showed such that the electrode is inserted into two conductive microchambers and is separated from the microfluidicchamber by thin insulating barriers Consequently cell adherance to the microchip can be prevented To accumulate the target cell onto thecentral microelectrode a stepping electric field is generated such that the applied electric field is subsequently switched between the adjacentelectrode pair via relays Cell which experienced pDEPwill be guided along the direction of stepping electric field toward the center of circularelectrode (adapted with permission from [19] copyright 2012 Journal of Medical and Biological Engineering)

as a means of cell rupturing if electric field much higher than012MVm is applied [104]

324 Hydrodynamic Sorting Hydrodynamics is one of theimportant forces that govern the behaviour of microfluidicdevices Its fundamental concept is represented with Navier-Stokes equation which describes the motion of fluid sub-stances at a given point in space and time Since the build-in-channels of microfluidic devices have the dimension ofless than 1mm the flow generated within them is completelylaminar at Reynolds number below 2000 Reynolds number(Re) is a dimensionless parameter representing the inertial toviscous forces ratio in a flow At this low Reynolds numbera particle has been expected to follow fluid streamlinesuperposed with its intrinsic Brownian motion due to theirdifferences in size and density [105] By resorting to thisprinciple into the designated microdevice geometries thecell can be separated according to the different flow rates ofparallel fluid flow into the desired outlet Such a techniqueis termed as hydrodynamic sorting and it can be furtherclassified into 3 types which are pinched flow fractionationdeterministic lateral displacement and inertia separation

Pinched flow fractionation acts according to the Zwei-fach-Fung effect such that the blood cells are separatedthrough lateral migration due to the asymmetric bifurcationof laminar flow [106] Its microfluidic design features a mainchannel with multiple narrow channels branching at the endof pinch section as showed in Figure 5(a) Such a channeldesign has resulted in the difference in flow velocity as wellas particle volume fraction between the branches Dependingon the sizes and density of the particle they will be pushedinto specific flow streamlines following by separation Inthe literature this method is not much exploited in CTCsisolation Geislinger and Franke are the first and only groupwho demonstrated continuous separation of cultured can-cer cells (MV3-melanoma cell line) and blood cells usingpinching flow as driving force [34] The proposed microflui-dic device worked at much lower Reynolds numbers (Re lt1) and an expansion with smooth broadening was added toincrease the absolute distances to facilitate the isolation Theefficiency of 100 and medium purity of 66 are yieldedin blood suspension with 9 of haematocrit Although theresult showed excellent cell viability study has reported thatsuch a technique is more suitable in a dilute suspension ofblood sample To improve the separation performance this


Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)

(C)(B) Pinched segment

(C) Branch A (D) Branch D5mm

(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs

Cross-sectional viewOuter inletInner inlet

Outer wallInner wall

Outer outletInner outlet

1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued


Cancer cellsWhite blood cells

Red blood cellsParticle flow

Focusing flow

Cancer ceWhit

Reicle flow

s1

s2

Expansion regionContraction region

Focusing flow

Particle flow

Y

X

107mm 500120583m

Dean flowDean drag force (FD)Inertial lift force (FI)

(d)

Figure 5 Types of hydrodynamic cell sorting (a) Pinched flow fractionation In both microfluidic design by (i) and (ii) Takagi et al [20]multiple branch channels with different channel dimensions were arranged at the end of the pinched segment thus resulted flow ratedistribution to each channel was different Cell would then enter into their respectively outlet in accordance to their size (Reprinted withpermission from [20] Copyright 2005 The Royal Society of Chemistry) As illustrated in (iii) [21] the smaller particle will closely followthe fluid streamline and move toward the upper portion of the exit area while the larger particles move closer to the center This is dueto the smaller cells tends to move faster under hydrodynamic force which resulted them to press closer to the wall as the flow ratio increase(Reprinted with permission from [21] Copyright 2013 JohnWiley and Sons) (b) Deterministic lateral displacementThe presence of array ofmicroposts (which each row of posts is slightly offset laterally with respect to the previous row)will cause the cells below critical hydrodynamicdiameters (such as WBC and RBC) to follow the streamlines cyclically through the gaps Meanwhile cell above critical hydrodynamic postssuch as CTCs will be moved by lateral drag into sequential streamline at each post (reprinted with permission from [22] copyright 2013 AIPPublishing LLC) (c) Inertial separation When the blood sample is pumped into the spiral channel the centrifugal acceleration of fluid flowwill result in the formation of two symmetrical counter-rotating vortices across the channelThe smaller cell such as RBC andWBCwill movealong the vortices toward the inner wall and back to the outer wall while larger CTCs will focus along the inner wall due to the additionalstrong inertial lift forces (reprinted by permission from Macmillan Publishers Ltd Scientific Reports [23] copyright 2013) (d) Contractionexpansionmicrofluidic device Dean drag forces are induced at the entrance of contraction region and thus result in blood sample which flowthrough this region to have an influence by inertial lift force RBC andWBCwill move toward 119904

2while cancer cells move toward 119904

1(reprinted

with permission from [24] copyright 2013 American Chemical Society)

technique is normally integrated with dielectrophoresis andmagnetophoresis method

Apart of pinched flow fractionation another promisingcandidate of hydrodynamic based separation is deterministiclateral displacement This device is composed of an array ofposts (see Figure 5(b)) inwhich each row is displaced at a dis-tance Δ120582 from the previous row Similar to microfiltrationthe dimension for both gap distance and obstacles size playspivotal role in the separation processWhen a given particle issmaller than the critical size the particle will flow accordingto themainstream line around the obstacle with nondeviatingtrajectory perpendicular to streamline However if the parti-cle is bigger than the critical size it will collide with the obsta-cle and results in streamline switch with a constant angleSuch a motion is termed as lateral displacement By tuningthe size of obstacle the gap between them and the shift in thearray particles from different size can be separated laterallyUsing this technique Davis et al showed how the whole

blood could be continuously separated into its constituentsof erythrocytes leukocytes and plasma [105] Despite its highsensitivity this study has highlighted that the flow rate usedin the device is relatively low compared to what would benecessary for rare cells sorting To investigate the optimalflow rate for separation based on deterministic displacementLoutherback et al have demonstrated the isolation of CTCsfrom blood sample within a long flow chamber filled withmirrored array of 58120583m triangular posts with 42 120583m gapsTheir result presented a very good CTCs capture efficiency(gt85) at volumetric flow rate of 10mLmin with no effecton cell viability [22]

Inertial separation is another method which stands alonetomanipulate particles in a continuous phase flow field Con-trary to both pinching flow fractionation and deterministiclateral displacement this method dominates the fluid inertiallift force to accelerate the flowing particle to their preferentialequilibrium points Such a phenomenon can be observed in


the spiral microchannel The presence of curvature structurewill generate rotational flow within the channel and thusresult in the formation of two symmetrical counter-rotatingvortices (top and bottom) across the channel cross sectiontermed as Dean vortices [106] This motion entrains particleto move back and forth along the channel width Coupledwith inertial lift force the larger particles (such as tumor cells)will occupy a single equilibrium position near the inner wallwhile the smaller particle (red blood cells and white bloodcells) will migrate to the outer half of the channel resulting inthe formation of distinct particle streams which are collectedin two separate outputs Recently a spiral device with 500120583mwide and 160 120583m high using Dean flow induced motionwas developed by Hou et al for CTCs isolation purposeWhen the device performance was tested with 3mL bloodsuspension spiked with cancer cells (sim20 hematocrit) itachieved high cancer cell recovery of gt85 and high cellviability (gt98) Besides followed by preliminary clinicaltesting this device too had successfully detected CTCs in allblood samples collected from 20 patients withmetastatic lungcancer ranging from 5 to 88 CTCs per mL [107] As opposedto Hou et al Warkiani et al investigated the performanceseparation within a spiral channel with trapezoidal crosssection [35] Their device had shown to achieve recoveryrate of 80 for 75mL of blood within 8 minutes Similarto Hou et al they also verified their method successfullybeing applicable to patient samples However such a devicerequires proper geometrical design For instance a deeperinner wall compared to the outer will cause strong vorticesforces to be formed at the inner side of the channel resultingin all particles to be trapped despite their different in sizeand flow rate [108] Apart of spiral structure Lee et al haddemonstrated their contraction-expansion array microchan-nel device (Figure 5(d)) to be able to sort different cancer celllines out of the whole blood suspensions with a very highcapture efficiency up to 995 [24 109]

33 Summary of Device Performance For benchtop CTCsrsquodetection device it is necessary to analyze and optimize thedevicesrsquo performance before they are employed as clinicaldiagnostic tools The key performance metrics which arewidely highlighted in the literature included capture effi-ciency (or recovery rate) throughput viability and purityCapture efficiency refers to the fraction of captured targetcells relative to the total captured cells It is usually expressedas a percentage ()Thismeasurement is important in access-ing the total of CTCs which has been lost in the isolationprocess A high capture efficiency represents less cell lossesand thus provides clinician accurate information about theamount of CTCs discovered from patientsrsquo blood sampleThroughput indicates the speed at which the system canprocess a sample In most experiments high throughputdevice is favored by most researches as it allows a givensample to be analyzed within the short timeThis parameter istypically reported as either the volumetric flow or number ofcells processed per second Meanwhile viability can be bestdescribed as the proportion of cells that remain ldquoaliverdquo andldquoviablerdquo This viability is commonly assayed by dye exclusiontechniques where cells are incubated with a dilute solution of

dye which only enters dead cells Maintenance of cell viabilityis an imperative feature of cell separation procedure Forinstance the capture living cells can be used for downstreamphenotypic and genotypic analyses Furthermore the cellbehavior can be correlated with cell number providing amore accurate picture of cell Purity is defined as the percen-tage of cells in separated population that are detected ashaving certain design characteristics It is an important indi-cator for the purpose of downstream analysis as a high purityindicates the cell subsets are not contaminated by nontargetcells Subsequently it will increase the sensitivity of assayused in postseparation analysis as well as provide accurateinformation of targeted cell Purity assessment is typicallyconductedwith a flow cytometer inwhich target cells labelledwith fluorescent markers are analyzed and the proportion ofeach cell type in sample is calculated

4 Conclusion and Future Outlook

In this review paper we have highlighted the development ofseveral methods that exploit the physical characteristic of thecells to isolate and differentiate CTCs Although these devicesdemonstrate significant progresses for CTCs isolation thedevelopment of these technologies is still in diverging phasein which methods described above are still at the proof ofconcept levelThere are some challenges that remain for thesedevices to be fully employed in point-of-care applicationincluding the usage of cultured cancer cell line the used ofwhole blood sample and low throughput

Cultured cancer cell line is themostwidely used in bench-top morphological model for CTCs rarely yet on clinicalsamples Its model is generated by isolating a tumor cell froma surgery sample and growing them in a controlled artificialenvironment Unlike primary tumors which can only bemaintained for a relatively short period of time in a reformu-lated culture system [110] cancer cell line is immortalized andgenetically modified to proliferate indefinitely Consequentlyit gives rise to a clonal population and thus allows variousscientific experiments to be carried out on these geneticallyidentical cancer cells In spite of their important contributionto cancer biology several drawbacks are reported in theliterature Review written by both Okita and Yamanaka [111]and Holmberg and Perlmann [112] has indicated that tumordoes not proliferate at the same rate as cultured cells In factcultured cells are grown rapidly with doubling time muchshorter than those of cancer cell in vivo When culturedcancer cells are incubated in a growth promoting solutionfor a long period of time they will induce occurrence ofgenomic changes such as copy number variations as wellas transcriptomic drift This condition undoubtedly affectsheterogeneity of cancer cell line and was proven in studyconducted byAuman andMcLeod [113] Henceforth the sep-aration efficiency will be lower when the device is tested onthe clinical sample in contrast to the cancer cell line To tacklethis problem effort should be made to develop new cell linesthat exhibit the genomic and transcriptomic heterogeneity ofcancer cell Readers are encouraged to refer to Holmberg andPerlmann paper for more information [112]


As previously mentioned CTCs can be detected in thebloodstream as early as before a primary tumor is detectedwith conventional clinical screening methods As a resultcomplete blood analysis is of prime interest for most CTCsdetection devices Although the use of whole blood sampleis preferable for device performance evaluation fractioningvarious target components from it has been a technical chal-lenge due to its massive number and wide diversity of celltype Up to the moment only two techniques (such as mem-brane microfilter device [15 28 29] and inertial separationdevice [24 34]) are reported with high capture efficiencywhen the devices are tested with the whole blood samples(refer to Table 1) Majority of techniques have demonstrateda dramatic reduction in capture efficiency when the wholeblood sample is employed To circumvent this problem theblood samples are generally diluted with isotonic diluentbefore proceeding to benchtop separation process Dilutionreduces the concentration of cell per unit volume and thusallows for rapid detection of certain particles from densecolonies In fact the presence of CTCs is extremely rare 0to 1 cells per millimeters of whole blood Therefore the useof dilution buffer will be able to decrease the number ofCTCs within a sample which in turn prolong the analysistime Furthermore study conducted by Takaori in 1979 hasindicated that excessive dilution will cause a progressivedecrease in blood sample pH Since blood cells respondquickly to changes in microenvironment their biologicalcharacteristic might change with regard to the reductionof pH For instance recent mathematical model developedby Wolf and DeLand has indicated that hematocrit varieswith the change of blood pH [114] In order to allow wholeblood sample to be analyzed with bench-separation deviceintegration of multiple function such as enrichment anddetection method onto single chips should be attemptedThe enrichment technique (eg magnetophoresis) can beemployed in first stage to increase the sensitivity of the assayby separating RBC from a human whole blood sample Thisis followed by a detection step (eg dielectrophoresis oracoustophoresis) to separating target rare cells out from othernucleated cells and residual RBCs

Multiples studies have suggested that the use of microflu-idic devices are extremely attractive for blood analysis as theseplatforms allow miniaturization and integration of complexfunctions As a result the complete lab for blood analysis isfeasible to be brought to patientrsquos bedside Besides microflu-idic revolves around the precise manipulation of fluid flowwithin channels in which the dimension of the channelcross section is smaller than 1mm2 Therefore only minuteamounts of blood are needed for analysis and repetitivesampling at multiple time points can be conducted Howeverthere is a drawback of using these devices such that mostof them are still rather limited in throughput As shown inTable 1 the throughput is achievable by most microfluidicdevices lying around processing of 2 to 3mL of bloodper hours Noteworthy these devices are normally operatedslowly to maintain separation efficiency However for diag-nostic application 75mL of whole blood sample from patie-nts is typically employed for CTCs enumeration Henceforththe flow rate required for optimum throughput is probably

insufficient as it will take hours to complete the analysis Inthis case the separation time is increased which leads to aloss of cell viability Consequently there is a need for resear-chers to develop a microfluidic architecture which allow forhigh throughput as well as enhancing the separation effi-ciency and viability of CTCs

In fact a few microfluidic techniques discussed in thispaper has been commercialized for clinical diagnostic deviceto isolate live CTCs from epithelial and nonepithelial malig-nancies These devices include ApoStream and ScreenCell

ApoStream system as illustrated in Figure 6 has beenlaunched by ApoCell laboratories in 2010 [115] This tech-nology leverages the dissimilarities of electrical propertiesof different cell types to isolate CTCs with the assistance ofhydrodynamic force to position the cell at a defined level in afluid velocity gradient Interdigitated electrodes are fabricatedon the chamber floor where the mixed cell population flowsover the electrode structure Since CTC population expressespositive DEP it is pulled along the floor flowing near theelectrode plane and collected through a port located in thechamber floor while the other cells (exhibiting negativeDEP)are levitated and carried away by the eluent Validation studyof ApoStream system conducted by Gupta et al using humanblood sample spiked with breast cancer cell line (MDA-MB-231) has successfully demonstrated a cell viability of 976with recovery efficiency of 866 [25] In recognition of itspotential in CTCs isolation in December 2014 ApoStream isreported to be employed in BEACON (a Phase 3 open-labeland multicenter study of Etirinotecan pegol versus treatmentof physicianrsquos choice (TPC) in patients with metastatic breastcancer) to isolate CTCs for endpoint biomarker analysis [116]

The ScreenCell device is developed to support down-stream analysis of captured CTCs such as immunohisto-chemistry and nucleic acid isolation This device consists ofa filtration reservoir with a microporous membrane filter in aremovable nozzle holder The whole blood sample is neededto be diluted with standardized filtration buffer before it isplaced on the filtration reservoir for 3 minutes sample pro-cessing time To improve the throughput vacuum is appliedto the outflow side of the filtration membrane Similar to theaforementioned membrane microfilter operating principleCTCs will be retained on the surface of filtration membranewhile small nucleated blood cell can easily pass through thefiltrationmembraneThemembrane with captured CTC cellscan be removed from the device in the end of filtrationprocess for further manipulation In 2011 ScreenCell studyconducted by Desitter et al on whole blood sample spikedwith lung cancer cells has demonstrated a high recovery rateof 91 and 74 for 5 and 2 cancer cells respectively permL of blood [117] This device is further tested by Desittergroup in enumeration of CTCs from 23 cutaneousmelanomapatients (CTCs does not express EpCAM) for the purposeof cytological analysis and analysis of genetic mutations Thepostsurgery median of CTC recovery is 1 CTC per 2mL with15 of 23 patients being tested to be CTC-positive

In conclusion CTCs play an important role in both theresearch lab and the clinic Due to their role in metastasisinformation acquired from CTCs has the potential to assistin prognosis as well as helping to define individualized

16 BioMed Research InternationalTa

ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]

Microfiltration(poretype)

Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter

(i)73plusmn13(fixedMCF

-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]

Microfiltration(3Dmem

brane)

Hum

anprostateadenocarcino

ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]

Microfiltration(pillar

type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]

Microfiltration(hetero-packed

bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Hydrodynamicsorting(pinched

flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting

(deterministiclateral

displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]

Hydrodynamicsorting(in

ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to

12mLh(processing42times10

7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]


Elute buffer flow PBMCs

CTCs Waste

Sample injection

CTC collection

DEP signal generatorGear pump

Buffer inlet

Buffer Syringe drive

Flow chamber

injection Sample vial Syringe

drive aspirate

Waste

Buffer outletParticle counter

Figure 6 Illustration of ApoStream device The elution buffer is introduced at the upstream end of the flow chamber with a computer-controlled gear pumpThe blood sample is injected with a high precision syringe pump at a low flow rate into the bottom of the flow chamberto reduce the cell levitation and to ensure that cells stay within effective DEP field Under DEP field the DEP forces will attract cancer cellstoward the electrodes on the chamber floors and vice versa to others cells Cancer cells will withdraw through the collection port which islocated close to the chamber floor Meanwhile other blood cells will be levitated and flow into the waste container via a second outlet port(reprinted with the permission from [25] copyright 2012 AIP Publishing LLC)

therapeutic regimensWhilemethods based onEpCAMenri-chment have brought CTC to point of care applicationNewertechnologies which can facilitate CTC isolation as well assupport downstream analysis techniques are in dire needThefeatures of user friendly and flexibility should be adapted intothe CTC isolation device design consideration to make thetechnology available to more labs and clinic

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgment

This study is financially supported by University of MalayaHigh Impact Research Grant (UMCHIRMOHEENG44)fromMinistry of Higher Education Malaysia

References

[1] N Gaudin and V Terrasse Latest World Cancer StatisticsGlobal Cancer Burden Rises to 141 Million New Cases in 2012Marked Increase in Breast Cancers Must Be Addressed vol 2014International Agency for Research on Cancer Lyon France2013

[2] H Esmaeilsabzali T V Beischlag M E Cox A MParameswaran and E J Park ldquoDetection and isolation ofcirculating tumor cells principles and methodsrdquo BiotechnologyAdvances vol 31 no 7 pp 1063ndash1084 2013

[3] B L Ziober M G Mauk E M Falls Z Chen A F Zioberand H H Bau ldquoLab-on-a-chip for oral cancer screening anddiagnosisrdquo Head and Neck vol 30 no 1 pp 111ndash121 2008

[4] Y Husemann J B Geigl F Schubert et al ldquoSystemic spreadis an early step in breast cancerrdquo Cancer Cell vol 13 no 1 pp58ndash68 2008

[5] M Farooqui M A Hassali A Knight et al ldquoA qualitativeexploration of Malaysian cancer patientsrsquo perceptions of cancerscreeningrdquo BMC Public Health vol 13 article 48 2013

[6] D Antolovic L Galindo A Carstens et al ldquoHeterogeneousdetection of circulating tumor cells in patients with colorec-tal cancer by immunomagnetic enrichment using differentEpCAM-specific antibodiesrdquoBMCBiotechnology vol 10 article35 2010

[7] R F Swaby and M Cristofanilli ldquoCirculating tumor cells inbreast cancer a tool whose time has come of agerdquo BMCMedicine vol 9 article 43 2011

[8] J E Allen and W S El-Deiry ldquoCirculating tumor cells andcolorectal cancerrdquo Current Colorectal Cancer Reports vol 6 no4 pp 212ndash220 2010

[9] T M Geislinger and T Franke ldquoHydrodynamic lift of vesiclesand red blood cells in flowmdashfrom Fahraeligus amp Lindqvist tomicrofluidic cell sortingrdquo Advances in Colloid and InterfaceScience vol 208 pp 161ndash176 2014

[10] D L Adams P Zhu O V Makarova et al ldquoThe systematicstudy of circulating tumor cell isolation using lithographicmicrofiltersrdquo RSC Advances vol 4 no 9 pp 4334ndash4342 2014

[11] S Zheng H K Lin B Lu et al ldquo3D microfilter device forviable circulating tumor cell (CTC) enrichment from bloodrdquoBiomedical Microdevices vol 13 no 1 pp 203ndash213 2011

[12] A Yusa M Toneri T Masuda et al ldquoDevelopment of a newrapid isolation device for circulating tumor cells (CTCs) using


3Dpalladiumfilter and its application for genetic analysisrdquoPLoSONE vol 9 no 2 Article ID e88821 2014

[13] M D Zhou S Hao A J Williams et al ldquoSeparable bilayermicrofiltration device for viable label-free enrichment of circu-lating tumour cellsrdquo Scientific Reports vol 4 article 7392 2014

[14] H Hisamoto T Horiuchi M Tokeshi A Hibara and TKitamori ldquoOn-chip integration of neutral ionophore-based ionpair extraction reactionrdquoAnalytical Chemistry vol 73 no 6 pp1382ndash1386 2001

[15] J Chung H Shao T Reiner D Issadore R Weissleder and HLee ldquoMicrofluidic cell sorter (120583FCS) for on-chip capture andanalysis of single cellsrdquo Advanced Healthcare Materials vol 1no 4 pp 432ndash436 2012

[16] J S Shim and C H Ahn ldquoAn on-chip whole bloodplasma sep-arator using hetero-packed beads at the inlet of amicrochannelrdquoLab on a ChipmdashMiniaturisation for Chemistry and Biology vol12 no 5 pp 863ndash866 2012

[17] F Petersson L Aberg A-M Sward-Nilsson and T LaurellldquoFree flow acoustophoresismicrofluidic-basedmode of particleand cell separationrdquo Analytical Chemistry vol 79 no 14 pp5117ndash5123 2007

[18] N Lewpiriyawong K Kandaswamy C Yang V Ivanov and RStocker ldquoMicrofluidic characterization and continuous separa-tion of cells and particles using conducting poly(dimethyl silox-ane) electrode induced alternating current-dielectrophoresisrdquoAnalytical Chemistry vol 83 no 24 pp 9579ndash9585 2011

[19] C-T Huang T G Amstislavskaya G-H Chen H-H ChangY-H Chen and C-P Jen ldquoSelectively concentrating cervicalcarcinoma cells from red blood cells utilizing dielectrophoresiswith circular ITO electrodes in stepping electric fieldsrdquo Journalof Medical and Biological Engineering vol 33 no 1 pp 51ndash592013

[20] J Takagi M Yamada M Yasuda and M Seki ldquoContinuousparticle separation in a microchannel having asymmetricallyarranged multiple branchesrdquo Lab on a ChipmdashMiniaturisationfor Chemistry and Biology vol 5 no 7 pp 778ndash784 2005

[21] J F Ashley C N Bowman and R H Davis ldquoHydrody-namic separation of particles using pinched-flow fractionationrdquoAIChE Journal vol 59 no 9 pp 3444ndash3457 2013

[22] K Loutherback J DrsquoSilva L Liu A Wu R H Austin and J CSturm ldquoDeterministic separation of cancer cells from blood at10mLminrdquoAIP Advances vol 2 no 4 Article ID 042107 2012

[23] H W Hou M E Warkiani B L Khoo et al ldquoIsolation andretrieval of circulating tumor cells using centrifugal forcesrdquoScientific Reports vol 3 article 1259 2013

[24] M G Lee C Y Bae S Choi H-J Cho and J-K Park ldquoHight-throughput inertial separation of cancer cells from humanwhole blood in a contraction-expansion arraymicrochannelrdquo inProceedings of the 15th International Conference onMiniaturizedSystems for Chemistry and Life Sciences (MicroTAS rsquo11) pp2065ndash2067 Seattle Wash USA October 2011

[25] V Gupta I Jafferji M Garza et al ldquoApoStream a newdielectrophoretic device for antibody independent isolation andrecovery of viable cancer cells frombloodrdquoBiomicrofluidics vol6 no 2 Article ID 024133 2012

[26] O Lara X Tong M Zborowski and J J Chalmers ldquoEnrich-ment of rare cancer cells through depletion of normal cells usingdensity and flow-through immunomagnetic cell separationrdquoExperimental Hematology vol 32 no 10 pp 891ndash904 2004

[27] R Konigsberg E Obermayr G Bises et al ldquoDetectionof EpCAM positive and negative circulating tumor cells in

metastatic breast cancer patientsrdquo Acta Oncologica vol 50 no5 pp 700ndash710 2011

[28] B Lu T Xu S Zheng A Goldkorn and Y-C Tai ldquoParylenemembrane slot filter for the capture analysis and culture ofviable circulating tumor cellsrdquo in Proceedings of the 23rd IEEEInternational Conference on Micro Electro Mechanical Systems(MEMS rsquo10) pp 935ndash938 Wanchai Hong Kong January 2010

[29] H K Lin S Zheng A J Williams et al ldquoPortable filter-basedmicrodevice for detection and characterization of circulatingtumor cellsrdquo Clinical Cancer Research vol 16 no 20 pp 5011ndash5018 2010

[30] C Arya J G Kralj K Jiang et al ldquoCapturing rare cells fromblood using a packed bed of custom-synthesized chitosan mic-roparticlesrdquo Journal of Materials Chemistry B vol 1 no 34 pp4313ndash4319 2013

[31] H-SMoonKKwon S-I Kimet al ldquoContinuous separation ofbreast cancer cells from blood samples using multi-orifice flowfractionation (MOFF) and dielectrophoresis (DEP)rdquo Lab on aChip vol 11 no 6 pp 1118ndash1125 2011

[32] P Augustsson C Magnusson M Nordin H Lilja and TLaurell ldquoMicrofluidic label-free enrichment of prostate cancercells in blood based on acoustophoresisrdquo Analytical Chemistryvol 84 no 18 pp 7954ndash7962 2012

[33] M A Burguillos C Magnusson M Nordin et al ldquoMicrochan-nel acoustophoresis does not impact survival or function ofmicroglia leukocytes or tumor cellsrdquo PLoS ONE vol 8 no 5Article ID e64233 2013

[34] T M Geislinger and T Franke ldquoSorting of circulating tumorcells (MV3-melanoma) and red blood cells using non-inertialliftrdquo Biomicrofluidics vol 7 no 4 Article ID 044120 2013

[35] M E Warkiani G Guan K B Luan et al ldquoSlanted spiralmicrofluidics for the ultra-fast label-free isolation of circulatingtumor cellsrdquo Lab on a Chip vol 14 no 1 pp 128ndash137 2014

[36] K A Rejniak ldquoInvestigating dynamical deformations of tumorcells in circulation predictions from a theoretical modelrdquoFrontiers in Oncology vol 2 article 111 2012

[37] I Dobrzynska E Skrzydlewska and Z A FigaszewskildquoChanges in electric properties of human breast cancer cellsrdquoJournal of Membrane Biology vol 246 no 2 pp 161ndash166 2013

[38] V A Loiko G L Ruban A G Olga V V Berdnik and N VGoncharovaMononuclear Cells Morphology for Cells Discrimi-nation by Angular Structure of Scattered Light Electromagneticand Light Scattering Bodrum Turkey 2007

[39] K A Gutowski J L Hudson andD Aminoff ldquoFlow cytometricanalysis of human erythrocytes I Probed with lectins andimmunoglobulinsrdquo Experimental Gerontology vol 26 no 4 pp315ndash326 1991

[40] E C Lynch ldquoPeripheral blood smearrdquo in Clinical Methods TheHistory Physical and Laboratory Examinations H K WalkerW D Hall and J W Hurst Eds Butterworths Boston MassUSA 1990

[41] J Seifter D Sloane and A Ratner ldquoBloodrdquo in Concepts inMedical Physiology B Sun Ed p 128 Lippincott Williams ampWilkins Philadelphia Pa USA 2005

[42] D H Cormack Essential Histology Lippincott Williams ampWilkins Philadelphia Pa USA 2001

[43] P Pal Textbok of Practical Physiology Orient Blackswan 2006[44] L E Eichelberger M O Koch J N Eble T M Ulbright B E

Juliar and L Cheng ldquoMaximum tumor diameter is an inde-pendent predictor of prostate-specific antigen recurrence inprostate cancerrdquo Modern Pathology vol 18 no 7 pp 886ndash8902005


[45] RHass andC Bertram ldquoCharacterization of humanbreast can-cer epithelial cells (HBCEC) derived from long term culturedbiopsiesrdquo Journal of Experimental amp Clinical Cancer Researchvol 28 no 1 article 127 2009

[46] M P Wong ldquoCirculating tumor cells as lung cancer biomark-ersrdquo Journal of Thoracic Disease vol 4 no 6 pp 631ndash634 2012

[47] V J Hofman M I Ilie C Bonnetaud et al ldquoCytopathologicdetection of circulating tumor cells using the isolation by sizeof epithelial tumor cell method promises and pitfallsrdquo TheAmerican Journal of Clinical Pathology vol 135 no 1 pp 146ndash156 2011

[48] C L Sommers E W Thompson J A Torri R Kemler E PGelmann and SW Byers ldquoCell adhesionmolecule uvomorulinexpression in human breast cancer cell lines relationship tomorphology and invasive capacitiesrdquo Cell Growth amp Differen-tiation vol 2 no 8 pp 365ndash372 1991

[49] P A Kenny G Y Lee C A Myers et al ldquoThe morphologiesof breast cancer cell lines in three-dimensional assays correlatewith their profiles of gene expressionrdquoMolecular Oncology vol1 no 1 pp 84ndash96 2007

[50] S Park R R Ang S P Duffy et al ldquoMorphological differencesbetween circulating tumor cells from prostate cancer patientsand cultured prostate cancer cellsrdquo PLoS ONE vol 9 no 1Article ID e85264 2014

[51] D Marrinucci K Bethel R H Bruce et al ldquoCase study ofthe morphologic variation of circulating tumor cellsrdquo HumanPathology vol 38 no 3 pp 514ndash519 2007

[52] S Tripathy and E J Berger ldquoMeasuring viscoelasticity of softsamples using atomic force microscopyrdquo Journal of Biomechan-ical Engineering vol 131 no 9 Article ID 094507 2009

[53] P Roca-Cusachs I Almendros R Sunyer N Gavara R Farreand D Navajas ldquoRheology of passive and adhesion-activatedneutrophils probed by atomic force microscopyrdquo BiophysicalJournal vol 91 no 9 pp 3508ndash3518 2006

[54] C-L ChenDMahalingam POsmulski et al ldquoSingle-cell anal-ysis of circulating tumor cells identifies cumulative expressionpatterns of EMT-related genes in metastatic prostate cancerrdquoProstate vol 73 no 8 pp 813ndash826 2013

[55] A L Weisenhorn M Khorsandi S Kasas V Gotzos and H-JButt ldquoDeformation and height anomaly of soft surfaces studiedwith anAFMrdquoNanotechnology vol 4 no 2 article 106 pp 106ndash113 1993

[56] M Mak and D Erickson ldquoA serial micropipette microfluidicdevice with applications to cancer cell repeated deformationstudiesrdquo Integrative Biology (Camb) vol 5 no 11 pp 1374ndash13842013

[57] A Mohammadalipou Y E Choi F Benencia M Burdickand D F J Tees ldquoInvestigation of mechanical properties ofbreast cancer cells usingmicropipette aspiration techniquerdquoTheJournal of the Federation of American Societies for ExperimentalBiology vol 26 pp 905ndash909 2012

[58] S Byun S Son D Amodei et al ldquoCharacterizing deformabilityand surface friction of cancer cellsrdquo Proceedings of the NationalAcademy of Sciences of the United States of America vol 110 no19 pp 7580ndash7585 2013

[59] H W Hou Q S Li G Y H Lee A P Kumar C N Ongand C T Lim ldquoDeformability study of breast cancer cells usingmicrofluidicsrdquo Biomedical Microdevices vol 11 no 3 pp 557ndash564 2009

[60] C Jin S M McFaul S P Duffy et al ldquoTechnologies forlabel-free separation of circulating tumor cells from historical

foundations to recent developmentsrdquo Lab on a Chip vol 14 no1 pp 32ndash44 2014

[61] T G Kuznetsova M N Starodubtseva N I Yegorenkov S AChizhik and R I Zhdanov ldquoAtomic force microscopy probingof cell elasticityrdquoMicron vol 38 no 8 pp 824ndash833 2007

[62] S E Cross Y-S Jin J Rao and J K Gimzewski ldquoNanomechan-ical analysis of cells from cancer patientsrdquo Nature Nanotechnol-ogy vol 2 no 12 pp 780ndash783 2007

[63] A Salmanzadeh M B Sano R C Gallo-Villanueva P CRoberts E M Schmelz and R V Davalos ldquoInvestigatingdielectric properties of different stages of syngeneic murineovarian cancer cellsrdquo Biomicrofluidics vol 7 Article ID 0118092013

[64] R O Becker ldquoThe bioelectric factors in amphibian-limbregenerationrdquo The Journal of Bone amp Joint SurgerymdashAmericanVolume vol 43 pp 643ndash656 1961

[65] R O Becker ldquoThe electrical control of growth processesrdquoMedical Times vol 95 no 6 pp 657ndash669 1967

[66] R O Becker and D G Murray ldquoThe electrical controlsystem regulating fracture healing in amphibiansrdquo ClinicalOrthopaedics and Related Research vol 73 pp 169ndash198 1970

[67] R O Becker ldquoThe basic biological data transmission andcontrol system influenced by electrical forcesrdquo Annals of theNew York Academy of Sciences vol 238 pp 236ndash241 1974

[68] R O Becker and G Selden The Body Electric W Morrow ampCompany New York NY USA 1985

[69] J C Cure ldquoElectrical characteristics of cancerrdquo in Proceedingsof the 2nd International Congress of Electrochemical Treatmentof Cancer Jupiter Fla USA 1991

[70] G Qiao W Duan C Chatwin A Sinclair and W WangldquoElectrical properties of breast cancer cells from impedancemeasurement of cell suspensionsrdquo Journal of Physics ConferenceSeries vol 224 no 1 Article ID 012081 2010

[71] F F Becker X-B Wang Y Huang R Pethig J Vykoukal andP R C Gascoyne ldquoSeparation of human breast cancer cellsfrom blood by differential dielectric affinityrdquo Proceedings of theNational Academy of Sciences of the United States of Americavol 92 no 3 pp 860ndash864 1995

[72] GE Healthcare Ficoll Paque PLUS for In Vitro Isolation ofLymphocytes GE Healthcare 2007

[73] D W Fawcett B L Vallee and M H Soule ldquoA method forconcentration and segregation of malignant cells from bloodypleural and peritoneal fluidsrdquo Science vol 111 no 2872 pp 34ndash36 1950

[74] S H Seal ldquoSilicone flotation a simple quantitative methodfor the isolation of free-floating cancer cells from the bloodrdquoCancer vol 12 no 3 pp 590ndash595 1959

[75] N Ahmadbeigi M Soleimani F Babaeijandaghi et al ldquoTheaggregate nature of humanmesenchymal stromal cells in nativebone marrowrdquo Cytotherapy vol 14 no 8 pp 917ndash924 2012

[76] R T van Beem A Hirsch I M Lommerse et al ldquoRecoveryand functional activity of mononuclear bone marrow andperipheral blood cells after different cell isolation protocolsused in clinical trials for cell therapy after acute myocardialinfarctionrdquo EuroIntervention vol 4 no 1 pp 133ndash138 2008

[77] M Aktas T F Radke B E Strauer P Wernet and G KoglerldquoSeparation of adult bone marrow mononuclear cells using theautomated closed separation system Sepaxrdquo Cytotherapy vol10 no 2 pp 203ndash211 2008

[78] C Posel K Moller W Frohlich I Schulz J Boltze and D-CWagner ldquoDensity gradient centrifugation compromises bone


marrow mononuclear cell yieldrdquo PLoS ONE vol 7 no 12Article ID e50293 2012

[79] W M Ellis G M Georgiou D M Roberton and G RJohnson ldquoTheuse of discontinuous percoll gradients to separatepopulations of cells from human bone marrow and peripheralbloodrdquo Journal of Immunological Methods vol 66 no 1 pp 9ndash16 1984

[80] Y Chang P-H Hsieh and C C-K Chao ldquoThe efficiency ofPercoll and Ficoll density gradientmedia in the isolation ofmar-row derived human mesenchymal stem cells with osteogenicpotentialrdquo Chang Gung Medical Journal vol 32 no 3 pp 264ndash275 2009

[81] R Gertler R Rosenberg K Fuehrer M Dahm H Nekardaand J Siewert ldquoDetection of circulating tumor cells in bloodusing an optimized density gradient centrifugationrdquo inMolecu-lar Staging of Cancer H Allgayer M Heiss and F SchildbergEds vol 162 of Recent Results in Cancer Research pp 149ndash155Springer Berlin Germany 2003

[82] L ZabagloM G OrmerodM Parton A Ring I E Smith andM Dowsett ldquoCell filtration-laser scanning cytometry for thecharacterisation of circulating breast cancer cellsrdquo CytometryPart A vol 55 no 2 pp 102ndash108 2003

[83] J-S Shiau C-H Tang T-Y Lin and D-M Wang ldquoA modelfor resistance growthduring proteinmicrofiltrationrdquo SeparationScience and Technology vol 38 no 4 pp 917ndash932 2003

[84] H M Ji V Samper Y Chen C K Heng T M Lim and LYobas ldquoSilicon-based microfilters for whole blood cell separa-tionrdquo Biomedical Microdevices vol 10 no 2 pp 251ndash257 2008

[85] T A Crowley and V Pizziconi ldquoIsolation of plasma fromwholeblood using planar microfilters for lab-on-a-chip applicationsrdquoLab on a Chip vol 5 no 9 pp 922ndash929 2005

[86] C-C Wu ldquoBlood cell-free plasma separated from blood sam-ples with a cascading Weir-type microfilter using dead-endfiltrationrdquo Journal of Medical and Biological Engineering vol 32no 3 pp 163ndash168 2012

[87] W Chen N-T Huang B Oh et al ldquoSurface-micromachinedmicrofiltrationmembranes for efficient isolation and functionalimmunophenotyping of subpopulations of immune cellsrdquoAdvanced Healthcare Materials vol 2 no 7 pp 965ndash975 2013

[88] S J Tan R L Lakshmi P Chen W-T Lim L Yobas andC T Lim ldquoVersatile label free biochip for the detectionof circulating tumor cells from peripheral blood in cancerpatientsrdquo Biosensors and Bioelectronics vol 26 no 4 pp 1701ndash1705 2010

[89] W Sheng T Chen R Kamath X XiongW Tan and Z H FanldquoAptamer-enabled efficient isolation of cancer cells from wholebloodusing amicrofluidic devicerdquoAnalytical Chemistry vol 84no 9 pp 4199ndash4206 2012

[90] W Ma D Liu H Shagoshtasbi et al ldquoExperimental andtheoretical study of hydrodynamic cell lysing of cancer cellsin a high-throughput Circular Multi-Channel Microfiltrationdevicerdquo in Proceedings of the 8th IEEE International Conferenceon NanoMicro Engineered and Molecular Systems (NEMS rsquo13)pp 412ndash415 IEEE Suzhou China April 2013

[91] F Petersson A Nilsson C Holm H Jonsson and T LaurellldquoContinuous separation of lipid particles from erythrocytes bymeans of laminar flow and acoustic standing wave forcesrdquo Labon a Chip vol 5 no 1 pp 20ndash22 2005

[92] L M Broche N Bhadal M P Lewis S Porter M P Hughesand F H Labeed ldquoEarly detection of oral cancermdashis dielec-trophoresis the answerrdquo Oral Oncology vol 43 no 2 pp 199ndash203 2007

[93] H JMulhall FH Labeed B KazmiD E CosteaM PHughesand M P Lewis ldquoCancer pre-cancer and normal oral cells dis-tinguished by dielectrophoresisrdquo Analytical and BioanalyticalChemistry vol 401 no 8 pp 2455ndash2463 2011

[94] F Fabbri S Carloni W Zoli et al ldquoDetection and recovery ofcirculating colon cancer cells using a dielectrophoresis-baseddevice KRAS mutation status in pure CTCsrdquo Cancer Lettersvol 335 no 1 pp 225ndash231 2013

[95] C Huang H Liu N H Bander and B J Kirby ldquoEnrich-ment of prostate cancer cells from blood cells with a hybriddielectrophoresis and immunocapture microfluidic systemrdquoBiomedical Microdevices vol 15 no 6 pp 941ndash948 2013

[96] P R C Gascoyne X-B Wang Y Huang and R F BeckerldquoDielectrophoretic separation of cancer cells from bloodrdquo IEEETransactions on Industry Applications vol 33 no 3 pp 670ndash678 1997

[97] AMenachery andR Pethig ldquoControlling cell destruction usingdielectrophoretic forcesrdquo IEE Proceedings Nanobiotechnologyvol 152 no 4 pp 145ndash149 2005

[98] D M Vykoukal P R C Gascoyne and J Vykoukal ldquoDielectriccharacterization of completemononuclear and polymorphonu-clear blood cell subpopulations for label-free discriminationrdquoIntegrative Biology vol 1 no 7 pp 477ndash484 2009

[99] Y Huang J Yang X-B Wang F F Becker and P R C Gas-coyne ldquoThe removal of human breast cancer cells from hema-topoietic CD34+ stem cells by dielectrophoretic field-flow-fractionationrdquo Journal of Hematotherapy amp Stem Cell Researchvol 8 no 5 pp 481ndash490 1999

[100] A Menachery C Kremer P E Wong et al ldquoCounterflowdielectrophoresis for trypanosome enrichment and detection inbloodrdquo Scientific Reports vol 2 article 775 2012

[101] X Xing M Zhang and L Yobas ldquoInterdigitated 3-D siliconring microelectrodes for DEP-based particle manipulationrdquoJournal of Microelectromechanical Systems vol 22 no 2 ArticleID 6341027 pp 363ndash371 2013

[102] L Wang J Lu S A Marchenko E S Monuki L A Flana-gan and A P Lee ldquoDual frequency dielectrophoresis withinterdigitated sidewall electrodes for microfluidic flow-throughseparation of beads and cellsrdquo Electrophoresis vol 30 no 5 pp782ndash791 2009

[103] T-S Leu and Z-F Liao ldquoSeparating plasma and blood cells bydielectrophoresis inmicrofluidic chipsrdquo International Journal ofModern Physics vol 19 pp 185ndash189 2012

[104] B I Morshed M Shams and T Mussivand ldquoElectrical lysisdynamics revisited and advances in on-chip operationrdquo CriticalReviews in Biomedical Engineering vol 41 no 1 pp 37ndash50 2013

[105] J A Davis D W Inglis K J Morton et al ldquoDeterministichydrodynamics taking blood apartrdquo Proceedings of the NationalAcademy of Sciences of the United States of America vol 103 no40 pp 14779ndash14784 2006

[106] A A S Bhagat H W Hou L D Li C T Lim and J HanldquoPinched flow coupled shear-modulated inertial microfluidicsfor high-throughput rare blood cell separationrdquo Lab on a Chipvol 11 no 11 pp 1870ndash1878 2011

[107] L Ying and Q Wang ldquoMicrofluidic chip-based technologiesemerging platforms for cancer diagnosisrdquo BMC Biotechnologyvol 13 article 76 2013

[108] G Guan L Wu A A Bhagat et al ldquoSpiral microchannelwith rectangular and trapezoidal cross-sections for size basedparticle separationrdquo Scientific Reports vol 3 article 1475 2013


[109] M G Lee J H Shin C Y Bae S Choi and J-K Park ldquoLabel-free cancer cell separation from human whole blood usinginertial microfluidics at low shear stressrdquo Analytical Chemistryvol 85 no 13 pp 6213ndash6218 2013

[110] D L Holliday and V Speirs ldquoChoosing the right cell line forbreast cancer researchrdquo Breast Cancer Research vol 13 no 4article 215 2011

[111] K Okita and S Yamanaka ldquoInduced pluripotent stem cellsopportunities and challengesrdquo Philosophical Transactions of theRoyal Society B Biological Sciences vol 366 no 1575 pp 2198ndash2207 2011

[112] J Holmberg and T Perlmann ldquoMaintaining differentiatedcellular identityrdquoNature ReviewsGenetics vol 13 no 6 pp 429ndash439 2012

[113] J T Auman and H L McLeod ldquoColorectal cancer cell lineslack the molecular heterogeneity of clinical colorectal tumorsrdquoClinical Colorectal Cancer vol 9 no 1 pp 40ndash47 2010

[114] M BWolf and E C DeLand ldquoAmathematical model of blood-interstitial acid-base balance application to dilution acidosisand acid-base statusrdquo Journal of Applied Physiology vol 110 no4 pp 988ndash1002 2011

[115] A Tan ldquoApoCell receives $1 million Small Business InnovationResearch (SBIR) contract to develop a point-of-care clinicaldevice for Circulating Tumor Cell (CTC) Isolationrdquo BioNewsTexas 2013

[116] S Hanner ldquoBiomarker-driven trial of novel topo-1 inhibitordebutedrdquo Targeted Oncology In press

[117] I Desitter B S Guerrouahen N Benali-Furet et al ldquoA newdevice for rapid isolation by size and characterization of rarecirculating tumor cellsrdquo Anticancer Research vol 31 no 2 pp427ndash441 2011

Submit your manuscripts athttpwwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Anatomy Research International

PeptidesInternational Journal of


Hindawi Publishing Corporation httpwwwhindawicom

International Journal of

Volume 2014

Zoology


Molecular Biology International

GenomicsInternational Journal of


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


BioinformaticsAdvances in

Marine BiologyJournal of



Signal TransductionJournal of


BioMed Research International

Evolutionary BiologyInternational Journal of



Biochemistry Research International

ArchaeaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Genetics Research International


Advances in

Virolog y

Hindawi Publishing Corporationhttpwwwhindawicom

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational



Enzyme Research



Microbiology


free survival (PFS) and overall survival rates were twice ashigh for patient with less than 3 CTCs per 75mL of bloodthus it has confirmed the previous findings Additionallythis group also presented significant data which showed thatpatients with elevated CTCs density after therapy wouldhave poor survival rate [8] Nevertheless simple enumerationof CTCs is inadequate because cancer is a constellation ofdiseases with various pathologic alterations that might causeprognosis Since the ability in analyzing proliferation of viableCTCs has still been lacking it is difficult to assess CTC infor-mation which is the representative of cellular informationavailable in primary tumor In fact the dimension of CTCbiological feature is especially significant for basic researchpertaining to metastasis as well as drug development Tofurther complicate matters the recent appreciation of geneticalterations and biomarker expression for instance KRASwithin tumors means a single biopsy sample is no longersufficient [9] Henceforth detecting and analyzing these cellson a sample of blood may shed new light on circumventsclinical need to improve therapeutic efficacy as well as theoverall patient survival rate Despite its high potential incancer treatments the detection of CTCs from whole bloodsample is particularly challenging due to their extremely rarepresence which is only 1 to 10 CTCs per billion normalblood cells in patients with advanced cancer Besides thelarge morphological variability among CTCs has imposedtechnical challenge to isolate the whole population wheresophisticated algorithms are required to identify the CTCsAforesaid CTC counts are associated with patient prognosisThus a highly sensitive detectionmethod is vital to accuratelycharacterize and enumerate the CTC

Numerous methods have been employed and discussedin the literature for CTCs separation To date the benchtopdevice developed by Veridex is the sole system approvedby United States Food and Drug Administration for theclinical monitoring of CTC counts in patients withmetastaticbreast colon and prostate cancerThrough this approach theCTCs are distinguished from other blood cells based on theirimmunoaffinity properties The selection is carried out byusing antibodies coupled with magnetic beads targeting thetumor-associated antigen onCTCs surface (EpCAMCK andCD45)This system is proved to have sensitivity of 877withthe capability to detect of approximately 5 CTCs per 75mL ofwhole blood Despite the proven clinical ability of this systema few data have been published concerning the inconsistentlymarkers expressed by all tumor cells Noteworthy EpCAM isnot expressed in all histological tumor types and thus mightresult in false-negativepositive result and limit the purity ofthe enriched CTC fraction Using affinity capture methodin macroscale analytical system will also lead to permanentattachment of target cells to marker protein on CTC Con-sequently the downstream options for the extraction andsubsequent characterization of CTC will be narrowed down

In addition to immune-affinity based separation in fact alarge panel of approaches for CTC isolation that are indepen-dent of cell surface antigens have been developed on the basisof its physical properties The majority of the physical meth-ods for CTCs isolation exploit the differences in mechanicaland physical properties including cell size deformability

electrical polarizability and magnetic susceptibility In con-trast to the immune-affinitymethodwhere epithelial antigensare needed to mediate the intercellular adhesion this tech-nique is label-free Therefore the interference such as samplecontamination due to the tagging molecules can be avoidedSince CTCs are unmodified by physical separation cells iso-lated using these methods are compatible with wider range ofanalyses including those requiring viable cells Consequentlyit allows large range of molecular biological analysis of CTCsto be implemented and thus provides clinician with furtherinsight into CTCs role in tumor formation

In this paper we will present an overview about the pot-ential application of benchtopCTCdetection device based ontheir physical properties in clinical oncology Both detaileddescription of various separation principle and comparisonsbased on separation metrics such as efficiency viability andthroughput are included in this review Additionally thechallenges faced by each technique will too be succinctly dis-cussedThe readers are encouraged to read the original papersfor additional details

2 Comparison of Biophysical andBiomechanical Properties betweenCTCs and Normal Blood Cells

According to most CTCs histological study researchers haveobserved changes in cell cytoskeleton contents and theircytoskeletal structures as it progresses toward a cancerousstate For instance CTCs are found to have greater nuclear tocytoplasmic ratio larger size and distinct nuclear morphol-ogy in contrast to the normal cells In fact these cytoskeletalchanges have resulted in changes in the overall mechanicalproperties of cells Computational mathematical model con-ducted by Rejniak has demonstrated the insight of CTCrsquosdeformation trajectories within blood vessel [36] His find-ing highlighted that CTCsrsquo cellular biomechanics such asdeformability played an important role in metastasis As anexample in order to invade distal sites the stiffness of CTCscytoskeleton is modified subsequently in a very dynamic wayso that it could squeeze across small spaces in extracellularmatrix and endothelial cell-cell junctions as well as circulat-ing through the small capillaries In order to successfully per-form the metastatic cascade the CTCs cell membrane mustwithstand hemodynamic forces and overcome the effects offluid shearswithin the blood vessel Consequently it alters theconservation of membrane structure which in turn affectsthe surface charge (electrical charge) in contrast to normalblood cells [37] These distinct physical differences in sizedeformability and electrical properties between tumor cellsand blood cells have allowed researchers to employ themas physical-based CTC detection method The physical andbiomechanical properties of CTCs will be briefly discussedin the following section

21 Size Inmorphological studies themeasurementmethodsuch as flow cytometry [38] blood smear or detailed exami-nation with microscopic study are broadly employed to ana-lyze the cell size of CTCs in comparison to normal cells



















Blood sample

Ficoll

Plasma CTCs

WBC

Ficoll

RBC neutrophils

Centrifuge

















(a)

Gap distance

350120583m

25120583m 10120583m 10120583m

0bottom 8120583m

0top 40120583m


Red blood cells

Leukocytes

Parylene C-membrane

(b)

(c)

Slot-shaped filter

(d)

Pump Drain

(e)

Large cells

Small cells



Inlet 1(samples)

Inlet 2(buffer)A

Physicalbarrier

A


GapA998400

A998400

(f)

Figure 2 Continued


Microchannel


Bead-packed inlet

1mm

(g)







Flow

Separated particles

Sample inlet


Clean fluid inlet

Sample inlet

(a)

Flow

(b)


















Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)


using pipette

ITO



Inertial

14

8

Voltage applied

(b)






Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



















Blood sample

Ficoll

Plasma CTCs

WBC

Ficoll

RBC neutrophils

Centrifuge

















(a)

Gap distance

350120583m

25120583m 10120583m 10120583m

0bottom 8120583m

0top 40120583m


Red blood cells

Leukocytes

Parylene C-membrane

(b)

(c)

Slot-shaped filter

(d)

Pump Drain

(e)

Large cells

Small cells



Inlet 1(samples)

Inlet 2(buffer)A

Physicalbarrier

A


GapA998400

A998400

(f)

Figure 2 Continued


Microchannel


Bead-packed inlet

1mm

(g)







Flow

Separated particles

Sample inlet


Clean fluid inlet

Sample inlet

(a)

Flow

(b)


















Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)


using pipette

ITO



Inertial

14

8

Voltage applied

(b)






Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology










(a)

Gap distance

350120583m

25120583m 10120583m 10120583m

0bottom 8120583m

0top 40120583m


Red blood cells

Leukocytes

Parylene C-membrane

(b)

(c)

Slot-shaped filter

(d)

Pump Drain

(e)

Large cells

Small cells



Inlet 1(samples)

Inlet 2(buffer)A

Physicalbarrier

A


GapA998400

A998400

(f)

Figure 2 Continued


Microchannel


Bead-packed inlet

1mm

(g)







Flow

Separated particles

Sample inlet


Clean fluid inlet

Sample inlet

(a)

Flow

(b)


















Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)


using pipette

ITO



Inertial

14

8

Voltage applied

(b)






Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Microchannel


Bead-packed inlet

1mm

(g)







Flow

Separated particles

Sample inlet


Clean fluid inlet

Sample inlet

(a)

Flow

(b)


















Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)


using pipette

ITO



Inertial

14

8

Voltage applied

(b)






Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Flow

Separated particles

Sample inlet


Clean fluid inlet

Sample inlet

(a)

Flow

(b)


















Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)


using pipette

ITO



Inertial

14

8

Voltage applied

(b)






Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology










Outlet A

PDMS

AgPDMSelectrodes

n-DEP p-DEP

(a)


using pipette

ITO



Inertial

14

8

Voltage applied

(b)






Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Branch 1

Pinched segment

Inlet 1

Inlet 2Branch 5

Branch 2

Branch 3

Branch 4

Inlet 1

Inlet 2

Branch A1A2 A3 A4

A5

A6

A7

A8

A9A10A11A12

(B)

(D)



(i) (ii)

(iii)

106mm3mm

15mm50120583m

20120583m

100120583m

200120583m

200120583m

(a)

RBC

Flow

CTC

WBC

(b)

Sample

Inlets

Y

Y

X

X

ZZ

Sheath

Isolated CTCs

Outlets

Bloodcells

CTCsLeukocytes

RBCs




1de

an cy

cle12

dean

cycle

FL

FD

FDFD

FD

FD

FL

FD

FD

FD

(c)

Figure 5 Continued




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology




Focusing flow

Cancer ceWhit

Reicle flow

s1

s2


Focusing flow

Particle flow

Y

X

107mm 500120583m


(d)



1(reprinted






















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

















ble1Perfo

rmance

specificatio

nof

physicalbasedCT

Csseparatio

nmetho

ds

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences

Centrifu

gatio

n(Ficoll-H

ypaque)

Hum

anbreastcancer

cells

(MCF

-7)

sim46

mdash

90

mdashLara

etal[26]

Centrifu

gatio

n(O

ncoQ

uick)

MCF

-7Z

R-75-1and

Hs578T

MCF

-7578

9ZR

-75-15980

Hs578T

5714

mdash

mdashmdash

Konigsberg

etal[27]


Hum

anbreastcancer

cells

(MCF

-7M

DA-

MB-453and

SK-BR-3)

andhu

man

prostate

adenocarcino

mac

ancerc

ells

(LNCA

PPC

3)

7120583m

pore

filter

(i)98plusmn2

(fixedMCF

-7cells)

(ii)8

5plusmn3

(unfi

xedMCF

-7cells)

8120583m

pore

filter


-7cells)

(ii)5

0plusmn14(unfi

xedMCF

-7cells)

mdashmdash

sim10mLmin

(con

centratio

n10

4

cellsm

L)Ad

amse

tal[10]

Hum

anprostatecancer

PC-3

cellsgt90

gt90

mdashmdash

Luetal[28]


brane)

Hum


ma

celllin

e(LN

CaP)

andhu

man

breastadenocarcino

mac

ellline

(MCF

-7)

gt86

85

mdashsim375

mLmin

(con

centratio

n10

6

cellsm

L)Zh

engetal[11]

Hum

angastr

iccarcinom

acell

line(NCI

N-87)

gt85

mdashmdash

25m

Lmin

(con

centratio

n25times

105cellsm

L)Yu

saetal[12]

Breastcancer

celllin

esM

CF-7

MDA-

MB-231

78ndash83

71ndash74

mdashmdash

(1120583cancerc

ellsspiked

into

1mL

ofun

dilutedbloo

d)Zh

ouetal[13]

Microfiltration(w

eir-type)

Hum

anA431cancerc

ells

gt95

mdashgt98

sim5m

Lhr

(10tumorsc

ellsmLof

who

lebloo

d)Ch

ungetal[15]


type)

MCF

-7SK-

BR-3J82T

24R

T4

andLN

CaP

gt90

ge93

gt90

mdashLinetal[29]


bead

type)

Hum

anbreastcancer

cells

(MCF

-7)

21ndash4

0mdash

02m

Lhr

(con

centratio

n10

6 mL)

Aryae

tal[30]

Dielectroph

oresis(planar

confi

guratio

n)Hum

anbreastcancer

cells

(MCF

-7)

(i)MCF

-77518

(ii)R

BC992

4(iii)WBC

9423

mdash(i)

MCF

-71624

(ii)R

BCmdash

(iii)WBC

mdashsim250120583

Ls

Moo

netal[31]

Dielectroph

oresis(con

tactless

mod

e)Hum

ancervicalcarcinom

acell

(HeLac

ells)

645

mdashmdash

mdash(con

centratio

nin

diluted

solutio

n25times10

5HeLac

ellsmL

325times10

6RB

Ccellsm

L325times

103WBC

cellsm

L)

Huang

etal[19]

Acou

stoph

oresis

Prostatecancer

celllin

e

DU145PC

3andLN

CAP

PFA-

fixed

cellsample

(i)DU145andPC

3854at119864ac=120Jm

3 966at

119864ac=188Jm

3

(ii)L

NCa

P555at119864ac=120Jm

3 788at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

361

at119864ac=66Jm

3 837at

119864ac=103Jm

3

Nosig

nificantchang

eswere

observed

fora

coustoph

oresis

treated

andun

treated

samples

PFA-

fixed

cellsample

(i)DU145andPC

3993at119864ac=120Jm

3 979at

119864ac=188Jm

3

Non

-PFA

-fixedcellsample

(i)DU145

995at119864ac=66Jm

3 93at

119864ac=103Jm

3

sim450120583

Lmin

(25times10

5tumor

cellsm

Lspiked

with

10-fo

lddilutedwho

lebloo

d)Au

gusts

sonetal[32]

DU145PC

3LN

CAPand

VCaP

mdashTh

eaverage

celldead

islt1

for

DU145andPC

3whilele3

for

LNCA

PandVC

aPmdash

sim100120583

Lmin

Burguillo

setal[33]


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology


Table1Con

tinued

Metho

dTargetof

cancer

celllin

eCT

Csseparatio

neffi

ciency

Cellviability

Purity

Flow

rate(w

holebloo

dtim

e)Re

ferences


flowfractio

natio

n)MV3-melanom

acells

100

sim100

(i)04

hematocrit666

(ii)4

hematocrit154

(iii)9

hematocrit55

600120583

Lh(con

centratio

nin

dilutedsolutio

n8times10

5tumor

cellsm

L6times10

8RB

Ccellsm

L)Geisling

erandFranke

[34]

Hydrodynamicsorting


displacement)

Hum

anbreastcancer

celllin

eMDAMB2

31M

CF10Aand

PC3ge85

ge95

mdash10mLmin

(con

centratio

n10

7cancer

cells

in1m

Lof

bloo

ddilutedwith

buffer)

Loutherbacketal[22]


ertia

llift

force)

MCF

-7gt85

gt98

50

3mLhr

(con

centratio

n10ndash100

CTCs

mLof

dilutedwho

lebloo

dsamples)

Hou

etal[23]


ertia

force)

MCF

-7T

24and

MDA-

MB-231gt80

gt80

sim4logdepletionof

WBC

s20m

Lmin

(500

tumou

rcellsper

75mLof

who

lebloo

d)Warkianietal[35]


ertia

force)

MCF

-7cancer

cells

995

mdashmdash

3to


7

cellsm

into

2times10

7cellsm

in)

Leee

tal[24]



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology



CTCs Waste

Sample injection

CTC collection


Buffer inlet


Flow chamber


drive aspirate

Waste






Acknowledgment


References





































































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

























































































































Volume 2014

Zoology






















Advances in

Virolog y



Volume 2014




Enzyme Research



Microbiology

review article benchtop technologies for circulating tumor

Documents