British Journal of Neurosurgery 2002; 16(2): 110118

ISSN 02688697 print/ISSN 1360046X online/02/02011009 The Neurosurgical FoundationDOI: 10.1080/0268869022013170 5

ORIGINAL ARTICLE

The advantages of frameless stereotactic biopsy over frame-based biopsy

N. L. DORWARD1, T. S. PALEOLOGOS2, O. ALBERTI2 & D. G. T. THOMAS2

Departments of Neurosurgery, 1Royal Free Hospital, and 2National Hospital for Neurology and Neurosurgery, London, UK

AbstractA comparison study is presented, which examines the outcome, complications and cost of stereotactic brain biopsyperformed with a frameless versus a frame-based method. The technique of frameless stereotactic biopsy has beenshown previously, in both laboratory and in vivo studies, to achieve a level of accuracy at least equal to frame-basedbiopsy. The investigators have validated the technique in a large clinical series. The frameless and frame-based serieswere concurrent, comprising 76 and 79 cases, respectively. The frameless stereotactic technique involved standardneedle biopsy, targeted by an image-guidance system and directed by a novel rigid adjustable instrument-holder.Frame-based biopsies were performed with the CRW and Leksell systems. There were no significant differences in thedemographics, lesion site, size and pathologies between the groups. Operating theatre occupancy and anaesthetic timewere both significantly shorter for the frameless series than the frame-based series (p < 0.0001). In addition, thecomplication rate in the frameless biopsy series was significantly lower than in the frame-based series (p = 0.018). Thisresulted in lower ITU bed occupancy (p = 0.02), shorter mean hospital stay (p = 0.0013) and significant cost savings(p = 0.0022) for the frameless stereotactic biopsy group, despite the greater use of more expensive MRI in these cases.This comparison study demonstrates that the superior imaging, target visualization and flexibility of the technique offrameless stereotactic biopsy translates into tangible advantages for safety, time and cost when compared with thecurrent gold-standard of frame-based biopsy. The principles are discussed and the authors propose a definition for theterm frameless stereotaxy.

Key words: Comparison study, Computer-assisted surgery, frameless stereotaxy, neuronavigation, stereotactic biopsy.

Introduction

Frame-based stereotactic biopsy is the current gold-standard technique for the retrieval of histologicalspecimens from targets within the brain. This methodprovides the neurosurgeon with a safe (mortality95%)18 means for biopsy retrieval which has trans-formed the outcome of this procedure compared withfreehand (CT-directed) burr-hole biopsy (mortality>5%, morbidity 15%, diagnostic yield 85%).9,10

However, the frame-based technique is cumbersome,restricted to point targeting and still carries a small,but significant complication rate.

We have developed a method for frameless stere-otactic biopsy that combines the precision ofstereotaxy with the ergonomics and practicality ofneuronavigation. The technique has been describedin detail previously,11 and the results of laboratory

phantom accuracy assessments reported.12 Theaccuracy of the frameless stereotactic techniquewas found to be at least the equivalent of contem-porary frame-based methods (mean error ofphantom frameless stereotactic biopsy = 1.8 mm,frame-based biopsy = 2.1 mm, for equivalentimaging parameters).1214 A preliminary clinicalstudy of this technique revealed how improvedimage presentation and manipulation facilitatedselection of the optimal biopsy target and demon-strated a significant reduction in the duration ofanaesthesia. 15 The assessments of accuracyincluded a study in which the position of the actualbiopsy site was identified with post-operative MRIand compared with the intended target site viaimage fusion. By this means the mean in vivo accu-racy of frameless stereotactic biopsy was shown tobe highly satisfactory at 2.3 mm.12

Correspondence: Mr N. L. Dorward, Department of Neurosurgery, Royal Free Hospital, Pond Street, London NW3 2QG, UK. Tel: 0207830 2097. Fax: 020 7830 2560. E-mail: neil.dorward@rfh.nthames.nhs.uk.

Received for publication 8 January 2001. Accepted 14 January 2002.

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

Advantages of frameless stereotaxy 111

In this paper we present a clinical comparativestudy, contrasting this novel technique with theestablished technique of frame-based biopsy inconcurrent series.

Materials and methods

Patient population

Between September 1996 and April 1999, 155 stere-otactic biopsy procedures were performed. Seventy-nine of these were undertaken with a stereotacticframe and 76 were conducted with the novel tech-nique of frameless stereotactic biopsy. Selection ofthe technique for each case was determined by theavailability of the image-guidance system (and IGSresearch fellow). Thus, throughout this period thepreferred technique was frameless biopsy with frame-based biopsy employed when this was unavailable(i.e. there was no case selection for either procedure).The patients in the frame-based series (Table I)comprised 49 men and 30 women, aged between 15and 96 years (mean 52.1). Those in the framelessseries comprised 42 men and 34 women, agedbetween 25 and 79 years (mean 54.9). The imagingmodality employed to target the biopsy was CT in89% of the frame-based procedures and 32% of theframeless operations. Conversely, the imagingmodality was MRI in 11% of the frame-based casesand 68% of the frameless ones.

Technique of frame-based stereotactic biopsy

The stereotactic frames employed were the CosmanRobertsWells (CRW, Radionics, Burlington, USA)and the Leksell (Elekta Instruments, Stockholm,Sweden) systems. All biopsies were performed undergeneral anaesthetic, induced prior to application ofthe base ring. Patients were transferred to thescanner with full monitoring and intravenous anaes-thesia, returning to the operating room for thesurgical procedure. The CT protocol comprisedacquisition of a block of axial slices of 2 mm slicethickness with 3 mm slice spacing, following intrave-nous injection of contrast. The MRI protocol

comprised acquisition of a full head gadolinium-enhanced T1-weighted volume of 80 slices giving aslice thickness of 1.5 mm.

Technique of frameless stereotactic biopsy

The neuronavigation system employed was Easy-Guide NeuroTM (Philips Medical SystemsNederland BV, Best, The Netherlands), an infra-redLED-based system. A stereotactic guide wasdeveloped,11 which would adapt to a variety ofinstruments, was freely adjustable to reach all partsof the cranium, would lock in place rigidly andallowed fine correction of the trajectory setting(Fig. 1). The guide arm was fixed to the Mayfieldclamp (OMI Surgical Products, Cincinnati, USA)and the joints of the arm locked simultaneouslywithout producing torsional movement.

The method developed for frameless stereotaxycomprised six stages, namely (i) image acquisition,(ii) image to patient registration, (iii) entry pointselection, (iv) burr hole construction, (v) target andtrajectory definition, and (vi) biopsy retrieval. Thistechnique has been described in detail previously.11,16

In summary, self-adhesive scalp fiducials wereapplied prior to imaging with either Gadolinium-enhanced MRI or contrast-enhanced CT withstandard IGS protocols (MR; SPGR sequence withminimum TR and TE, flip 50, matrix 256 3 256,FOV 24 cm and slice thickness 1.5 mm, CT; helicalscan with matrix of 512 3 512, FOV 20.5 cm, gantrytilt 0 and slice thickness 3 mm). The patient waspositioned and a Mayfield head clamp applied. Imageto patient registration was performed with a fiducial-based method. The entry point was selected using ahand-held pointer and virtual elongation to displaythe trajectory and potential target sites (Fig. 2).Distance to target was calculated, depth stop

TABLE I. Patient demographics and imaging modalities forthe frameless and frame-based biopsy series

Frame-based series

Frameless series

Combined series

Number n 79 76 155Gender Male 49 42 91

Female 30 34 64Age (years) Range 1596 2579 1596

Mean 52.8 54.9 53.8SD 15.9 14.7 15.3

Scan Modality CT 89% 32% 59%MRI 11% 68% 41%

FIG. 1. Target and trajectory selection with the framelessstereotactic technique. The instrument holder is attachedto the Mayfield head clamp (foreground) and the arm islocked in position over the burr-hole. An IGS pointer isseen docked in the trapped ball, which allows fineadjustment of the trajectory.

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

112 N. L. Dorward et al.

positioned on the biopsy needle (SedanNasholdside cutting needle, Radionics Inc., Burlington,USA) and specimens retrieved from each quadrantas performed normally. (Fig. 3).

Cost analysis

The cost of inpatient days, surgery for stereotacticbiopsy and imaging studies were obtained from thehospital finance department (Table II). The numberof days spent by each patient on the ward and inITU was determined from inpatient and ward

records. Additional imaging was defined as thatwhich was obtained during the immediate postoper-ative period through clinical necessity (i.e. mainlyfollowing the occurrence of a complication). Theminor differences in the protocols for MRI and CTacquisition for the two biopsy techniques did nothave any financial impact, as scanning times weresimilar. However, the fiducial markers each cost 1leading to an average additional cost of 10 perframeless stereotaxy case.

Capital costs per case were derived for each tech-nique by determination of the purchase costs,annual service charges, equipment lifetime andnumber of cases per annum. The CRW lifetimewas known to be 10 years, as this was the intervalfrom first acquisition to replacement in our Institu-tion. The image-guidance lifetime was presumed tobe similar. The proportion of the capital costsborne by the frameless biopsy cases was calculatedfrom the ratio of biopsies per annum (frameless andframe-based) to open image-guided cases perannum (50%). The proportion of the capital costsborne by the frame-based biopsy cases was 100% ofthe frame costs.

Image and data analysis

Much of the patient and clinical data were collectedprospectively and supplemented by review of thecase notes for each patient, examination of thepreoperative brain scans and pathology fromthe histology database. The maximum dimension ofeach lesion was measured, the depth of the lesionbelow the nearest skin surface ascertained (methoddescribed in detail previously12) and the lesionclassified according to the lobe(s) involved, side,eloquence and brain structure(s) affected. The dura-tion of the anaesthetic was obtained from theanaesthetic records and the duration of surgery fromthe operating room computer log. Intra-operativecomplications were discerned from the surgeon oroperation notes and postoperative complicationsclinically, from the ward records and pathologyservice records.

The results were analysed for patterns, correlationand significance as a whole, and when segregatedaccording to operation type. The statistical tests of

FIG. 2. MR images from a case in the framelessstereotactic biopsy series. The pointer tip (at the skinsurface) is extended with virtual pointer elongation todisplay the target and trajectory for biopsy. Sagittal,coronal and axial reformats reveal the detailed anatomy ofthe target site, whilst the perpendicular pointer view (lowerleft image) allows the surgeon to walk along the trajectoryto the target and beyond.

FIG. 3. Biopsy retrieval with the frameless stereotaxytechnique. The adjustable, adaptable instrument-holderlocked in position rigidly allowing reliable and repeatabletarget localization.

TABLE II. Hospital costings (GB) for in-patient stay,surgery and imaging

Cost ()

Admission (Neurosurgical Ward 24 h) 380Admission (ITU 24 h) 1350Operation (frame-based stereotactic biopsy) 4650Operation (frameless stereotactic biopsy) 4650CT scan 150MRI scan 500CT scan (+ fiducials) 160MRI scan (+ fiducials) 510

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

Advantages of frameless stereotaxy 113

significance employed were unpaired two-tailed t-tests for normally distributed data and c 2 tests fornon-continuous data (significance established whenp < 0.05).

Results

Radiological characteristics of the lesions

Examination of the imaging studies revealed that thedistribution, dimensions and locations of the lesionswere similar for the frame-based and frameless series(Table III). The mean lesion diameter was 37 mm(range 880 mm, SD 15.8), the lesion was in thecerebral hemispheres in 87%, the diencephalon in

9% and the brainstem/cerebellum in 4%. The lesionwas on the right in 45% of cases, on the left in 42%,midline in 10% and bilateral in the remaining 3%.In 88% of cases the lesion was single and in 12%there was more than one lesion. The lesions werecortical in 8%, subcortical in 32% and deep in 60%.This distribution was seen in both the frame-basedand frameless series cases. The depth of the lesionsbelow the nearest skin surface was found to besignificantly greater in the frame-based series (mean47.2 mm, SD 20.6 mm) than in the frameless seriesof cases (mean 35.2, SD 18.6, p < 0.0001). Therewas, however, no difference between the groups inthe frequency of lesions within eloquent regions.

Pathological diagnosis

The spectrum of disease encountered in each of thebiopsy series was representative of a general, unse-lected group of neurosurgical patients. There wereno significant differences between the proportions ofpathologies between the frame-based and framelessseries. The final pathological diagnosis was classifiedinto high grade glioma, low grade glioma, othertumour, inflammatory and non-diagnostic groups.In the frame-based series there were 47 high gradegliomas (60%), 13 low grade gliomas (16%), 10other tumours (13%), five inflammatory conditions(6%) and four cases (5%) without a pathologicaldiagnosis (Table IV). In the frameless series therewere 52 high grade gliomas (68%), six low gradegliomas (8%), 12 other tumours (16%), five inflam-matory conditions (7%) and one case without apathological diagnosis (1%).

Comparison of the smear specimen diagnosiswith the final pathological diagnosis for each caseand for each series revealed similar results for thetwo series. The correct diagnosis was reached onsmear examination in 130 (87%) of the 150 diag-nostic specimens. These rates were similar for bothframe-based and frameless series, and compare wellwith the rates quoted in the literature.1719

TABLE III. Imaging characteristics for the lesions in eachseries

Frame-based series

Frameless series

Combined series

LateralityRight 41% 47% 45%Left 41% 42% 42%Midline 14% 8% 10%Bilateral 4% 3% 3%

LocationTelencephalon 84% 90% 87%Diencephalon 10% 8% 9%Rhombencephalon 6% 2% 4%Multifocal 34% 14% 24%

SiteCortical 4% 11% 8%Subcortical 27% 36% 32%Deep 69% 53% 60%

Number of lesionsSingle 88% 87% 88%Multiple 12% 13% 12%

Lesion diameter (mm)Range 880 879 880Mean 34.8 38.7 37SD 16.2 15.4 15.8

Lesion depth (mm)Range 11102 7102 7102Mean 47.2 35.2 40.5SD 20.6 18.6 20.3

TABLE IV. Histological results for the frame-based, frameless and combined series with pathological classification

Classification Pathology Frame-based series Frameless series Combined series

High grade glioma Glioma grade IIIIV 47 52 99Low grade glioma Glioma grade III 13 6 19

Lymphoma metastasis 4 5 93 5 8

Other tumours Pineal tumour 2 1 3Neuroma craniopharyngioma 1 1

1 1Abscess 1 1 2

Inflammatory conditions Multiple sclerosis 1 1 2PML/HIV 1 1 2Cysticercosis 1 1Radionecrosis 1 1Vasculitis 1 1 2

Non-diagnostic No result 4 1 5

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

114 N. L. Dorward et al.

Duration of operation

Whilst the duration of anaesthesia and surgery werealmost identical in the frameless series, in the frame-based series the anaesthetic time included scanacquisition and was therefore considerably longerthan the duration of surgery. For the frameless casesthe procedure (operation and anaesthetic) lasted forbetween 20 and 180 min (mean 54.2, SD 23.6),whilst for the frame-based procedures the duration ofanaesthesia ranged from 80 to 235 minutes (mean127.4, SD 31.9). The duration of the operation in theframe-based series ranged from 25 to 145 minutes(mean 56.3, SD 17.5). Thus, the mean duration ofthe anaesthetic was significantly longer in the frame-based than the frameless series (p < 0.0001), butthere was no significant difference in the duration ofsurgery between the two groups (Fig. 4).

Complications

In the frame-based series 22 patients experiencedone or more complication (Table V). There werefive haemorrhages with one fatality from a deephemisphere haematoma. In one of these cases theburr hole was extended to a craniectomy to controlbiopsy site bleeding. Three patients suffered perma-nent neurological deficits from haemorrhage andone patient suffered a transient deficit. In sevenother cases there was a transient deterioration inneurological status due to increased vasogenic

oedema following surgery and two patients, who hadnot previously had fits, suffered epileptic seizures.There were eight postoperative infections (threechest infections, two urinary tract infections, onewound infection, one peritonitis following duodenalulcer perforation, one septicaemia secondary toprolonged ventilation after haemorrhage). Therewere also two cardiac complications (one transientleft ventricular failure and one sinus arrest) and onegastro-intestinal complication (the perforation of aduodenal ulcer).

In the frameless biopsy series 11 patientsexperienced complications. There were no casesof clinically significant haemorrhage. Seizuresoccurred in five cases following surgery, three ofwhom had not previously experienced fits. In fourcases there was deterioration in neurological statusdue to increased vasogenic oedema (one permanentdeterioration, three transient). One patient sufferedan infectious complication, succumbing to pneu-monia 2 months postoperatively, following aprolonged period on the ITU. One patient experi-enced gastritis that resolved following withdrawalof dexamethasone and one patient developedhypertension, requiring the introduction of newmedication.

Whilst the overall complication rate for theframe-based series was 22% and that for the frame-less series was significantly lower at 14% (p =0.0183), the rate of surgical complications was 8.8and 6.6%, respectively (Table V). There was onefatality in each series; there were significantly morehaemorrhagic complications and significantly morepermanent neurological deficit in the frame-basedseries than the frameless series (overall significancep = 0.0008). There were also more infectivecomplications in the frame-based series.

Cost analysis

The hospital accountants made no distinctionbetween frame-based and frameless biopsy forprocedural costs (4650 each, Table VI. AlthoughMRI was used more often and there was a smalladditional cost (10) for each frameless biopsy dueto the use of fiducials, the mean cost per case forframeless stereotactic biopsy (6723) was signifi-

FIG. 4. Bar graph showing the mean duration ofanaesthesia for both the frame-based and frameless series(error bars denote 1 SD).

TABLE V. Complications occurring in the Frameless and Frame-Based biopsy series

Complication Frame-based series Frameless series Combined series (%)

Death 1 1 2 (1.3)Haemorrhage (clinically significant) 5 0 5 (3.2)Permanent neurological deficit 3 1 4 (2.5)Transient neurological deficit 8 3 11 (7.0)Epilepsy (new) 2 5 7 (4.5)Infection 10 1 11 (7.0)Other system complications 3 2 5 (3.2)Total 22 11 33 (21.3)

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

Advantages of frameless stereotaxy 115

cantly lower than the mean cost per case for frame-based biopsy (7731, p = 0.0022). This arosethrough the significantly longer mean duration ofhospital stay in the frame-based series (6.5 days)than in the frameless series (mean 4.1 days, p =0.0013). Since the hospital stay for uncomplicatedcases was similar for both series this differencereflected the higher incidence of complications inthe frame-based series. In addition, the mean ITUstay in the frame-based group (0.33 days) was alsosignificantly greater than that for the frameless group(0.07 days, p = 0.02).

Capital costs for the frame-based series werepurchase cost (86,698 for CRW equipment) andservice contract (2500 per annum). The framewas replaced at 10 years and we performed 60biopsies per annum, resulting in a per case capitalcost of 186.16 for frame-based biopsy. Capitalcosts for the frameless series were purchase cost(107,000 for the basic cranial system) and service

contract (12,500 per annum). We performed 60biopsies and 120 open image-guided operations perannum, thus attributing 50% of the navigationequipment costs to the biopsy cases. The equip-ment lifetime was taken as 10 years resulting in aper case capital cost of 193.33 for framelessbiopsy.

Discussion

Development of biopsy techniques

Even in the CT era freehand burr-hole biopsycarried a mortality rate of over 5% and diagnosticyield of only 85%.9 Fortunately, this technique hasbeen superseded by the advent of stereotacticbiopsy, which carries a mortality rate of below 1%,a serious morbidity rate of 34% and a diagnosticyield of over 95%.18 There are, however, intrinsiclimitations to frame-based biopsy. Specificallyframe-based systems are point-based, requirecomplex calculations for each target, employ largecumbersome structures and require scanningwithin the frame (i.e. at the time of surgery). Overthe last decade image-guidance systems (IGS) havebeen developed which provide rapid image refor-mats to display target positions interactively. Thesesystems are intuitive, accurate and rapid. However,image-guidance workstations are essentially scan-display systems and are not stereotactic devices. Anumber of groups have used IGS to direct hand-held biopsy retrieval,20 but this is a retrograde step,whereby the proven accuracy of frame-based biopsyis abandoned for freehand biopsy (albeit withimage-guidance).

Our group have developed a system that combinesimage-guidance and stereotaxy to exploit theadvantages of each and overcome their intrinsicdisadvantages. Concurrently, other groups have alsodeveloped arm-based systems for frameless stereo-tactic biopsy,2123 but none have satisfactorilydemonstrated the laboratory and in vivo accuracy oftheir method, and several of these arms clearly eitherlack manoeuvrability or sufficient rigidity for wide-spread adoption. By contrast our system has a verysimple and ergonomic design and has proven accu-racy that compares favourably with frame-basedbiopsy.

TABLE VI. Mean costings (GB) per case for both of the biopsy series

n Frame-based Cost () n Frameless Cost ()

Operative procedure 1 4650 1 4650ITU stay (mean days) 0.34 456 0.07 96Ward stay (mean days) 6.22 2365 4.03 1531CT scans (mean number) 1.19 179 0.43 69MRI scans (mean number) 0.16 81 0.74 377Capital costs (per case) 193 186Total 7731 6723

FIG. 5. Bar graph demonstrating the mean depth of lesionbelow skin surface for the frameless and frame-basedbiopsy series. There was a significant difference betweenthe means (p < 0.0001, two-tailed t-test), but considerableoverlap (error bars denote 1 SD).

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

116 N. L. Dorward et al.

Defining frameless stereotaxy

Whilst there used to be no doubt over whether atechnique was stereotactic or not (a stereotacticframe was either used or was not) the introductionof IGS has brought some confusion in this regard.If, however, stereotaxy is defined as a system bywhich an instrument may be advanced directly to apre-selected discrete target, without deviation orcollateral brain injury it becomes clear that aninstrument guide that fixes rigidly, aligns accuratelyand guides instruments reliably is fundamental.Hand-held methods are ipso facto not stereotactic.Our group were able to successfully combine IGSand stereotaxy through development of an armwhich satisfied these requirements11 and we proposethat frameless stereotaxy is a useful term whenapplied strictly to point-targeted arm-based IGStechniques.

Accuracy

The mean error of the technique of frameless stereo-taxy assessed via laboratory tests12,24 was 1.3 mm(SD 0.6 mm). The mean error for 3 mm helical CTwas 0.9 mm (SD 0.5 mm) and that for 1.5T MRIwas 1.3 mm (SD 0.6 mm). In these tests, accuracyclosely paralleled the voxel size of the scansemployed. Slice thickness was similarly found to bethe major determinant of error by Maciunas and co-workers in their landmark tests of frame accu-racy.13,14 They found the application accuracy of theCRW frame to be 1.8 mm (SD 1.1, 95% CI3.6 mm) and with comparable scanning parametersthe accuracy of frameless stereotaxy was 1.1 mm(SD 0.5 mm, 95% Cl 2.1 mm). Our studies alsorevealed the mean in vivo error of frameless stereo-taxy to be 2.3 mm (SD 2.0 mm) with both CT andMRI guidance.12 There are no comparable in vivoaccuracy measurements available for frame-basedbiopsy with which to compare. A further study25

examined the impact of postimaging brain distortionin a wide variety of IGS cases and these results indi-cate that brain shift would have no significant impacton the accuracy of frameless stereotaxy.

Advantages of frameless stereotaxy

The technique of frameless stereotaxy was univer-sally popular in our Department, owing mainly to(i) the temporal separation of imaging from surgery(making the surgical procedure rapid and simple)and (ii) the enhanced image presentation for targetselection. In addition, the frameless techniqueenabled MRI-directed biopsy to be employed moreoften (there is no need for MRI-compatible equip-ment), multiple target selection was undemandingand frame fixation was not required. In this study,comparing 79 frame-based biopsies with 76 frame-

less stereotactic biopsies we found distinctadvantages with the frameless technique. Anaes-thetic time (and so theatre occupancy) weredramatically reduced (>50% reduction, p < 0.0001)as previously described,15 complications were signifi-cantly fewer (50% reduction, p = 0.0008) and thesecombined to make frameless stereotactic biopsysignificantly cheaper than frame-based biopsy (15%cost saving, p = 0.018). Most interestingly, analysisof the capital costs per case for these techniquesrevealed that there was little difference betweenframe-based and frameless costs when the additionaluses of the navigation equipment were taken intoaccount. Although the rate of complications forframe-based biopsy appears relatively high thisreflects the inclusion of all clinical events whetherdirectly related to surgery or not. In addition, detec-tion of complications with prospective collation isinevitably greater than in the retrospective reportsfound in the literature26 and certain complications(e.g. haemorrhage) have been shown to occur at avery high incidence, but usually remain unde-tected.27 Nonetheless, the possibility remains thatthe high rate of adverse events in the frame-basedseries is related to relative inexperience of the opera-tors, whereas the novel technique may have beenmore closely supervised. Although some of the timedifferential between the techniques is dependentupon the use of a general anaesthetic and patienttransfer, the advantages of temporal separation ofscanning and imaging plus the detailed imagepresentation and ease of target selection remainpertinent for frameless biopsy, whichever anaesthetictechnique is employed.

MRI was employed more often in the framelessseries, taking advantage of the superior anatomicalinformation and avoidance of ionising radiationpossible by this technique, without the penalties ofmajor image-distortion, equipment costs andcomplexity associated with frame-based MRIbiopsy.28 Whilst MRI is acknowledged to haveinferior spatial geometry compared with CT, theseerrors are low at the centre of a target volume, maybe minimized by careful selection of the imagingparameters and are much reduced in framelesscases by the absence of frame-induced distor-tion.29,30 In addition, CT scans are by no meansfree of geometric distortion29 and the improvedvisualization of targets may in large part explain thesuperior results of frameless stereotaxy.

In this study two techniques of stereotacticbiopsy were compared. The most appropriateimaging modality and target were chosen in eachindividual case with no attempt to standardizethese. Thus, MRI was more readily available in theframeless series with the advantages stated above.Similarly, the target was selected by reference to2D slices in the frame-based series and withorthogonal reformats in the frameless series. The

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

Advantages of frameless stereotaxy 117

consequence of this was that the series did notcompare two techniques of reaching the sametarget, rather two methodologies were compared,including their intrinsic constraints or lack ofrestrictions. This enabled the relative merits offrameless stereotactic biopsy to be demonstratedfully.

Conclusions

Frameless stereotactic biopsy is a simple, effectiveand intuitive technique which provides the advan-tages of reduced anaesthetic time, lower cost andfewer complications when compared with traditionalframe-based stereotactic biopsy. The term framelessstereotaxy, should only be applied to methods ofimage-guidance where instruments are directed withstereotactic precision to a preselected target by anavigation system, and not to hand-held methods.

Acknowledgements

This work was performed within the context of theEASI project European Applications for SurgicalInterventions, supported by the European Commis-sion under contract HC1012 in their 4thFramework Telematics Applications for HealthRTD programme. The partners in the EASI consor-tium are The National Hospital for Neurology andNeurosurgery in London, the Laboratory forMedical Imaging Research of the Katholieke Univer-siteit Leuven, the Image Sciences Institute of UtrechtUniversity & University Hospital Utrecht, PhilipsMedical Systems Nederland B.V., Philips ResearchLaboratory Hamburg, and the Image ProcessingGroup of Radiological Sciences at UMDS of Guysand St Thomas Hospitals in London.

None of the authors has any financial interest inthe devices presented in this article.

References

1 Apuzzo MLJ, Chandrasoma PT, Cohen D, et al.Computed imaging stereotaxy: experience andperspective related to 500 procedures applied to brainmasses. Neurosurgery 1987;20:9307.

2 Hall WA. The safety and efficacy of stereotactic biopsyfor intracranial lesions. Cancer 1998;82:174955.

3 Ostertag CB, Mennel HD, Kiessling M. Stereotacticbiopsy of brain tumours. Surg Neurol 1980;14:27583.

4 Ranjan A, Rajshekhar V, Joseph T, et al. Nondiag-nostic CT-guided stereotactic biopsies in a series of407 cases: influence of CT morphology and operatorexperience. J Neurosurg 1993;79:83944.

5 Regis J, Bouillot P, Rouby-Volot F, et al. Pineal regiontumours and the role of stereotactic biopsy: review ofthe mortality, morbidity and diagnostic rates in 370cases. Neurosurgery 1996;39:90712.

6 Swain PD, Hitchon PW, Follett KA, et al. Computedimaging-assisted sterotactic brain biopsy: a riskanalysis of 225 consecutive cases. Surg Neurol 1998;49:6409.

7 Thomas DGT, Nouby RM. Experience in 300 casesof CT-directed stereotactic surgery for lesion biopsyand aspiration of haematoma. Br J Neurosurg 1989;3:3215.

8 Wild AM, Xuereb JH, Marks PV, et al. Computer-ized tomographic stereotaxy in the management of200 consecutive intracranial mass lesions. Analysis ofindications, benefits and outcome. Br J Neurosurg1990;4:40715.

9 Marshall LF, Jennett B, Langfitt TW. Needlebiopsy for the diagnosis of malignant glioma. JAMA1974;228:141718.

10 Wen DY, Hall WA, Miller DA, et al. Targeted brainbiopsy: a comparison of freehand computed tomog-raphy-guided and stereotactic techniques. Neurosurgery1993;32:40713.

11 Dorward NL, Alberti O, Dijkstra A, et al. ClinicalIntroduction of an adjustable rigid instrument-holderfor frameless stereotactic interventions. Comput AidedSurg 1997;2:1805.

12 Dorward NL, Alberti O, Palmer JD, et al. Accuracy oftrue frameless stereotaxy; in vivo measurementand laboratory phantom studies. J Neurosurg 1999;90:1608.

13 Maciunas RJ, Galloway RL, Latimer JW, et al. Anindependent application accuracy evaluation of stereo-tactic frame systems. Stereotact Funct Neurosurg 1992;58:1037.

14 Maciunas RJ, Galloway RLJ, Latimer JW. The appli-cation accuracy of stereotactic frames. Neurosurgery1994;35:68294.

15 Alberti O, Dorward NL, Kitchen ND, et al. Neuron-avigation impact on operating time. Stereotact FunctNeurosurg 1997;68:448.

16 Dorward NL. Frameless stereotactic biopsy with theEasyGuide. Medica Mundi 1998;41:337.

17 Firlik KS, Martinez AJ, Lunsford LD. Use of cytolog-ical preparations for the intra-operative diagnosis ofstereotactically obtained brain biopsies: a 19 year expe-rience and survey of neuropathologists. J Neurosurg1999;91:4548.

18 Greene GM, Hitchon PW, Schelper RL, et al. Diag-nostic yield in CT-guided stereotactic biopsy ofgliomas. J Neurosurg 1989;71:4947.

19 Revesz T, Scaravilli F, Coutinho L, et al. Reliability ofhistological diagnosis including grading in gliomasbiopsied by image-guided stereotactic technique.Brain 1993;116:78193.

20 Gumprecht HK, Widenka DC, Lumenta CB.BrainLab VectorVision Neuronavigation System:technology and clinical experiences in 131 cases.Neurosurgery 1999;44:97104.

21 Barnett GH, Kormos DW, Steiner CP, et al. Use ofa frameless, armless stereotactic wand for braintumor localization with two-dimensional and threedimensional neuroimaging. Neurosurgery 1993;33:6748.

22 Germano IM, Queenan JV. Clinical experience withintracranial brain needle biopsy using framelesssurgical navigation. Comput Aided Surg 1998;31:339.

23 Patel N, Sandeman D. A simple trajectory guidancedevice that assists freehand and interactive imageguided biopsy of small deep intracranial targets.Comput Assisted Surg 1997;2:186192.

24 Dorward NL. Development and evaluation of image-guided neuroendoscopy, with investigation of post-imagingbrain distortion and accuracy of frameless stereotaxy.Master of Surgery Thesis, University of London, 1999.

25 Dorward NL, Alberti O, Velani B, et al. Post-imagingbrain distortion; magnitude, correlates and impact onneuronavigation. J Neurosurg 1998;88:65662.

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.

118 N. L. Dorward et al.

26 Allibone JB, Harkness WJ, Hayward RD. Craniotomy-related complications in a paediatric neurosurgeryunit a prospective study. Br J Neurosurg 1999;13:14853.

27 Kulkarni AV, Guha A, Lozano A, et al. Incidence ofsilent haemorrhage and delayed deterioration afterstereotactic brain biopsy. J Neurosurg 1998;89:315.

28 Bradford R, Thomas DGT, Bydder GM. MRI-directed stereotactic biopsy of cerebral lesions. ActaNeurochir Suppl (Wien) 1987;39:257.

29 Breeuwer M, Zylka W, Wadley JP, et al. Detection andcorrection of geomatric distortion in 3D CT images, inLemke HU, Vannier MW, Inamura K, Framan AG(eds) Computer Assisted Radiology and Surgery.Amsterdam: Elsevier Science, 1999, pp. 7882.

30 Walton L, Hampshire A, Forster DMC, et al. Stereo-tactic localisation with magnetic resonance imaging: aphantom study to compare the accuracy obtainedusing two-dimensional and three-dimensional imagingacquisitions. Neurosurgery 1997;41:1319.

Br J

Neu

rosu

rg D

ownl

oade

d fro

m in

form

ahea

lthca

re.c

om b

y U

nive

rsita

t de

Giro

na o

n 01

/12/

15Fo

r per

sona

l use

onl

y.