focus on technology: how important is resolution in telecytopathology?
TRANSCRIPT
Focus on Technology
How Important Is Resolution in Telecytopathology?
Rachel Q. McMahon, BS, CT1; Erin E. McCarthy, BS, CT2; Scott J. Hetzel, MS3; Kasturi Das, MD4;
and Jimmie Stewart III, MD5
BACKGROUND: The authors conducted an analysis of 2 telepathology systems with different resolutions to determine
how resolution affects the pathologists’ ability to provide preliminary diagnoses for fine-needle aspirations (FNA). METH-
ODS: FNA cases evaluated by telepathology between February 1, 2011 and January 18, 2012 were reviewed. Concordance
indices between preliminary and final diagnoses were calculated for cases assessed with two proprietary systems (the
Remote Meeting Technologies iMedHD system and the Olympus NetCam system) using 3 diagnostic classifications (nega-
tive, atypical, and suspicious/positive). A Wilcoxon rank-sum test was used to compare the number of passes necessary
to determine adequacy. RESULTS: In total, 298 NetCam cases and 26 iMedHD cases were evaluated. The concordance
index, which was calculated using the 3 classifications, was 0.943 (95% confidence interval, 0.922-0.963) for NetCam
compared with 0.951 (95% confidence interval, 0.898-1.000) for iMedHD. The mean value for the number of passes
required to determine adequacy was 2.2 for NetCam and 2.1 for iMedHD (P 5.838). CONCLUSIONS: The results from sta-
tistical analyses demonstrated no difference in the concordance indices between preliminary and final diagnoses or in the
number of passes necessary to render adequacy between the 2 telepathology systems. However, because it had higher
resolution along with other features, the iMedHD system achieved greater user satisfaction. Cancer (Cancer Cytopathol)
2014;122:546-52. VC 2014 American Cancer Society.
KEY WORDS: telepathology; telecytopathology; immediate evaluations; preliminary diagnoses; fine-needle aspirations.
INTRODUCTION
Telepathology is defined by the College of American Pathologists as the practice of pathology in which a pathol-
ogist views digitized or analog images (video or still) and renders an interpretation for inclusion in a formal diag-
nostic report or for documentation in the patient record. The College of American Pathologists further
categorizes telepathology into different modes, including static telepathology (interpretation based on prese-
lected, still images), dynamic telepathology (viewing real-time images), and whole-slide imaging.1 The use of
telepathology as an educational and diagnostic tool has been on the rise since the mid 1980s.2,3 It has recently
proven to be sufficient for making timely and accurate preliminary diagnoses of specimens obtained by fine-nee-
dle aspiration (FNA).2,4–6 For the purposes of this article, this technology is referred to hereinafter as telecytopa-
thology, because it is solely used for the assessment FNA specimens.
Studies increasingly are demonstrating high concordance and accuracy rates for preliminary diagnoses
provided by dynamic telecytopathology versus direct microscopic evaluation. One 2008 study reported a
Corresponding author: Rachel McMahon, BS, CT, Cytopathology Laboratory/D4-259, University of Wisconsin Hospital and Clinics, 600 Highland
Avenue, Madison, WI 53792; Fax: (608) 263-6453; [email protected]
1Department of Cytopathology, University of Wisconsin Hospital and Clinics, Madison, Wisconsin; 2Department of Cytology, Wisconsin State
Laboratory of Hygiene, Madison, Wisconsin; 3Department of Biostatistics and Medical Informatics, University of Wisconsin, Madison, Wisconsin;4Department of Pathology, North Shore-LIJ Hospital Systems, Lake Success, New York; 5Department of Pathology and Laboratory Medicine,
University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin
Received: October 30, 2013; Revised: December 16, 2013; Accepted: December 20, 2013
Published online March 27, 2014 in Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/cncy.21404, wileyonlinelibrary.com
546 Cancer Cytopathology July 2014
Original Article
diagnostic concordance rate of 97% and a diagnostic
accuracy of 99% when comparing diagnoses rendered by
telecytopathology with the original diagnoses of previ-
ously finalized cases.4 Alsharif et al observed a discrepancy
rate as low as 1.8% between the preliminary diagnosis
established by telecytopathology and the final diagnosis of
371 cases.5 Kim et al reported that there was no significant
statistical difference between preliminary diagnoses estab-
lished by telecytopathology and on-site rapid diagnoses
compared with the respective final diagnoses of 40 pancre-
atic specimens obtained using endoscopic ultrasound
(EUS).6 The use of telecytopathology also reduces work
flow interruptions for pathologists whose offices may be
at a great distance from the FNA procedure area(s).7–9
Heimann et al reported that providing immediate assess-
ments by dynamic telecytopathology saved pathologists
30 minutes on average per procedure compared with pro-
viding on-site immediate assessments.2 A 2012 study
indicated that the mean procedure time for pathologists
was 0.2 hours when telecytopathology was used versus
0.74 hours for conventional on-site evaluations.10 Finally,
Kim et al reported an average total of 7.5 minutes spent
by the cytopathologist in rendering a rapid assessment for
FNA procedures by telecytopathology.6
In light of the increasing popularity of telecytopa-
thology, several imaging systems have become available,
making it difficult to decide which system would best suit
a growing cytopathology FNA service. The cytopathology
laboratory at the University of Wisconsin Hospital and
Clinics (UWHC) has access to 2 different dynamic telecy-
topathology systems, allowing for a direct comparison to
be made between the 2 in terms of diagnostic accuracy,
ease of use, and cost effectiveness. The current retrospec-
tive study was performed to compare the RMT iMedHD
system (Remote Meeting Technologies, Melville, NY)
with the Olympus NetCam system (Olympus, Center
Valley, Pa) largely with regard to resolution using statisti-
cal analysis as well as practical value.
MATERIALS AND METHODS
Background
The cytopathology laboratory at UWHC has been using
dynamic telecytopathology to provide preliminary diagno-
ses of cytologic specimens since 2005 using the Olympus
NetCam system for FNAs guided by ultrasound (US),
computed tomography (CT), EUS, and endobronchial
ultrasound (EBUS). In February 2011, the UWHC cyto-
pathology department began providing preliminary diag-
noses for FNA procedures at Watertown Regional Medical
Center (WRMC), a facility located approximately 47 miles
from UWHC, using the RMT iMedHD system. Table 1
outlines the specifications of the 2 systems.
A cytotechnologist is responsible for attending FNA
procedures at which he or she makes slides from the FNA
material (1 is air dried and 1 is fixed in ethanol for each
pass), stains the air-dried slides using Hema-Diff stain,
and screens them for adequacy. Before screening the slides
TABLE 1. Comparison of the Remote Meeting Tech-nologies iMedHD System and the Olympus NetCamSystem
RMT iMedHD (Remote
Meeting Technologies,Melville, NY)
Olympus NetCam(Olympus, Center Valley, Pa)
Workstation components
Microcast HD Pro-Lite
Camera and control box
(OPTRONICS, Goleta,
Calif); rmtiServ-900 HD
video encoder; cat5e ether-
net cable; microscope with
C-mount; desktop PC and
monitor
DP 70-72 digital camera (Olym-
pus, Center Valley, Pa); Micro-
Suite Five Software; PCI
graphics card with hard drive;
microscope with C-mount;
desktop PC and monitor
Resolution
1920 3 1080 Pixels, 60
frames per second; HD
performance (see Fig. 1)
800 3 600 Pixels; optimized for
static imaging (see Fig. 2)
Software
Web-based system, no
software required
Requires the Olympus MicroSuite
Five imaging software to be
loaded on each transmitting
workstation
Controls
Independent viewing
controls allow each guest
viewer to adjust the color,
brightness, and contrast
Only the system operator has
control over the camera
parameters
Security
Invitation and
authentication are required
for any participant to view
the streaming images,
providing heightened
security
No invitation or authentication
required, but all participants
must log into the hospital
intranet to view the images
Cost
Approximately $30,000
(includes charges for initial
project management,
intelligent video equipment
solution package and
image-capture feature) and
approximately $6600 per
year for ongoing yearly
service/support and RMT
MultiView5
Pricing information is not
available, because this system
is no longer available for
clinical use
Abbreviations: Cat5e, enhanced category 5 cable; HD, high definition; PC,
personal computer; PCI, peripheral component interconnect.
Focus on Technology/McMahon et al
Cancer Cytopathology July 2014 547
at 1 of 5 telecytopathology stations located outside of the
laboratory, the cytotechnologist will open the telecytopa-
thology program on the computer so the cytopathologist
on service can view the material from the laboratory or
from his or her office. The cytotechnologist is then able to
display diagnostic areas for the cytopathologist while pro-
viding pertinent clinical information and discussing the
case over the telephone. The clinician who performed the
FNA can see what the cytopathologist is viewing on the
computer monitor and can communicate by telephone as
well if needed. Once a preliminary diagnosis is established
by the cytopathologist, this information is conveyed to
the performing clinician by the cytotechnologist or
directly over the telephone by the cytopathologist.
Because the cytopathologist does not physically
attend the FNA procedures, it is possible to have 2 or
more procedures in different departments happening
simultaneously, thereby increasing the number of cases
for which he or she is available for guidance. Providing an
immediate adequacy assessment improves patient care
and clinician satisfaction by reducing the number of
repeat FNA procedures, and the addition of a preliminary
diagnosis allows for early triage of patients for the appro-
priate treatment plan. For example, patients in UWHC’s
endocrine thyroid FNA clinic may be scheduled for surgi-
cal evaluation if a preliminary diagnosis of atypical or
greater is reached. However, no definitive surgery is per-
formed until the final diagnosis is made.
Analysis
The study began with a query of the specimen database
for all of the FNA cases on which telecytopathology was
used at UWHC and WRMC between February 1, 2011
and January 18, 2012. At the time of data collection, there
were 3 cytopathologists on staff at UWHC with an experi-
ence range of approximately 8 to 25 years, and there were
6 cytotechnologists with an experience range of 6 months
to 20 years. Only the organ systems that are routinely
aspirated at both locations, including thyroid, lung,
lymph nodes, liver, and salivary gland, were included. All
of the specimens were obtained by the radiology depart-
ments of both institutions using CT or US guidance.
Cases that were given a preliminary diagnosis of
“nondiagnostic” were excluded, because as these results
were attributed to sampling error.
A statistical analysis was performed to determine
whether there was a higher concordance rate between pre-
liminary and final diagnoses for those cases that were
assessed using the RMT iMedHD system compared with
the cases that were assessed using the Olympus NetCam
system. The concordance indices and 95% confidence
intervals (CIs) between the preliminary and final diagno-
ses were calculated using the methods described in Harrel
et al.11 The indices were calculated using 3 classifications
(negative, atypical, and suspicious/positive). We chose to
combine the suspicious and positive categories because, in
many cases, a preliminary diagnosis of suspicious was
given only because the pathologist wanted to review the
fixed slides (stained with Papanicolaou stain) and/or cell
block material before making a positive diagnosis. To
compare the preliminary and final diagnoses of our cases,
both sets of data had to have matching diagnostic catego-
ries. Because the “nondiagnostic” category was eliminated
from our analysis, this led to the exclusion of 1 UWHC
case for the calculation of concordance indices; that case
was given a preliminary diagnosis of “atypical” but later
was finalized as “nondiagnostic.” The overall percentage
of preliminary diagnoses that matched the corresponding
final diagnoses also was calculated. Finally, a Wilcoxon
rank-sum test was used to compare the number of passes
necessary before a diagnosis was rendered at each location.
All statistical tests were 2-sided, with the significance level
set at 0.05.
RESULTS
In total, 298 cases from UWHC and 26 cases from
WRMC were evaluated (Table 2). The concordance index
for UWHC was 0.943 (95% CI, 0.922-0.963) compared
with a concordance index of 0.951 (95% CI, 0.898-
1.000) for WRMC. These results indicate that there was
no significant difference in concordance between prelimi-
nary and final diagnoses at the 2 locations.
There was no significant difference (P 5 0.838) in
the number of passes necessary to render adequacy between
the 2 sites. The mean 6 standard deviation number of
passes for UWHC was 2.2 6 1.5 passes, and the median
was 2 passes.1–3 For WRMC, the mean 6 standard devia-
tion number of passes was 2.1 6 1.2 passes, and the median
was 1.5 passes1–3 (Table 3). Table 4 provides a breakdown
of preliminary diagnoses in relation to the respective final
diagnoses. Table 5 indicates the percentage of cases in
which the preliminary and final diagnoses matched.
It is worth noting that the disparity in the number of
cases evaluated by the Olympus NetCam system and the
Original Article
548 Cancer Cytopathology July 2014
RMT iMedHD system could be considered a limitation
in this study. Although the statistical analyses conducted
in this study are valid even with a significant difference in
sample sizes, a similar study using more equivalent sample
sizes could be beneficial to confirm our results.
DISCUSSION
The results of our statistical analysis indicate no signifi-
cant difference in the concordance indices or in the num-
ber of passes necessary to render adequacy between the
Olympus NetCam system at UWHC and the RMT
iMedHD system at WRMC. This is somewhat surprising,
because it is counterintuitive to observe no difference
between a system with a substantially higher resolution
and another system with a much lower resolution. A par-
tial reason for this is the presence of a cytotechnologist on
site at the FNA procedure. Several studies have acknowl-
edged that a competent on-site operator is essential to the
successful use of a telecytopathology system.2–5,7,10 Per-
sonnel trained in cytomorphology can rapidly locate and
emphasize diagnostic areas on a slide. Transmitting these
areas on high magnification allows the cytopathologist to
make a preliminary diagnosis with relatively high accuracy
and few FNA passes. If individuals with a variably lower
level of training were used to operate the telecytopathol-
ogy system, then the lower resolution might have caused a
more significant difference.
Overall, our data demonstrate that both telecytopa-
thology systems allow cytopathologists to make accurate
preliminary diagnoses. Table 4 indicates the number of
cases in each diagnostic category that were upgraded,
downgraded, or remained in the same category once final-
ized. One case, a lymph node specimen from UWHC
(which subsequently was excluded from the concordance
index calculations), was given a preliminary diagnosis of
“atypical” and was later finalized as “nondiagnostic.” The
preliminary diagnosis was based on the presence of
necrotic material, which can be suggestive of carcinoma.
Upon review of all of the slides, the case was deemed
“nondiagnostic” because of the lack of viable cellular
material; the slides contained only inflammation and
necrosis. Similar discrepancies because of a paucity of
diagnostic material are not uncommon in telecytopathol-
ogy.2,4,5 Lack of diagnostic material is also a universal
problem for on-site evaluations. Another outlier on Table
4 was a lymph node specimen from UWHC that was
given a preliminary diagnosis of “negative” and was later
finalized as “positive.” The cytopathologist arrived at this
diagnosis after reviewing the flow cytometry results, which
revealed a lymphoproliferative disorder. Flow cytometric
analysis is routinely used to determine lymphoprolifera-
tive disorders that morphology cannot reliably diagnose.
Heimann et al reported a similar discrepancy in which a
bone aspirate originally called “blood and increased
white blood cells present” was later proven to be a B-cell
TABLE 2. Total Number of Cases From WatertownRegional Medical Center and University of Wiscon-sin Hospital and Clinics by Specimen Type
No. of Cases
Specimen Type UWHC WRMC Total
Liver 7 3 10
Lung 104 8 112
Lymph node 105 2 107
Salivary gland 17 2 19
Thyroid 65 11 76
Total 298 26 324
Abbreviations: UWHC, University of Wisconsin Hospital and Clinics;
WRMC, Watertown Regional Medical Center.
TABLE 3. Number of Passes Required to Make aPreliminary Diagnosis
No. of Passes
Variable UWHC (n 5 298) WRMC (n 5 26)
Median 2 1.5
Interquartile range 1-3 1-3
Mean 2.2 2.1
Standard deviation 1.5 1.2
Abbreviations: UWHC, University of Wisconsin Hospital and Clinics;
WRMC, Watertown Regional Medical Center.
TABLE 4. Cases Categorized by Preliminary Diag-nosis and Respective Final Diagnosis
Final Diagnosis
PreliminaryDiagnosis Nondiagnostic Negative Atypical
Suspicious/Positive
UWHC
Negative 0 65 12 1
Atypical 1 6 19 23
Suspicious/
positive
0 0 4 167
WRMC
Negative 0 9 2 0
Atypical 0 0 1 1
Suspicious/
positive
0 0 0 13
Abbreviations: UWHC, University of Wisconsin Hospital and Clinics;
WRMC, Watertown Regional Medical Center.
Focus on Technology/McMahon et al
Cancer Cytopathology July 2014 549
lymphoma after flow analysis.2 Table 5 demonstrates that
the majority of cases with a preliminary diagnosis of
“suspicious/positive” or “negative” were finalized in the
same diagnostic categories. There were more discrepancies
in the “atypical” category, because those diagnoses were
often upgraded or downgraded after the eventual review
of fixed slides and/or cell block material. This appears to
be a common scenario; Kern et al reported in a similar
study that 40% of 48 specimens that were reported as
inadequate during rapid assessment using telecytopathol-
ogy were shifted later into a diagnostic category upon
review of additional material or results from ancillary
studies.9
Although both telecytopathology systems were suffi-
cient for the pathologists to make an accurate preliminary
diagnosis, there are several other factors to consider when
comparing the Olympus NetCam and the RMT
iMedHD systems (Table 1). Most notably, we have
observed that the higher resolution (1920 3 1080 pixels)
and high-definition performance of the RMT iMedHD
system provides a much clearer image than the Olympus
NetCam system (800 3 600 pixels). This allows for better
TABLE 5. Percentage of All Cases in Which the Preliminary and Final Diagnoses Matched
UWHC WRMC
Diagnosis Preliminary Final Matched, % Preliminary Final Matched, %
Negative 78 65 83.3 11 9 81.8
Atypical 49 19 38.8 2 1 50
Suspicious/positive 171 167 97.7 13 13 100
Total 298 251 84.2 26 23 88.5
Abbreviations: UWHC, University of Wisconsin Hospital and Clinics; WRMC, Watertown Regional Medical Center.
Figure 1. This screenshot was taken of the cytopathologist’s monitor during a telecytopathology session using the RMT iMedHD
system (Remote Meeting Technologies, Melville, NY). This field of view represents papillary thyroid carcinoma at high-power
magnification (340). The iMedHD system displays a full screen image of the field of view, providing a clear view of the diagnostic
cells.
Original Article
550 Cancer Cytopathology July 2014
discernment of cellular detail and, in some cases, leads to a
shorter wait time before a preliminary diagnosis is estab-
lished (Figs. 1, 2). Figure 1 is a screenshot taken of the
cytopathologist’s monitor during a telecytopathology ses-
sion using the RMT iMedHD system, and Figure 2 is a
screenshot that was taken on the same monitor using the
Olympus NetCam system. Both images represent the
same field of view at high power (original magnification,
340) of papillary thyroid carcinoma. The RMT
iMedHD system displays a full screen image of the field of
view, providing a clear view of the diagnostic cells;
whereas the NetCam system displays a smaller image
within the browser window.
No software additions to the telecytopathology
workstations are needed when using the web-based RMT
iMedHD system, unlike the Olympus NetCam, which
requires loading the Olympus MicroSuite Five imaging
software program into each transmitting workstation.
The RMT iMedHD system is somewhat easier to use for
both the pathologist and the cytotechnologist, in that it
has independent viewing controls for each guest viewer, so
that he or she can adjust the color, brightness, and contrast
of the transmitted image as desired. The Olympus Net-
Cam MicroSuite Five imaging software only allows the
cytotechnologist driving the microscope to have control
over the camera parameters. Although this causes no great
difficulty in most cases, it can lead to delays if a cytopa-
thologist wants an image adjustment and requires the
cytotechnologist to have a higher level of knowledge about
the workings of the MicroSuite Five software. The RMT
iMedHD system provides more internal security than the
Olympus NetCam, because it requires an invitation and
authentication for any participant to view the streaming
images, which are streamed under appropriate encryption.
Therefore, violations of the Health Insurance Portability
and Accountability Act are not an issue with RMT
iMedHD. However, such violations are not a concern
with the Olympus NetCam either, because all participants
must log onto the hospital intranet to view the cases. If a
faculty member is outside the hospital and needs to view a
case using the Olympus NetCam, then he or she must use
a virtual private network provided by the hospital
Figure 2. This screenshot was taken of the cytopathologist’s monitor during a telecytopathology session using the Olympus Net-
Cam system (Olympus, Center Valley, Pa). This field of view represents papillary thyroid carcinoma at high-power magnification
(340). The NetCam system displays a smaller image within the browser window, and nuclear detail is more difficult to discern
than with the iMedHD image (see Fig. 1).
Focus on Technology/McMahon et al
Cancer Cytopathology July 2014 551
information technology department and sign into the
monitored hospital intranet. Finally, no patient informa-
tion is transmitted by telecytopathology on either system,
because any pertinent patient identification is communi-
cated by a concurrent telephone call.
From our own experience, we have observed that,
when trying to access the images from an off-site location,
the Olympus NetCam has a more pronounced lag time
than the RMT iMedHD system. The images streaming
on the RMT iMedHD system are refreshed at a higher
rate (60 frames per second) than those on the Olympus
NetCam, making the RMT iMedHD system more suita-
ble for use at remote locations. This is likely because of
the high-definition video camera purchased with the
RMT iMedHD system. Some lag time does occur with
this system and camera as well, but it is largely negligible
on a high-bandwidth network connection.
One drawback to the RMT iMedHD system is that
it comes at a higher cost than other telecytopathology sys-
tems on the market. The cost of 1 RMT iMedHD work-
station at UWHC is approximately $30,000 (this
includes charges for initial project management, intelli-
gent video equipment solution package, and image cap-
ture feature). The cost for ongoing, yearly service and
support and RMT MultiView5 (which allows for up to 5
concurrent viewers) is approximately $6600 per year. A
direct price comparison with the Olympus NetCam sys-
tem is not available, because that system is no longer mar-
keted for clinical use. A system comparable in price to the
Olympus NetCam is the Nikon DS-Fi2-L3 (Nikon,
Tokyo, Japan), a firmware-based system available for
approximately $7000. This system offers a maximum
resolution of 2560 3 1920 pixels with a maximum of 10
frames per second as per promotional sales information.
However, anecdotally, this resolution is not typically used
during telecytopathology sessions because of the low
frames per second available. A maximum of 37 frames per
second can be achieved during live display mode with a
lower resolution, allowing for faster transmission of
images.
In conclusion, our analysis demonstrates that a lower
resolution system can be used to provide preliminary diag-
noses with about the same statistical accuracy as a higher
resolution system. Many factors aside from resolution
contribute to the success of using telecytopathology for
preliminary diagnosis of FNA specimens. These factors
include ease of software use, internet/intranet bandwidth,
security factors, cost of the system, and personnel on site
at the FNA procedure. Although the lower resolution sys-
tem is sufficient for use on our FNA service, the RMT
iMedHD telecytopathology system allows the cytopatho-
logists at UWHC to make preliminary diagnoses with
considerable ease and confidence.
FUNDING SUPPORT
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
REFERENCES
1. Commission on Laboratory Accreditation-Laboratory AccreditationProgram. Laboratory General Checklist. Northfield, IL: College ofAmerican Pathologists; 2011.
2. Heimann A, Maini G, Hwang S, Shroyer RR, Singh M. Use oftelecytology for the immediate assessment of CT guided and endo-scopic FNA cytology. Diagn Cytopathol. 2010;40:575-581.
3. Williams S, Henricks WH, Becich MJ, Toscano M, Carter AB.Telepathology for patient care: what am I getting myself into? AdvAnat Pathol. 2010;17:130-149.
4. Kerr SE, Bellizzi AM, Stelow EB, Frierson HF Jr, Policarpio-Nicholas ML. Initial assessment of fine-needle aspiration specimensby telepathology: validation for use in pathology resident-facultyconsultations. Am J Clin Pathol. 2008;130:409-413.
5. Alsharif M, Carlo-Demovich J, Massey C, et al. Telepathology forimmediate evaluation of fine-needle aspiration specimens. CancerCytopathol. 2010;118:119-126.
6. Kim B, Chhieng DC, Crowe DR, et al. Dynamic telecytopathol-ogy of on site rapid cytology diagnosis for pancreatic adenocarci-noma [serial online]. Cytojournal. 2006;3:27.
7. Khurana KK. Telecytology and its evolving role in cytopathology.Diagn Cytopathol. 2012;40:498-502.
8. Kaplan KJ. Telecytopathology for immediate evaluation of fineneedle aspiration specimens. Cancer Cytopathol. 2010;118:115-118.
9. Kern I, Gabric S, Triller N, Pozek I. Telecytology for rapid assess-ment of cytological specimens. J Telemed Telecare. 2012;18:86-89.
10. Marotti JD, Johncox V, Ng D, Gonzalez JL, Padmanabhan V.Implementation of telecytology for immediate assessment of endo-scopic ultrasound-guided fine-needle aspirations compared to con-ventional on-site evaluation: analysis of 240 consecutive cases. ActaCytol. 2012;56:548-553.
11. Harrell FE, Califf RM, Pryor DB, Lee KL, Rosati RA. Evaluatingthe yield of medical tests. JAMA. 1982;247:2543-2546.
Original Article
552 Cancer Cytopathology July 2014