Indian Journal of Pathology and Microbiology

: 2021  |  Volume : 64  |  Issue : 1  |  Page : 78--83

Validation of Whole Slide Imaging for primary surgical pathology diagnosis of prostate biopsies

Vidya Rao, Pavitra Subramanian, Akash P Sali, Santosh Menon, Sangeeta B Desai 
 Department of Pathology, Tata Memorial Centre, Homi Bhabha National Institute, Mumbai, Maharashtra, India

Correspondence Address:
Sangeeta B Desai
Department of Pathology, Annexe Building, Tata Memorial Hospital, E. Borges Road, Parel, Mumbai - 400 012, Maharashtra


Context: Whole slide imaging (WSI) is an important component of digital pathology which includes digitization of glass slides and their storage as digital images. Implementation of WSI for primary surgical pathology diagnosis is evolving, following various studies which have evaluated the feasibility of WSI technology for primary diagnosis. Aims, Settings and Design: The present study was a single-center, observational study which included evaluation by three pathologists and aimed at assessing concordance on specialty-specific diagnosis and comparison of time taken for diagnosis on WSI and conventional light microscopy (CLM). Materials and Methods: Seventy prostate core biopsy slides (reported between January 2016 and December 2016) were scanned using Pannoramic MIDI II scanner, 3DHISTECH, Budapest, Hungary, at 20× and 40×. Sixty slides were used for validation study following training with 10 slides. Statistical Analysis Used: Intraobserver concordance for diagnosis between the two platforms of evaluation was analyzed using Cohen's κ statistics and intraclass correlation coefficient (ICC); observation time for diagnosis was compared by Wilcoxon signed-rank test. Results: Interpretation on WSI using 20× and 40× was comparable with no major discordance. A high level of intraobserver agreement was observed between CLM and WSI for all three observers, both for primary diagnosis (κ = 0.9) and Grade group (κ = 0.7-0.8) in cases of prostatic adenocarcinoma. The major discordance rate between CLM and WSI was 3.3%–8.3%, which reflected the expertise of the observers. The time spent for diagnosis using WSI was variable for the three pathologists. Conclusion: WSI is comparable to CLM and can be safely incorporated for primary histological diagnosis of prostate core biopsies.

How to cite this article:
Rao V, Subramanian P, Sali AP, Menon S, Desai SB. Validation of Whole Slide Imaging for primary surgical pathology diagnosis of prostate biopsies.Indian J Pathol Microbiol 2021;64:78-83

How to cite this URL:
Rao V, Subramanian P, Sali AP, Menon S, Desai SB. Validation of Whole Slide Imaging for primary surgical pathology diagnosis of prostate biopsies. Indian J Pathol Microbiol [serial online] 2021 [cited 2023 Dec 1 ];64:78-83
Available from:

Full Text


Since the introduction of the first automated, high-resolution whole slide imaging (WSI) system by Art Wetzel and John Gilbertson in 1999, there have been rapid advancements in informatics and technology leading to the development of high throughput WSI scanners, revolutionizing the field of digital pathology with several applications thereof.[1]Current WSI systems utilize high speed, high-resolution image acquisition via refined illumination systems, microscope optics, and focusing systems followed by digital assembly to finally reproduce the entire tissue specimen on glass slide in a digital format which can be manipulated using a viewer software, closely simulating conventional light microscopy (CLM). Virtual microscopy serves as a high-quality solution for teleconsultation (including frozen section diagnosis and referral consultation in specialty practice), educational training, quality assurance and proficiency testing, archiving, and research. The College of American Pathologists (CAP) published digital pathology guidelines in 2013 recommending validation of WSI system for each of the intended use separately, before incorporation into routine workflow.[2]Recently, systematic analysis of diagnostic concordance between WSI and CLM by Goacher et al.[3]along with large cohort studies[4],[5] has demonstrated non-inferiority of WSI diagnosis over CLM diagnosis with no clinical difference and complete concordance qualifying FDA approval for digital pathology systems.

Despite the growing number of validation studies in primary diagnosis, relatively fewer studies in subspecialties, adequately including similar specimen types and addressing specialty-specific challenges, have contributed to reluctance in the full implementation of WSI in routine clinical practice. Evaluation of prostate core biopsies invariably includes informative prognostic parameters (Gleason score, grade group, perineural invasion (PNI), and extra-prostatic extension (EPE) if any, in the presence of adenocarcinoma) in addition to primary diagnosis.

Reproducibility of diagnosis on prostate core biopsies, particularly Gleason grading system has been previously studied in the past, with varying case selection, evaluator experience, and results ranging from 10% to 78% interobserver agreement.[6],[7],[8],[9],[10],[11],[12],[13],[14],[15],[16],[17],[18],[19],[20] However, there are fewer studies comparing primary diagnosis of prostate core biopsies rendered on CLM and WSI and far fewer including adequate sample size and uropathologists among evaluators.[21],[22],[23],[24],[25]

The primary aim of this study was to assess the concordance of specialty-specific diagnosis on WSI with CLM or “true diagnosis,” which is the gold standard. The study was designed to recapitulate the real-world diagnostic setting including pathologists with varying specialty experience and a sample set adequately representing the spectrum and complexity of diagnoses likely to be encountered during routine practice. Intraobserver agreement for diagnosis was established for each pathologist, in concordance with CAP guidelines for WSI validation.

 Materials and Methods

A total of 70 prostate core biopsies representing benign and malignant prostate pathology were included in this single-institution validation study, after approval from the Institutional Review Board. A single slide per case was selected. Coded glass slides were digitized at both 20× and 40× using Pannoramic MIDI II scanner (3DHISTECH; Budapest, Hungary) yielding high-resolution image acquisition (spatial sampling period: 0.274 μm/pixel at 20× and 0.139 μm/pixel at 40×). The images captured were assigned unique identifiers before evaluation, archived, and visualized using a web-based viewer software.

Evaluation of 60 core biopsies (study set) by three pathologists (2 uropathologists and 1 trainee pathologist) was undertaken after evaluation of 10 core biopsies (training set) since there was no prior experience with WSI interpretation. The three study pathologists were blinded to clinical reports and evaluated each of the 60 biopsies, first by virtual microscopy (WSI) and later by CLM after a wash-out period of 4 weeks, to minimize recall bias.

Parameters evaluated included diagnosis (benign/malignant), specific diagnosis, tumor quantitation (number of cores involved, percentage tumor infiltration), presence of PNI, lymphovascular emboli (LVI), and EPE if any. Diagnosis of prostatic adenocarcinoma was further qualified with Gleason score (GS) and Grade group (GG) as per WHO 2016 classification[26]and observation time, i.e., time taken for diagnosis was recorded.

Major discordances were defined as differences with significant clinical impact including misclassification of cases between benign and malignant pathology, migration between GG1 and GG2, or a difference of more than one grade group among the prostatic adenocarcinoma cases.

Minor discordances were defined as those with minimal clinical impact and included prostatic adenocarcinoma with difference by a single grade group (except migration between GG1 and GG2).


Data were analyzed using Cohen's κ statistic, a measure of agreement between observations, the magnitude of which reflects the strength of agreement.

Concordance for tumor quantitation parameters including total number of cores involved by the tumor and percentage of tumor infiltration was calculated using intraclass correlation coefficient (ICC).

Interpretation of strength of agreement was analyzed according to Douglas and Altman, where kappa value over 0.8 indicated excellent agreement, good level of agreement for kappa ranging between 0.6 and 0.8, moderate level of agreement for kappa ranging from 0.4 to 0.6, reasonably fair level of agreement for kappa between 0.4 and 0.2, and poor level of agreement for kappa <0.2.

The analyses of intraobserver agreement compared the WSI and CLM values recorded and were studied for each pathologist separately.

Time spent on arriving at a diagnosis using CLM and WSI was compared by Wilcoxon signed-rank test.

For all statistical analysis, IBM SPSS Statistics for Windows (V25.0; Armonk, NY: IBM Corp.) was used. Entire test was applied at a 5% significance level.


A total of 60 prostate core biopsies digitized at 20× and 40× were used for validation purposes. The images obtained by scanning the slides at 20× and 40× objectives were comparable in terms of quality, clarity, and resolution needed to report the digitized image. The average size of an image scanned at 20× objective was 200MB as compared to 554 MB when scanned at 40× objective. The observations made on both the sets of images were also comparable for all the three pathologists with no major discordances and only a few minor discordances. Hence, only the set of images scanned at 20× was used for the comparison with CLM.

The results for intraobserver agreement for different parameters are detailed in [Table 1] for each of the three pathologists. All three pathologists had an excellent agreement between CLM and virtual microscopy for primary diagnosis (benign/malignant) and specific diagnoses including benign prostate tissue, acute inflammation, chronic inflammation, acute and chronic inflammation, prostate atrophy, basal cell hyperplasia, atypical acinar cell proliferation, intraductal carcinoma, conventional acinar prostatic adenocarcinoma, and high-grade urothelial carcinoma (all κ < 0.9). Agreement for GS and GG was uniformly good with κ varying between 0.68 and 0.75 for GS and between 0.7 and 0.8 for GG.{Table 1}

Major and minor discordances between CLM and WSI diagnosis with the distribution of discrepant reads across all three pathologists are detailed in [Table 2].{Table 2}

Overall, there were 10 major discordances out of 180 reads across the three evaluating pathologists, with 5.56% discrepancy rate. The major discrepancy rate varied between 3.3% and 8.3% for individual pathologists, reflecting the level of expertise. All three pathologists had one case each which was misclassified between benign and malignant pathology on evaluation using conventional and digital platforms. Other major discrepancies were attributable to differences in GG < 1.

Minor discrepancies varied between 8.3% and 15% for individual pathologists and clustered around reads with GS 4 + 4 = 8/GG 4 (n = 16/22), with migration to lower grade group, GG3 (n = 7/16) or higher-grade group, GG5 (n = 9/16) on either WSI or CLM evaluation.

Concordance between WSI and CLM for additional prognostic indicators in cases of prostatic adenocarcinoma varied from moderate level to good (κ = 0.5–0.69) for PNI and a similar degree of agreement was noted for EPE (κ = 0.48–0.7) as well.

The presence of LVI was recorded as a single read by Pathologist 2 on CLM, whereas Pathologist 3 recorded the presence of tumor emboli in a single case on both CLM and virtual microscopy.

Tumor quantitation either as number of cores showing tumor or as percentage of tumor infiltration showed superior levels of agreement between WSI and CLM across all three pathologists with ICC ranging from 0.89 to 0.97 [Table 3].{Table 3}

The average time taken for diagnosis and comparison of the same on WSI evaluation and CLM are detailed in [Table 4].{Table 4}

For glass slides across all the three pathologists (180 reads), 79% cases required less than 60 s, 17% of the cases required 60 to 120 s and the remaining 4% cases required up to 180 s for diagnosis.

Comparably, for whole slide images, 67% cases required less than 60 s, 28% of the cases required time between 60 to 120 s, 4% cases required time up to 180 s, and 2 cases required additional time beyond 180 s.

There was a significant difference in time spent on arriving at a diagnosis for Pathologists 2 and 3 (P < 0.05) with WSI taking lesser time for Pathologist 2, whereas the same taking longer for Pathologist 3. There was no demonstrable significant difference between the two evaluation platforms in terms of observation time for Pathologist 1.


Implementation of WSI for routine evaluation of prostate core biopsies is highly feasible due to several contributing factors including small-sized digital files, informatics tools facilitating complex measurements, synchrony of H and E images with immunostained images, and a global view, thereby simplifying assessment of GS. Few studies collectively evaluating from 50 to over 800 cases have focused on evaluation of GS in prostatic core biopsies and scanning at 20× has been considered to be sufficient.[27]Similarly, in our study negligible discrepancies (0.6%–0.1%) were observed on WSI evaluation of images digitized at 20× and 40× while images scanned at 20× were digital files of comparably lesser size (200MB as compared to 550MB at 40×) and took lesser time for scanning (112 s average scanning time as compared to 323 s at 40×).

In the present study, there were very good comparable levels of intraobserver agreement among the three pathologists for evaluation by WSI and CLM for primary diagnosis, both broad diagnostic groups as well as specific diagnosis (κ < 0.9), GS individually (κ = 0.68–0.75) and similar when grouped under clinically relevant GG (κ = 0.7–0.8).

These results compare favorably with reproducibility studies for GS using CLM as well as with other WSI validation studies for prostatic biopsies [Table 5] and [Table 6].{Table 5}{Table 6}

One of the most important cytological features of prostate adenocarcinoma is prominent nucleoli. Nuclear enlargement, hyperchromasia, mitotic figures, and apoptotic bodies are other less important features of cancer. These features could have been missed or overlooked on WSI contributing to major discrepancies. Furthermore, identification of subtle features such as nucleolar prominence in suspected glands requires focusing in different planes of the section, which is possible in CLM, however, requires special features such as digitizing a glass slide at various Z-axis (vertical focal planes) in WSI, which was lacking in this study. In addition, a certain amount of discrepancy can be attributable to the lack of access to immunohistochemical stains in this study, which otherwise is a common aid for diagnosis in routine practice for difficult cases.

Major discordances that had GG difference of more than 1 grade or migration of GG with significant clinical implications for the three pathologists showed a declining trend with 8.3%, 5%, and 3.3% for Pathologists 1, 2, and 3, respectively, probably reflecting increasing experience in evaluation of prostate biopsies. In this case, it possibly reflected the evaluator's expertise rather than the difference between the two methodologies.

In the majority of cases with minor discordances, the difference was in assigning the secondary/higher pattern grade. Similar discrepancies were observed by Rodriguez et al.[21]and Chargari et al.[22]where there were higher levels of agreement for primary over secondary Gleason pattern.

Tumor quantification directly correlates with pathologic grade, stage and predicts biochemical recurrence and thus has consequently been incorporated into clinical nomograms for patient management with overwhelming consensus on additional means of tumor quantitation (percentage or mm of core involved by tumor) along with the number of cores involved.[28]

Invariably, given the refined tools for navigation at different magnification and digital measurements of tumor involvement, the accuracy and objectivity are highly reproducible with WSI systems.

Studies quantifying tumor involvement with WSI and CLM have shown comparable results with superior reproducibility of linear measurement of tumor attributable to digital measurement tools.[21],[22]

Perineural invasion is a commonly reported parameter assuming independent prognostic significance in patients groups stratified based on clinical stage, PSA levels, GS, and tumor extent despite conflicting data on its predictive value for EPE at radical prostatectomy or biochemical recurrence after surgery or radiation.[29]Similar to other parameters evaluated, there was relatively good agreement across the three pathologists further contributing to the non-inferiority of WSI compared to CLM. Interestingly, the lowest intraobserver agreement with Pathologist 1 (κ = 0.5) was attributable to higher observations of PNI made on WSI than on routine microscopy.

LVI is very rarely seen in needle biopsies mainly occurring in high-grade cancers and is a rarely noted parameter, thus not contributing much to the aim of the study. However, notably, there was perfect concordance for pathologist 3 on WSI and routine microscopy.

EPE has been associated with the significant predictive value of shorter time to biochemical progression following radical prostatectomy. This is a relatively infrequent finding, mainly seen in cases of large volume, high-grade cancer.

The intraobserver agreement for EPE was moderate to good (κ = 0.48–0.78) for each pathologist across routine and virtual microscopy, further contributing as additional evidence of their equality in evaluation beyond primary diagnosis.

Time for diagnosis using WSI and CLM across the three pathologists was variable with no difference, significantly lesser time with WSI and significantly more time with WSI, respectively. Despite comparable time taken for diagnosis using WSI as with CLM, in accordance with other studies evaluating prostatic core biopsies,[22],[23] time taken per case in a 12-core evaluation setting (which includes scanning time) has been a significant deterrent for integration into routine workflow necessitating high throughput scanners and superior interface features.

However, with increasing evidence validating equality of WSI with routine microscopy for primary diagnosis on prostate core biopsies, remote consultation for second opinions is still a feasible reality.

In addition, increasing experience among pathologists in using virtual microscopy along with technological innovations is expected to contribute to reduced turn-around time.

As an extended application of equality of digital evaluation with CLM, the unique feature of annotation of digital slides was utilized by the European Network of Urological Pathologists (ENUP) across 15 uropathology experts to generate summation areas of specific Gleason patterns similar to heat maps and data thus obtained helped in precisely identifying areas on which grading decisions were based on and foci requiring consensus on grading.[30]Thus, data generated from WSI in such settings can be used as an aid for standardization of Gleason grading.

Overall findings contribute to the growing evidence that histologic interpretation of routinely reported parameters on digital slides is comparable with routine microscopic evaluation even in a setting of specialty practice, with a number of immediate applications inherent to WSI.


We thank Mr. Akshay Ashok Govardhane, Application Specialist, Diagnostic Biosystems (India) for technical assistance.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Ho J, Parwani AV, Jukic DM, Yagi Y, Anthony L, Gilbertson JR. Use of whole slide imaging in surgical pathology quality assurance: Design and pilot validation studies. Hum Pathol 2006;37:322-31.
2Pantanowitz L, Sinard JH, Henricks WH, Fatheree LA, Carter AB, Contis L, et al. Validating whole slide imaging for diagnostic purposes in pathology: Guideline from the College of American Pathologists Pathology and Laboratory Quality Center. Arch Pathol Lab Med 2013;137:1710-22.
3Goacher E, Randell R, Williams B, Treanor D. The diagnostic concordance of whole slide imaging and light microscopy: A systematic review. Arch Pathol Lab Med 2016;141:151-61.
4Mukhopadhyay S, Feldman MD, Abels E, Ashfaq R, Beltaifa S, Cacciabeve NG, et al. Whole slide imaging versus microscopy for primary diagnosis in surgical pathology: A Multicenter blinded randomized noninferiority study of 1992 cases (Pivotal study). Am J Surg Pathol 2018;42:39-52.
5Snead DR, Tsang YW, Meskiri A, Kimani PK, Crossman R, Rajpoot NM, et al. Validation of digital pathology imaging for primary histopathological diagnosis. Histopathology 2016;68:1063-72.
6Svanholm H, Mygind H. Prostatic carcinoma reproducibility of histologic grading. Acta Pathol Microbiol Immunol Scand 1985;93:67-71.
7ten Kate FJ, Gallee MP, Schmitz PI, Joebsis AC, van der Heul RO, Prins ME, et al. Problems in grading of prostatic carcinoma: Interobserver reproducibility of five different grading systems. World J Urol 1986;4:147-52.
8Rousselet M, Saint-Andre J, Six P, Soret J. Reproducibility and prognostic value of Gleason's and Gaeta's histological grades in prostatic carcinoma. Ann Urol (Paris) 1986;20:317-22.
9de las Morenas A, Siroky MB, Merriam J, Stilmant MM. Prostatic adenocarcinoma: Reproducibility and correlation with clinical stages of four grading systems. Hum Pathol 1988;19:595-7.
10Di Loreto C, Fitzpatrick B, Underhill S, Kim DH, Dytch HE, Galera-Davidson H, et al. Correlation between visual clues, objective architectural features, and interobserver agreement in prostate cancer. Am J Clin Pathol 1991;96:70-5.
11Özdamar ŞO, Sarikaya Ş, Yildiz L, Atilla MK, Kandemir B, Yildiz S. Intraobserver and interobserver reproducibility of WHO and Gleason histologic grading systems in prostatic adenocarcinomas. Int Urol Nephrol 1996;28:73-7.
12Lessells AM, Burnett RA, Howatson SR, Lang S, Lee FD, McLaren KM, et al. Observer variability in the histopathological reporting of needle biopsy specimens of the prostate. Hum Pathol 1997;28:646-9.
13McLean M, Srigley J, Banerjee D, Warde P, Hao Y. Interobserver variation in prostate cancer Gleason scoring: Are there implications for the design of clinical trials and treatment strategies? Clin Oncol 1997;9:222-5.
14Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane CG, Epstein JI. Interobserver reproducibility of Gleason grading of prostatic carcinoma: General pathologist. Hum Pathol 2001;32:81-8.
15Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane CG, Amin MB, et al. Interobserver reproducibility of Gleason grading of prostatic carcinoma: Urologic pathologists. Hum Pathol 2001;32:74-80.
16Oyama T, Allsbrook WC Jr, Kurokawa K, Matsuda H, Segawa A, Sano T, et al. A comparison of interobserver reproducibility of Gleason grading of prostatic carcinoma in Japan and the United States. Arch Pathol Lab Med 2005;129:1004-10.
17Melia J, Moseley R, Ball R, Griffiths D, Grigor K, Harnden P, et al. A UK-based investigation of inter-and intra-observer reproducibility of Gleason grading of prostatic biopsies. Histopathology 2006;48:644-54.
18Ozkan TA, Eruyar AT, Cebeci OO, Memik O, Ozcan L, Kuskonmaz I. Interobserver variability in Gleason histological grading of prostate cancer. Scand J Urol 2016;50:420-4.
19Sadimin ET, Khani F, Diolombi M, Meliti A, Epstein JI. Interobserver reproducibility of percent Gleason pattern 4 in prostatic adenocarcinoma on prostate biopsies. Am J Surg Pathol 2016;40:1686-92.
20Shah RB, Li J, Cheng L, Egevad L, Deng F-M, Fine SW, et al. Diagnosis of Gleason pattern 5 prostate adenocarcinoma on core needle biopsy. Am J Surg Pathol 2015;39:1242-9.
21Rodriguez-Urrego PA, Cronin AM, Al-Ahmadie HA, Gopalan A, Tickoo SK, Reuter VE, et al. Interobserver and intraobserver reproducibility in digital and routine microscopic assessment of prostate needle biopsies. Hum Pathol 2011;42:68-74.
22Chargari C, Comperat E, Magné N, Védrine L, Houlgatte A, Egevad L, et al. Prostate needle biopsy examination by means of virtual microscopy. Pathol Res Pract 2011;207:366-9.
23Fine JL, Grzybicki DM, Silowash R, Ho J, Gilbertson JR, Anthony L, et al. Evaluation of whole slide image immunohistochemistry interpretation in challenging prostate needle biopsies. Hum Pathol 2008;39:564-72.
24Helin HO, Lundin ME, Laakso M, Lundin J, Helin HJ, Isola J. Virtual microscopy in prostate histopathology: Simultaneous viewing of biopsies stained sequentially with hematoxylin and eosin, and α-methylacyl-coenzyme A racemase/p63 immunohistochemistry. J Urol 2006;175:495-9.
25Van der Kwast TH, Evans A, Lockwood G, Tkachuk D, Bostwick DG, Epstein JI, et al. Variability in diagnostic opinion among pathologists for single small atypical foci in prostate biopsies. Am J Surg Pathol 2010;34:169-77.
26Moch H, Humphrey P, Ulbright T. WHO Classification of Genitourinary Tumors. Lyon: IARC; 2016.
27Saco A, Ramírez J, Rakislova N, Mira A, Ordi J. Validation of whole-slide imaging for histolopathogical diagnosis: Current state. Pathobiology 2016;83:89-98.
28Egevad L, Allsbrook WC Jr, Epstein JI. Current practice of diagnosis and reporting of prostate cancer on needle biopsy among genitourinary pathologists. Hum Pathol 2006;37:292-7.
29Harnden P, Shelley MD, Clements H, Coles B, Tyndale-Biscoe RS, Naylor B, et al. The prognostic significance of perineural invasion in prostatic cancer biopsies: A systematic review. Cancer 2007;109:13-24.
30Egevad L, Algaba F, Berney DM, Boccon-Gibod L, Compérat E, Evans AJ, et al. Interactive digital slides with heat maps: A novel method to improve the reproducibility of Gleason grading. Virchows Arch 2011;459:175-82.
31Al Nemer AM, Elsharkawy T, Elshawarby M, Al-Tamimi D, Kussaibi H, Ahmed A. The updated grading system of prostate carcinoma: An inter-observer agreement study among general pathologists in an academic practice. Apmis 2017;125:957-61.
32Mulay K, Swain M, Jaiman S, Gowrishankar S. Gleason scoring of prostatic carcinoma: Impact of a web-based tutorial on inter-and intra-observer variability. Indian J Pathol Microbiol 2008;51:22.
33Helin H, Lundin M, Lundin J, Martikainen P, Tammela T, Helin H, et al. Web-based virtual microscopy in teaching and standardizing Gleason grading. Hum Pathol 2005;36:381-6.
34Evans AJ, Henry PC, Van der Kwast TH, Tkachuk DC, Watson K, Lockwood GA, et al. Interobserver variability between expert urologic pathologists for extraprostatic extension and surgical margin status in radical prostatectomy specimens. Am J Surg Pathol 2008;32:1503-12.