Validation of whole slide imaging in the primary diagnosis of gynaecological pathology in a University Hospital

Abstract

AimsExperience in the use of whole slide imaging (WSI) for primary diagnosis in pathology is very limited. We aimed to determine the accuracy of interpretation of WSI compared with conventional light microscopy (CLM) in the diagnosis of routine gynaecological biopsies.MethodsAll gynaecological specimens (n=452) received over a 2-month period at the Department of Pathology of the Hospital Clinic of Barcelona were analysed blindly by two gynaecological pathologists, one using CLM and the other WSI. All slides were digitised in a Ventana iScan HT (Roche diagnostics) at 200×. All discrepant diagnoses were reviewed, and a final consensus diagnosis was established. The results were evaluated by weighted κ statistics for two observers.ResultsThe level of interobserver agreement between WSI and CLM evaluations was almost perfect (κ value: 0.914; 95% CI 0.879 to 0.949) and increased during the study period: κ value 0.890; 95% CI 0.835 to 0.945 in the first period and 0.941; 95%; CI 0.899 to 0.983 in the second period. Major discrepancies (differences in clinical management or prognosis) were observed in 9 cases (2.0%). All discrepancies consisted of small lesions (8 high grade squamous intraepithelial lesions of the uterine cervix, one lymph node micrometastasis of an ovarian carcinoma) underdiagnosed or missed in the WSI or the CLM evaluation. Discrepancies with no or minor clinical relevance were identified in 3.8% of the biopsies. No discrepancy was related to the poor quality of the WSI image.ConclusionsDiagnosis of gynaecological specimens by WSI is accurate and may be introduced into routine diagnosis.

A Systematic Analysis of Discordant Diagnoses in Digital Pathology Compared With Light Microscopy

“…The first thing we could do is ensure pathologists are aware that dysplasia is a potential pitfall. Ordi et al 34 describe an increase in glass-digital concordance for cervical dysplasia as their study progressed, suggesting that there is a significant learning curve effect for digital dysplasia diagnosis. This is an area of diagnosis that may need a longer settling-in period before pathologists can confidently and safely sign out digital cases.…”

Conventional Microscopical versus Digital Whole-Slide Imaging-Based Diagnosis of Thin-Layer Cervical Specimens: A Validation Study

“…[1234567] One of the objectives of viewing scanned whole slides from histological or cytological specimen is remote exchange of knowledge and expertise of professionals to increase diagnostic accuracy. [12689] As a result of the increased implementation of WSI in research applications and clinical practice, guidelines for validation of WSI for use in primary diagnostics of histological specimens were recently proposed by the College of American Pathologists (CAP).…”

Validation of Whole-Slide Imaging for Histolopathogical Diagnosis: Current State

“…However, some studies have shown that although diagnosis with WSI is initially more time-consuming, this time quickly decreases as pathologists become familiar with the use of the WSI viewer [30,31,32,33,34,35]. Thus, there is a learning curve in the use of WSI and the time required for making a diagnosis, and a recent study conducted at our institution confirmed that the diagnostic performance improved with practice [36]. Another limitation of WSI is the relatively high costs of the equipment.…”

“…Major discrepancies were found in only 2% of the cases, and none was related to poor image quality. Most discrepancies in this study were observed in biopsies of premalignant lesions of the uterine cervix, an area which has shown high inter- and intraobserver variability rates using CLM [36]. The magnification used in the study was ×200, and higher magnification did not seem to be required.…”