[2026] Validation of Two Whole Slide Imaging Scanners Based on the Draft Guidelines of the College of American Pathologists

Michael J Thrall. Methodist Hospital, Houston, TX

Background: Whole slide imaging (WSI) converts glass slides into digital images that can be viewed remotely. We have four branch hospitals dependent on a central academic hospital for slide production and expert consultation. We intend to use WSI to avoid glass slide courier delays. We validated using the draft guidelines released by the College of American Pathologists in 2011.
Design: 100 consecutive cases were selected from the following categories that mimic the intended use: consults, frozen sections, malignancies, and special or immunohistochemical stains. Key slides were selected for each case. The slides were scanned at 20x magnification using standard focus and no Z-axis on 2 Ventana iScan Coreo Au scanners (the same 100 cases scanned twice). Pathologists viewed half of the cases as glass slides first and half as WSI first, then switched to the other modality after a delay of at least 3 weeks. Brief history was provided. Diagnoses were documented and intraobserver agreement compiled after the completion of all slide and WSI review.
Results: 23 pathologists participated; most viewed 9 cases. Intraobserver agreement was seen in 165 cases (82.5%), with disagreement in 35 cases (17.5%). Our most striking finding was that 10 cases (5%) showed disagreement between WSI and glass slides because of focal findings not seen in the WSI modality.

Focal Findings Not Seen by Whole Slide Imaging
Specimen SourceGlass Slide DiagnosisWSI DiagnosisWSI Scanner With Unseen Finding
BoneOsteomyelitisOsteoarthritisScanner #1
BreastAtypical lobular hyperplasiaNegativeBoth scanners
ColonFocal active colitisNegativeBoth scanners
EsophagusCandida seenNo Candida seenBoth scanners
LiverIron 1-2+No increase in ironScanner #2
StomachSuspicious for adenocarcinomaNegativeScanner #1
StomachChronic active gastritisChronic inactive gastritisScanner #1
WSI = Whole Slide Imaging


Conclusions: Overall, WSI performs well when compared to glass slides in our validation. The most concerning cause of intraobserver variability was diagnostically important small foci seen on glass slides but not WSI. Relative lack of familiarity with WSI, the slowness of WSI which discourages thorough scanning, and the "low stakes" setting of validation may all have contributed to this problem. However, the lack of good definition of the images at high power must be considered as a factor. Images produced with the fastest scan times (20x magnification and automatic focus) may not be sufficient for a confident negative diagnosis, especially in cases with subtle findings.
Category: Quality Assurance

Tuesday, March 5, 2013 9:30 AM

Poster Session III # 250, Tuesday Morning

 

Close Window