Types and Frequency of Whole Slide Imaging Scan Failures in a Clinical High Through-Put Digital Pathology Scanning Laboratory

ABSTRACT

Background: Digital pathology adoption for clinical diagnostics continues to increase due to favorable regulatory environment and need for remote diagnosis during COVID-19 pandemic. Whole slide imaging (WSI) scan failure is fairly uncommon. However, there is limited literature about the true incidence of WSI scan failure rates and the impact on the daily operations in a setting of complete digital workflow. Our digital pathology scanning facility is one of the largest clinical digital pathology operating the world. Our facility routinely monitors scan failure data as a part of quality control and quality assurance. This study was undertaken to address the issues related to scan failure and to assess impact on turn-around times (TATs). This data would be beneficial to health care providers considering transition to a complete digital work-flow. Design: In 2017, we transitioned from scanning of archival slides to mostly new slides for primary diagnosis. We have operated 13 different Philips UFS scanners and scanned 2,289,266 slides representing nearly 233,864 cases. Scan failure data was collected from 3 resources (1) Errors detected by machine, (2) Retrospective quality control review and (3) Errors reported by pathologists. Every slide image is appraised by the scanner for defects including failed region of interest (ROI) detections, slides skipped, slides dropped, tissue not detected, and other faults. Each image is also checked by scan technician to determine if the ROI was correctly captured or not. Routinely 1.5% of the daily scans are inspected by senior staff for quality assurance. Slides are scored on a scale of 1 to 10 using different parameters and scans scoring <8 are designated as failed scans and slides typically get rescan. Total scan failure rates, re-scan (since 2019) rates were recorded and monitored. Results: Table 1 summarizes WSI scan failure data at our facility. Overall scan failure rate was just 1.19% with majority of the failures were attributable to machine error followed by failures due to slide preparation features. Most common machine error was failed ROI followed by skipped tissue error. Conclusions: WSI scan failure is extremely uncommon (1.19%) in a facility with experienced slide scanning staff and optimal slide preparations. Re-scanning was requested only for 1.19% cases and was feasible in 100% cases. Scanning of archival versus newly prepared slides did not have an impact on scan failure rates. Scan failure is not frequent enough to impact TATs and therefore need not be a concern for institutions considering transitioning to digital workflow. (Table Presented).