Whole Slide Imaging Telepathology (WSITP) for Primary Diagnosis in Surgical Pathology: A Comprehensive Validation Study at University Health Network (UHN)
Andrew J Evans, Sylvia L Asa. University Health Network, Toronto, ON, Canada
Background: UHN pathologists began using WSITP for making frozen section diagnosis in 2006. Given the success of this program, we are now considering using WSITP for routine primary diagnosis for surgical pathology cases originating at a partner hospital located 400 miles north of Toronto. 200-400 slides (25-40 cases) per day are sent from this facility to UHN for sub-specialty reporting. Prior to implementing WSITP, to avoid shipping these cases to UHN, there is an absolute need to ensure that the diagnostic performance of WSITP is at least equivalent to light microscopy (LM).
Design: As a practical approach to WSITP validation, we are determining whether the same pathologist can make the same diagnosis for cases reviewed initially by WSITP and then by LM. UHN pathologists, in 13 sub-specialty areas, prospectively select cases that represent what they encounter for primary diagnosis. Full cases are digitized using a scanner (Omnyx, LLC) and then reviewed using Omnyx viewing software. Preliminary diagnoses are recorded before the same pathologist reviews the cases by LM prior to final sign-out. There is no washout period between digital and LM review. The ability or inability to make a digital diagnosis along with diagnostic concordance between digital and LM are recorded. Discrepancies are considered major or minor, depending on their perceived impact on patient care.
Results: In the first 1.5 months of this 6-12 month study, 184 cases (> 1900 slides) from 8 sub-specialty areas have been reviewed. The cases range from small biopsies (1 slide with multiple levels) to large resection specimens (> 45 slides). Technical issues related to sub-optimal histology have prevented pathologists from making digital diagnoses in 2 cases and minor discrepancies have been noted in 3 cases. No major discrepancies have been observed. Complete agreement between digital and LM diagnosis has occurred in 179 cases (97%).
Conclusions: We have every indication that WSITP will be safe and accurate for making primary diagnoses on > 95% of cases originating at our northern partner site. It is likely that digital diagnoses would need to be deferred to glass slide review in only a small minority of cases. By no longer shipping all of these cases to UHN, WSITP will result in reduced transportation costs, removal of the risk of slides becoming lost or damaged during transport and reduced turnaround times with improved continuity of patient care.
Tuesday, March 20, 2012 1:00 PM
Poster Session IV # 230, Tuesday Afternoon