Publication
Validation of Digital Pathology Platform for Metabolic-Associated Steatohepatitis for Clinical Trials
medRxiv
Abstract
Aims: Determine if pathologic assessment of disease activity in steatohepatitis, performed using Whole Slide Images (WSIs) on the AISight Clinical Trials platform, yields results that are comparable to those obtained from the analysis performed using glass slides.
Methods and Results: The accuracy of scoring for steatohepatitis (NAS >4 with >1 for each feature and absence of atypical features suggestive of other liver disease) performed on the WSI viewing platform was evaluated against scoring conducted on glass slides. Both methods were assessed for overall percent agreement (OPA) with a consensus 'ground truth' (GT) score, defined as the median score of a panel of 3 expert pathologists on glass slides. Each case was also read by 3 different pathologists, once on glass and once using WSIs with a minimum 2-week washout period between glass and WSI reads. It was demonstrated that the average OPA across 3 pathologists of WSI scoring with GT was non-inferior to the average OPA of glass scoring with GT (non-inferiority margin of -0.05, difference of -0.001, 95% CI of (-0.027,0.026), and p<0.0001). For each pathologist, there was a similar average OPA of WSI and glass reads with glass GT (pathologist A 0.843 and 0.849, pathologist B 0.633 and 0.605 and pathologist C 0.755 and 0.780), with intra-reader, inter-modality agreements per histologic feature being greater than published intrareader agreements.
Conclusion: Accuracy of digital reads for steatohepatitis using WSIs is equivalent to glass reads in the context of a clinical trial for scoring using the Clinical Research Network scoring system.
View PublicationMethods and Results: The accuracy of scoring for steatohepatitis (NAS >4 with >1 for each feature and absence of atypical features suggestive of other liver disease) performed on the WSI viewing platform was evaluated against scoring conducted on glass slides. Both methods were assessed for overall percent agreement (OPA) with a consensus 'ground truth' (GT) score, defined as the median score of a panel of 3 expert pathologists on glass slides. Each case was also read by 3 different pathologists, once on glass and once using WSIs with a minimum 2-week washout period between glass and WSI reads. It was demonstrated that the average OPA across 3 pathologists of WSI scoring with GT was non-inferior to the average OPA of glass scoring with GT (non-inferiority margin of -0.05, difference of -0.001, 95% CI of (-0.027,0.026), and p<0.0001). For each pathologist, there was a similar average OPA of WSI and glass reads with glass GT (pathologist A 0.843 and 0.849, pathologist B 0.633 and 0.605 and pathologist C 0.755 and 0.780), with intra-reader, inter-modality agreements per histologic feature being greater than published intrareader agreements.
Conclusion: Accuracy of digital reads for steatohepatitis using WSIs is equivalent to glass reads in the context of a clinical trial for scoring using the Clinical Research Network scoring system.
Authors
Path AI