Table 1 Recommendations for validating WSI systems [25, 47]
Recommendation for validating WSI systems (PMID:2,363,490, | Training in WSI should be offered to participants In a 2022 study by Rizzo et al., in which 45 validation articles were reviewed for diagnostic issues 9% of the articles reported issues with misinterpretation of diagnosis and 6% of the articles reported issues with lack of confidence emphasizing the importance of training [46]. Also, a 2024 study by Koefoed-Nielsen et al., on implementation of digital pathology at two departments stressed the need for more system specific training before implementation [25] |
A sample set of atleast 60 routine cases for one application and another 20 cases for each additional application should be used | |
A washout period of atleast 2Â weeks between viewing the slide sets in each condition should be present | |
The performance of WSI review is considered non-inferior to light microscopy if the upper bound of the two sided 95% confidence interval of the difference between the overall major discrepancy rates of WSI review and light microscopy slides review diagnosis is 4% or less and if the upper bound of the two-sided 95% confidence interval of the overall major discrepancy rate of the WSI review diagnosis (relative to the reference diagnosis) is 7% or less for the same observer. If concordance is less than 95%, laboratories should investigate and take corrective action towards the cause |