Assessment of a computerized quantitative quality control tool for whole slide images of kidney biopsies

Abstract

Inconsistencies in the preparation of histology slides and whole‐slide images (WSIs) may lead to challenges with subsequent image analysis and machine learning approaches for interrogating the WSI. These variabilities are especially pronounced in multicenter cohorts, where batch effects (i.e. systematic technical artifacts unrelated to biological variability) may introduce biases to machine learning algorithms. To date, manual quality control (QC) has been the de facto standard for dataset curation, but remains highly subjective and is too laborious in light of the increasing scale of tissue slide digitization efforts. This study aimed to evaluate a computer‐aided QC pipeline for facilitating a reproducible QC process of WSI datasets. An open source tool, HistoQC, was employed to identify image artifacts and compute quantitative metrics describing visual attributes of WSIs to the Nephrotic Syndrome Study Network (NEPTUNE) digital pathology repository. A comparison in inter‐reader concordance between HistoQC aided and unaided curation was performed to quantify improvements in curation reproducibility. HistoQC metrics were additionally employed to quantify the presence of batch effects within NEPTUNE WSIs. Of the 1814 WSIs (458 H&E, 470 PAS, 438 silver, 448 trichrome) from n = 512 cases considered in this study, approximately 9% (163) were identified as unsuitable for subsequent computational analysis. The concordance in the identification of these WSIs among computational pathologists rose from moderate (Gwet's AC1 range 0.43 to 0.59 across stains) to excellent (Gwet's AC1 range 0.79 to 0.93 across stains) agreement when aided by HistoQC. Furthermore, statistically significant batch effects (p < 0.001) in the NEPTUNE WSI dataset were discovered. Taken together, our findings strongly suggest that quantitative QC is a necessary step in the curation of digital pathology cohorts. © 2020 The Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Computer extracted gland features from H&E predicts prostate cancer recurrence comparably to a genomic companion diagnostic test: a large multi-site study

“…UMAP embedding 41 was performed on the validation set to assess the inter-site variation between images prior to Histotyping analysis and to verify that Histotyping features were resilient to batch effects across multiple sites. Such sources of pre-analytic variation can arise from differences in specimen preparation and scanning between laboratories, are correlated with the site, and have been shown to degrade the performance of digital pathology analysis algorithms 34,42,43 . UMAP was used to reduce the feature space to two dimensions for evaluating the clustering between slides from different laboratories.…”

Assessment of glomerular morphological patterns by deep learning algorithms

“…Algorithms for segmentation of healthy kidney parenchyma have been previously successfully developed [4,5]. Similarly, computational morphologic analyses of diabetic nephropathy, mesangial proliferation, and IgA-Nephropathy pattern have been published [6][7][8][9]. However, until now, no CNN-based approach that simultaneously deals with various glomerular lesions and that can discern them from unaffected glomeruli has been reported.…”

Automated quality assessment of large digitised histology cohorts by artificial intelligence

“…Currently this is generally reliant upon manual curation of images being considered for analysis, which is time consuming and inefficient, particularly in the context of a lack of pathologist resource. Furthermore, the reliance upon human observers for QA is perhaps questionable given the recognised subjectivity and poor inter-observer concordance in such tasks, even amongst expert pathologists 16 , 17 . This issue is further complicated by the interpretation of what is appropriate quality or ‘usable’ which for the expert pathologist is dependent upon the diagnostic question.…”