Automated scoring of total inflammation in renal allograft biopsies

Abstract

Background Computer‐assisted scoring is gaining prominence in the evaluation of renal histology; however, much of the focus has been on identifying larger objects such as glomeruli. Total inflammation impacts graft outcome, and its quantification requires tools to identify objects at the cellular level or smaller. The goal of the current study was to use CD45 stained slides coupled with image analysis tools to quantify the amount of non‐glomerular inflammation within the cortex. Methods Sixty renal transplant whole slide images were used for digital image analysis. Multiple thresholding methods using pixel intensity and object size were used to identify inflammation in the cortex. Additionally, convolutional neural networks were used to separate glomeruli from other objects in the cortex. This combined measure of inflammation was then correlated with rescored Banff total inflammation classification and outcomes. Results Identification of glomeruli on biopsies had high fidelity (mean pixelwise dice coefficient of .858). Continuous total inflammation scores correlated well with Banff rescoring (maximum Pearson correlation .824). A separate set of thresholds resulted in a significant correlation with alloimmune graft loss. Conclusions Automated scoring of inflammation showed a high correlation with Banff scoring. Digital image analysis provides a powerful tool for analysis of renal pathology, not only because it is reproducible and can be automated, but also because it provides much more granular data for studies.

Artificial Intelligence Advances in Transplant Pathology

“…Table 2 provides a comparative overview of these models. The application of AI in transplant kidney pathology has been examined in a number of studies [ 4 , 32 , 33 , 34 ].…”

Multiple abnormal peripheral blood gene expression assay results are correlated with subsequent graft loss after kidney transplantation

“…eGFR at 12-monts post transplantation was lower in both the 1 Not-TX and >1 Not-TX cohorts when compared to the TX group. The eGFR at 12 months was 47 ( 16) mL/min/m 2 in the >1 Not-TX cohort and 59 (16) mL/min/1.73 m 2 in the TX cohort (p < .001) The eGFR at 12 months was also lower in the 1 Not-TX group (51 [16] mL/min/1.73 m 2 ) compared to the TX group (p = .003). We also performed a linear regression analysis with selected baseline variables associated with eGFR at 12 months including the number of Not-TX results, presence of DGF, donor age in years (as a surrogate for KDPI, so as to not remove all living donors from the analysis), HLA mismatch, and recipient age in years.…”

“…Non‐invasive assays have been developed to correlate with rejection on the biopsy. However, it is well‐known that the scoring of rejection by pathologists is subject to wide variability 15–17 . Thus, when using an established assay on a new dataset such as the one in the study, the performance of the assay (sensitivity/specificity) might vary because the pathologist's opinions likely differed from the pathologists who were part of the development of the original classifier.…”

“…However, it is well-known that the scoring of rejection by pathologists is subject to wide variability. [15][16][17] Thus, when using an established assay on a new dataset such as the one in the study, the performance of the assay (sensitivity/specificity) might vary because the pathologist's opinions likely differed from the pathologists who were part of the development of the original classifier. Visual estimation of Banff scores by pathologists is a flawed "gold standard" for non-invasive assays.…”

“…The “Gold Standard” for the diagnosis of acute rejection is the interpretation of the changes on kidney allograft biopsy. However, it is well‐known that the scoring of rejection by pathologists is subject to wide variability 15–17 . Studies of the performance of peripheral blood assays have been based on the interpretation of the kidney allograft biopsy; and this variation in classification based on the biopsy interpretation may account for a proportion of the apparent errors of the performance of peripheral blood biomarker assays.…”