Search

BackgroundDiagnosis of kidney transplant rejection currently relies on manual histopathological assessment, which is subjective and susceptible to inter-observer variability, leading to limited reproducibility. We aim to develop a deep learning system for automated assessment of whole-slide images (WSIs) from kidney allograft biopsies to enable detection and subtyping of rejection and to predict the prognosis of rejection.MethodWe collected H&E-stained WSIs of kidney allograft biopsies at 400x magnification from January 2015 to September 2023 at two hospitals. These biopsy specimens were classified as T cell-mediated rejection, antibody-mediated rejection, and other lesions based on the consensus reached by two experienced transplant pathologists. To achieve feature extraction, feature aggregation, and global classification, we employed multi-instance learning and common convolution neural networks (CNNs). The performance of the developed models was evaluated using various metrics, including confusion matrix, receiver operating characteristic curves, the area under the curve (AUC), classification map, heat map, and pathologist-machine confrontations.ResultsIn total, 906 WSIs from 302 kidney allograft biopsies were included for analysis. The model based on multi-instance learning enables detection and subtyping of rejection, named renal rejection artificial intelligence model (RRAIM), with the overall 3-category AUC of 0.798 in the independent test set, which is superior to that of three transplant pathologists under nearly routine assessment conditions. Moreover, the prognosis models accurately predicted graft loss within 1 year following rejection and treatment response for rejection, achieving AUC of 0.936 and 0.756, respectively.ConclusionWe first developed deep-learning models utilizing multi-instance learning for the detection and subtyping of rejection and prediction of rejection prognosis in kidney allograft biopsies. These models performed well and may be useful in assisting the pathological diagnosis.

Renal ischemia-reperfusion injury (IRI) contributes to acute kidney injury (AKI), increases morbidity and mortality, and is a significant risk factor for chronic kidney disease (CKD). Macrophage infiltration is a common feature after renal IRI, and infiltrating macrophages can be polarized into the following two distinct types: M1 macrophages, i.e., classically activated macrophages, which can not only inhibit infection but also accelerate renal injury, and M2 macrophages, i.e., alternatively activated macrophages, which have a repair phenotype that can promote wound healing and subsequent fibrosis. The role of TSC1, which is a negative regulator of mTOR signaling that regulates macrophage polarization in inflammation-linked diseases, has been well documented, but whether TSC1 contributes to macrophage polarization in the process of IRI is still unknown. Here, by using a mouse model of renal ischemia-reperfusion, we found that myeloid cell-specific TSC1 knockout mice (termed Lyz-TSC1 cKO mice) had higher serum creatinine levels, more severe histological damage, and greater proinflammatory cytokine production than wild-type (WT) mice during the early phase after renal ischemia-reperfusion. Furthermore, the Lyz-TSC1 cKO mice showed attenuated renal fibrosis during the repair phase of IRI with decreased levels of M2 markers on macrophages in the operated kidneys, which was further confirmed in a cell model of hypoxia-reoxygenation (H/R) in vitro. Mechanistically, by using RNA sequencing of sorted renal macrophages, we found that the expression of most M1-related genes was upregulated in the Lyz-TSC1 cKO group (Supplemental Table 1) during the early phase. However, C/EBPβ and CD206 expression was decreased during the repair phase compared to in the WT group. Overall, our findings demonstrate that the expression of TSC1 in macrophages contributes to the whole process of IRI but serves as an inflammation suppressor during the early phase and a fibrosis promoter during the repair phase.