Identification of glomerular lesions and intrinsic glomerular cell types in kidney diseases via deep learning

Abstract

Identification of glomerular lesions and structures is a key point for pathological diagnosis, treatment instructions, and prognosis evaluation in kidney diseases. These time‐consuming tasks require a more accurate and reproducible quantitative analysis method. We established derivation and validation cohorts composed of 400 Chinese patients with immunoglobulin A nephropathy (IgAN) retrospectively. Deep convolutional neural networks and biomedical image processing algorithms were implemented to locate glomeruli, identify glomerular lesions (global and segmental glomerular sclerosis, crescent, and none of the above), identify and quantify different intrinsic glomerular cells, and assess a network‐based mesangial hypercellularity score in periodic acid–Schiff (PAS)‐stained slides. Our framework achieved 93.1% average precision and 94.9% average recall for location of glomeruli, and a total Cohen's kappa of 0.912 [95% confidence interval (CI), 0.892–0.932] for glomerular lesion classification. The evaluation of global, segmental glomerular sclerosis, and crescents achieved Cohen's kappa values of 1.0, 0.776, 0.861, and 95% CI of (1.0, 1.0), (0.727, 0.825), (0.824, 0.898), respectively. The well‐designed neural network can identify three kinds of intrinsic glomerular cells with 92.2% accuracy, surpassing the about 5–11% average accuracy of junior pathologists. Statistical interpretation shows that there was a significant difference (P value < 0.0001) between this analytic renal pathology system (ARPS) and four junior pathologists for identifying mesangial and endothelial cells, while that for podocytes was similar, with P value = 0.0602. In addition, this study indicated that the ratio of mesangial cells, endothelial cells, and podocytes within glomeruli from IgAN was 0.41:0.36:0.23, and the performance of mesangial score assessment reached a Cohen's kappa of 0.42 and 95% CI (0.18, 0.69). The proposed computer‐aided diagnosis system has feasibility for quantitative analysis and auxiliary recognition of glomerular pathological features. © 2020 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of Pathological Society of Great Britain and Ireland.

Evaluation of Kidney Histological Images Using Unsupervised Deep Learning

“…Their main objective is to segment various structures present in the slide images, such as the glomeruli, 28 or to detect the glomeruli from the images, 29 or to extract novel pathologic findings from the glomeruli, 30 or to associate defined glomerular features with some pathologic findings or clinical variables. 5,31 Our study falls into the last category. One study used manually constructed features of the glomerulus images to detect proliferative lesions in the glomerulus, without using deep learning.…”

Multi-Task Learning-Based Immunofluorescence Classification of Kidney Disease

“…AI has been widely used in the recognition of renal pathology [29], such as the recognition of cortical or medulla, glomeruli, renal tubules, and renal arteries [30,31], as well as the recognition of internal glomerular structures, such as the podocytes, mesangial cells, and mesangial area [32]. However, in the final diagnosis of kidney disease, the type, location, and shape of immune complex deposits are still important bases and have distinct characteristics, especially in IgAN and MN.…”

The evolution and future of diabetic kidney disease research: a bibliometric analysis

“…Moreover, artificial intelligence is widely applied in the analysis of clinical indicators, digital imaging data, and digital pathological data in renal diseases [46,47] and improves the diagnosis and prognostication of DKD [48][49][50]. High-performance models built by artificial intelligence may contribute to more effective and accurate interventions in the clinical practice of DKD.…”