BackgroundInterstitial fibrosis, tubular atrophy (IFTA), and glomerulosclerosis are indicators of irrecoverable kidney injury. Modern machine learning (ML) tools have enabled robust, automated identification of image structures that can be comparable with analysis by human experts. ML algorithms were developed and tested for the ability to replicate the detection and quantification of IFTA and glomerulosclerosis that renal pathologists perform.MethodsA renal pathologist annotated renal biopsy specimens from 116 whole-slide images (WSIs) for IFTA and glomerulosclerosis. A total of 79 WSIs were used for training different configurations of a convolutional neural network (CNN), and 17 and 20 WSIs were used as internal and external testing cases, respectively. The best model was compared against the input of four renal pathologists on 20 new testing slides. Further, for 87 testing biopsy specimens, IFTA and glomerulosclerosis measurements made by pathologists and the CNN were correlated to patient outcome using classic statistical tools.ResultsThe best average performance across all image classes came from a DeepLab version 2 network trained at 40× magnification. IFTA and glomerulosclerosis percentages derived from this CNN achieved high levels of agreement with four renal pathologists. The pathologist- and CNN-based analyses of IFTA and glomerulosclerosis showed statistically significant and equivalent correlation with all patient-outcome variables.ConclusionsML algorithms can be trained to replicate the IFTA and glomerulosclerosis assessment performed by renal pathologists. This suggests computational methods may be able to provide a standardized approach to evaluate the extent of chronic kidney injury in situations in which renal-pathologist time is restricted or unavailable.

Background. Several groups have previously developed logistic regression models for predicting delayed graft function (DGF). In this study, we used an automated machine learning (ML) modeling pipeline to generate and optimize DGF prediction models en masse. Methods. Deceased donor renal transplants at our institution from 2010 to 2018 were included. Input data consisted of 21 donor features from United Network for Organ Sharing. A training set composed of ~50%/50% split in DGF-positive and DGF-negative cases was used to generate 400 869 models. Each model was based on 1 of 7 ML algorithms (gradient boosting machine, k-nearest neighbor, logistic regression, neural network, naive Bayes, random forest, support vector machine) with various combinations of feature sets and hyperparameter values. Performance of each model was based on a separate secondary test dataset and assessed by common statistical metrics. Results. The best performing models were based on neural network algorithms, with the highest area under the receiver operating characteristic curve of 0.7595. This model used 10 out of the original 21 donor features, including age, height, weight, ethnicity, serum creatinine, blood urea nitrogen, hypertension history, donation after cardiac death status, cause of death, and cold ischemia time. With the same donor data, the highest area under the receiver operating characteristic curve for logistic regression models was 0.7484, using all donor features. Conclusions. Our automated en masse ML modeling approach was able to rapidly generate ML models for DGF prediction. The performance of the ML models was comparable with classic logistic regression models.