Explainable Biomarkers for Automated Glomerular and Patient-Level Disease Classification

Artificial intelligence can accurately distinguish IgA nephropathy from diabetic nephropathy under Masson staining and becomes an important assistant for renal pathologists

Introduction

Hyperplasia of the mesangial area is common in IgA nephropathy (IgAN) and diabetic nephropathy (DN), and it is often difficult to distinguish them by light microscopy alone, especially in the absence of clinical data. At present, artificial intelligence (AI) is widely used in pathological diagnosis, but mainly in tumor pathology. The application of AI in renal pathological is still in its infancy.

Methods

Patients diagnosed as IgAN or DN by renal biopsy in First Affiliated Hospital of Zhejiang Chinese Medicine University from September 1, 2020 to April 30, 2022 were selected as the training set, and patients who diagnosed from May 1, 2022 to June 30, 2022 were selected as the test set. We focused on the glomerulus and captured the field of the glomerulus in Masson staining WSI at 200x magnification, all in 1,000 × 1,000 pixels JPEG format. We augmented the data from training set through minor affine transformation, and then randomly split the training set into training and adjustment data according to 8:2. The training data and the Yolov5 6.1 algorithm were used to train the AI model with constant adjustment of parameters according to the adjusted data. Finally, we obtained the optimal model, tested this model with test set and compared it with renal pathologists.

Results

AI can accurately detect the glomeruli. The overall accuracy of AI glomerulus detection was 98.67% and the omission rate was only 1.30%. No Intact glomerulus was missed. The overall accuracy of AI reached 73.24%, among which the accuracy of IgAN reached 77.27% and DN reached 69.59%. The AUC of IgAN was 0.733 and that of DN was 0.627. In addition, compared with renal pathologists, AI can distinguish IgAN from DN more quickly and accurately, and has higher consistency.

Discussion

We constructed an AI model based on Masson staining images of renal tissue to distinguish IgAN from DN. This model has also been successfully deployed in the work of renal pathologists to assist them in their daily diagnosis and teaching work.

Machine learning in renal pathology

Introduction

When assessing kidney biopsies, pathologists use light microscopy, immunofluorescence, and electron microscopy to describe and diagnose glomerular lesions and diseases. These methods can be laborious, costly, fraught with inter-observer variability, and can have delays in turn-around time. Thus, computational approaches can be designed as screening and/or diagnostic tools, potentially relieving pathologist time, healthcare resources, while also having the ability to identify novel biomarkers, including subvisual features.

Methods

Here, we implement our recently published biomarker feature extraction (BFE) model along with 3 pre-trained deep learning models (VGG16, VGG19, and InceptionV3) to diagnose 3 glomerular diseases using PAS-stained digital pathology images alone. The BFE model extracts a panel of 233 explainable features related to underlying pathology, which are subsequently narrowed down to 10 morphological and microstructural texture features for classification with a linear discriminant analysis machine learning classifier. 45 patient renal biopsies (371 glomeruli) from minimal change disease (MCD), membranous nephropathy (MN), and thin-basement membrane nephropathy (TBMN) were split into training/validation and held out sets. For the 3 deep learningmodels, data augmentation and Grad-CAM were used for better performance and interpretability.

Results

The BFE model showed glomerular validation accuracy of 67.6% and testing accuracy of 76.8%. All deep learning approaches had higher validation accuracies (most for VGG16 at 78.5%) but lower testing accuracies. The highest testing accuracy at the glomerular level was VGG16 at 71.9%, while at the patient-level was InceptionV3 at 73.3%.

Discussion

The results highlight the potential of both traditional machine learning and deep learning-based approaches for kidney biopsy evaluation.

The fractal and textural analysis of glomeruli in obese and non-obese patients

Background

Fractal dimension is an indirect indicator of signal complexity. The aim was to evaluate the fractal and textural analysis parameters of glomeruli in obese and non-obese patients with glomerular diseases and association of these parameters with clinical features.

Methods

The study included 125 patients mean age 46 ± 15.2 years: obese (BMI ≥ 27 kg/m²—63 patients) and non-obese (BMI < 27 kg/m²—62 patients). Serum concentration of creatinine, protein, albumin, cholesterol, trygliceride, and daily proteinuria were measured. Formula Chronic Kidney Disease Epidemiology Colaboration (CKD-EPI) equation was calculated. Fractal (fractal dimension, lacunarity) and textural (angular second moment (ASM), textural correlation (COR), inverse difference moment (IDM), textural contrast (CON), variance) analysis parameters were compared between two groups.

Results

Obese patients had higher mean value of variance (t = 1.867), ASM (t = 1.532) and CON (t = 0.394) but without significant difference (P > 0.05) compared to non-obese. Mean value of COR (t = 0.108) and IDM (t = 0.185) were almost the same in two patient groups. Obese patients had higher value of lacunarity (t = 0.499) in comparison with non-obese, the mean value of fractal dimension (t = 0.225) was almost the same in two groups. Significantly positive association between variance and creatinine concentration (r = 0.499, P < 0.01), significantly negative association between variance and CKD-EPI (r = -0.448, P < 0.01), variance and sex (r = -0.339, P < 0.05) were found.

Conclusions

Variance showed significant correlation with serum creatinine concentration, CKD-EPI and sex. CON and IDM were significantly related to sex. Fractal and textural analysis parameters of glomeruli could become a supplement to histopathologic analysis of kidney tissue.

•

Variance showed significant correlation with eGFR calculated by CKD- EPI formula.

•

Significant correlation between variance and serum creatinine was found.

•

Textural contrast and inverse difference moment were significantly related to sex.

•

Fractal analysis of glomeruli could become supplement to histopathologic analysis.

•

Textural analyses of kidney tissue should become useful to histopathologic analysis.