The Clinical Application of Artificial Intelligence Assisted Renal Biopsy Diagnosis System.

The Clinical Application of Artificial Intelligence Assisted Renal Biopsy Image Diagnosis System.

This research project aims to collect images from patients with chronic glomerulonephritis. For each subject, the images obtained from renal biopsy will undergo evaluation by both the diagnostic model and the 'gold standard' diagnosis by pathologists. The researchers will test the subjects using the diagnostic model and compare the results with the known 'gold standard' diagnosis, in order to evaluate the AUC, specificity, and sensitivity of the diagnostic model.

Renal biopsy pathology is an essential gold standard for the diagnosis of most glomerular diseases, relying on the comprehensive evaluation of H&E staining, special stains (such as PAS, PASM, and Masson), immunofluorescence, and the ultrastructural study under transmission electron microscopy (TEM). This research project aims to collect images from patients with chronic glomerulonephritis. For each subject, the images obtained from renal biopsy will undergo evaluation by both the diagnostic model and the 'gold standard' diagnosis by pathologists. The researchers will test the subjects using the diagnostic model and compare the results with the known 'gold standard' diagnosis, in order to evaluate the AUC, specificity, and sensitivity of the diagnostic model.

Inclusion Criteria:

1. Voluntary signing of informed consent form;
2. Patients clinically diagnosed or suspected of having chronic kidney disease according to the 2023 KDIGO Clinical Practice Guideline for the Evaluation and Management of Kidney Disease;
3. Undergoing renal biopsy and pathological specimen preparation.

Exclusion Criteria:

1. Biopsy tissue from donor kidney or transplanted kidney;
2. Poor quality of pathological specimen, unable to conduct pathological diagnosis.