Sarder Study Aims to Expand Horizon on Digital Pathology

“Our project aims to detect and quantify the earliest measurable glomerular structural changes from renal biopsies of DN patients, as well as identify patients at risk of renal failure, thus opening new windows to precision therapy,” Sarder says.

At the current rate, one in three U.S. adults will be diabetic by 2050. DN is a disease secondary to diabetes and accounts for half of the end-stage renal disease cases in the country.

Measurement of minute urinary albumin (microalbuminuria) is the most common non-invasive clinical biomarker of DN. In order to conclusively define DN severity, pathologists conduct qualitative manual estimation of glomerular structural damage in renal biopsies. 

However, renal glomerular structure in DN biopsies does not often correlate with less invasive clinical biometrics such as estimated glomerular filtration rate, urine protein, serum creatinine and glucose levels, Sarder says.

This traditional diagnostic method is approximate, subjected to user bias, time-consuming, and has low diagnostic precision in early disease stages, he notes, adding that manual hand identified features may not always accurately predict disease progression.

“Computational image analysis offers the opportunity to project clinical biometrics onto glomerular histological structures,” Sarder says. “This method provides finer precision in identifying structural changes that lead to physiological changes, which in turn reduces the required clinical resources and time for diagnosis, and provides clinicians with greater feedback to improve early intervention.”

“Our tools quantify glomerular features in histological renal tissue images more efficiently than manual methods. We have also derived a quantitative progression risk score describing DN progression risk estimated off only a single biopsy point,” he adds.

Specific aims of the study are to:

• develop a comprehensive, robust, and precise tool for automatic glomerular feature quantification from 2 µm renal tissue images
• computationally model pathological stages of DN biopsies
• computationally predict DN progression severity from initial renal biopsy and clinical follow-up biometrics

Sarder notes that his research team will employ and study a variety of tools from modern and classical artificial intelligence (AI) methods to achieve the project’s goals.

“UB is focusing on making an impact in health care via AI and digital pathology is one such tool,” he says. ”Our project is expanding the horizon on digital pathology as one of the few groups in the nation working on computational renal pathology.”

Sarder is principal investigator on the five-year, $1.5 million grant titled “Computational Imaging of Renal Structures for Diagnosing Diabetic Nephropathy,” and funded through the National Institute of Diabetes and Digestive and Kidney Diseases.

Co-investigators from the Jacobs School of Medicine and Biomedical Sciences are:

• John E. Tomaszewski, MD, SUNY Distinguished Professor and the Peter A. Nickerson, PhD, Professor and Chair of pathology and anatomical sciences
• Rabi Yacoub, MD, assistant professor of medicine in the Division of Nephrology

Co-investigators from other institutions are:

• Agnes B. Fogo, MD, of Vanderbilt University
• Sanjay Jain, MD, PhD, of Washington University School of Medicine in St. Louis
• Kuang-Yu Jen, MD, PhD, of University of California, Davis