Reliable, digital, automated detection of interstitial fibrosis and tubular atrophy (IFTA) has not yet been developed. Machine learning (ML) can reproduce the renal pathologist’s visual assessment of IFTA and glomerulosclerosis. Well-trained ML methods not only showed similar agreement to that seen among renal pathologists for the assessment of IFTA and glomerulosclerosis, but also equivalent statistical association with patient outcome. These methods can help expedite research on very large digital archives of renal biopsy specimens, and may also benefit clinical practice by acting as a stand-in reading for pathology scenarios where renal expertise is limited or unavailable.
Select References
B. Ginley, K. Y. Jen, S. S. Han, L. Rodrigues, S. Jain, A. Fogo, J. Zuckerman, V. Walavalkar, J. C. Miecznikowski, Y. Wen, F. Wen, D. Yun, K. C. Moon, A. Rosenberg, C. Parikh, and P. Sarder, “Automated computational detection of interstitial fibrosis, tubular atrophy, and glomerulosclerosis,” Journal of the American Society of Nephrology, vol. 32, no. 4, pp. 837-50, Apr. 2021.
Underline indicates corresponding author.
* indicates equal contribution.
Ω indicates Dr. Sarder’s faculty trainees.
§ indicates post doctoral associates.
† indicates Dr. Sarder’s graduate students.
‡ indicates Dr. Sarder’s undergraduate students.