Abstract: FR-OR082
Quantitative Glomerular Morphometry of Protocol Biopsies Predicts eGFR Decline in Living Donor Kidney Transplantation
Session Information
- The Changing Landscape of Kidney Transplantation: Clinical Innovation, Outcomes, and Health Policy
October 23, 2026 | Location: Mile High Ballroom 4A, Convention Center
Abstract Time: 05:30 PM - 05:40 PM
Category: Transplantation
- 2002 Transplantation: Clinical
Authors
- Krishnamoorthy, Anusha, Wallace H Coulter Department of Biomedical Engineering, Atlanta, Georgia, United States
- Tokuyama, Naoto, Emory University, Atlanta, Georgia, United States
- Abe, Yohei, Kagawa Daigaku Igakubu Daigakuin Igakukei Kenkyuka, Kita District, Kagawa Prefecture, Japan
- Fan, Fan, Wallace H Coulter Department of Biomedical Engineering, Atlanta, Georgia, United States
- Pathak, Tilak B., Emory University, Atlanta, Georgia, United States
- Subramanian, Ajay Rajaraman, Georgia Institute of Technology, Atlanta, Georgia, United States
- Dhamdhere, Rohan Narendra, Emory University, Atlanta, Georgia, United States
- Janowczyk, Andrew, Emory University, Atlanta, Georgia, United States
- Bingaman, Adam W., Emory University School of Medicine, Atlanta, Georgia, United States
- Larsen, Christian, Emory University School of Medicine, Atlanta, Georgia, United States
- Farris, Alton Brad, Emory University School of Medicine, Atlanta, Georgia, United States
- Madabhushi, Anant, Emory University, Atlanta, Georgia, United States
Background
Chronic allograft dysfunction remains a leading cause of graft loss, with glomerular injury central to its progression. However, Banff scoring relies on semi-quantitative visual assessment and does not capture detailed glomerular structural changes reproducibly. We developed a computational pipeline to characterize glomerular morphometry on routine H&E-stained 3-month protocol biopsies to predict subsequent eGFR decline.
Methods
We analyzed whole-slide images from 144 living-donor kidney transplant recipients from Kagawa University, with outcome defined as ≥20% eGFR decline by 3–24 months post-transplant. An average of 21 glomeruli per patient (range 8–49; total 2851) were analyzed. A computer vision-based pipeline segmented glomerular subcompartments to extract 131 morphometric features per glomerulus. Patients were split into training (n=96) and test (n=48) cohorts (2:1). Feature selection used univariate Cox regression followed by ElasticNet-penalized multivariable Cox regression, yielding a 5-feature morphometry risk score. Patients were dichotomized at the median predicted risk score and evaluated in the held-out test cohort using Kaplan–Meier analysis.
Results
Among 144 recipients, 55 (38.2%) experienced ≥20% eGFR decline (29.4 per 100 person-years). The final model selected 5 features: large-capillary fraction variability, capillary area variability, lumen spacing, wall–lumen ratio variability, and glomerular diameter. In the test cohort, the model identified a high-risk group with increased decline risk (HR 4.96, 95% CI 1.62–15.17; p=0.0019). The morphometry score remained independently associated with decline after adjustment (adjusted HR 6.26, 95% CI 1.75–22.43; p=0.0049) and improved discrimination versus clinical variables alone (C-index 0.671 vs 0.756).
Conclusion
Quantitative glomerular morphometry in early protocol biopsies predicts ≥20% eGFR decline by 24 months, enabling reproducible risk stratification beyond histopathology and supporting precision medicine in post-transplant surveillance.