Computerized scoring of histopathology for predicting coronary vasculopathy, validated by intravascular ultrasound

Background: Allograft coronary vasculopathy results from a complex interplay between immunologic and non-immunologic factors. We devised a computerized biopsy scoring method based on histopathology to predict the development of coronary vasculopathy. Methods: One hundred forty heart transplant recipients underwent serial intravascular ultrasound analysis at baseline (within 1 month) and at 1 year after transplantation and were evaluated for development of coronary vasculopathy (change in coronary maximal intimal thickness, CMIT). We evaluated serial endomyocardial biopsy specimens for cellular rejection, vascular rejection, ischemia, and fibrosis. In a mathematical model, we computed a biopsy score in each patient based on the duration and severity of histopathology. Results: We found a significant correlation between biopsy score (RY) and progression of coronary vasculopathy (r = 0.54, p = 0.001). Using a sensitivity analysis method, an RY value of ≥560 predicted development of coronary vasculopathy with a sensitivity of 86%, specificity of 62%, and diagnostic accuracy of 80%. Compared with patients with low-risk biopsy scores (RY < 560, n = 37), patients with high-risk biopsy scores (RY ≥ 560, n = 103) had increased progression of coronary vasculopathy (CMIT, 0.59 ± 0.29 vs 0.19 ± 0.10 mm, p < 0.001) and worse 7-year event-free survival (60% vs 91%, p = 0.01). Conclusion: The biopsy score is an effective method for predicting the development of coronary vasculopathy and for predicting outcome in cardiac transplant recipients.