Risk Prediction Model for Basal Cell Carcinoma in Cardiac Allograft Recipients

Abstract

Basal cell carcinoma (BCC) is the second most common skin cancers in posttransplant patients. Long-term immunosuppression predisposes the patients to higher risk. This study was undertaken to develop a risk prediction model using the United Network for Organ Sharing (UNOS) database. Heart transplant recipients (2000~2015) from the UNOS database were analyzed. The Cox proportional hazards model was applied to screen the predictors associated with the development of BCC. Stepwise forward selection with Akaike information criterion was done to obtain the multivariate model. Area under the curve was derived from the receiver operating characteristics curve to assess the quality of the prediction model. A risk scoring system was developed to stratify patients into different risk groups, and the occurrence rates of posttransplant BCC among different groups were compared. There were 24,374 patients who received heart transplantation within this study period, and 1211 recipients have been reported with BCC. The multivariate model provides area under the curves at 5, 8, and 10 years posttransplant of 0.77, 0.76, and 0.76, respectively, in the derivation set and 0.75, 0.74, and 0.74, respectively, in the validation set. The predicted and observed probabilities of developing BCC in 5 years agree well across different risk groups. Kaplan-Meier survival curves were generated, which demonstrate significant differences between subjects in different risk groups. A risk prediction model has been generated for the first time for BCC with a c-statistic of ≥0.74 in both derivation and validation sets, making it a good tool for risk stratification.