Nomogram that predicts graft survival probability following living-donor kidney transplant.

ABSTRACT

OBJECTIVES: The goal of this project was to develop a nomogram that predicts the probability of graft survival at 5 years. MATERIALS AND METHODS: From our dataset, 1,581 patients were used to construct a nomogram (modeling group), the remaining 319 patients (testing group) were used for its validation. Initially, the modeling group variables were correlated with graft survival by univariate analysis. Significant factors were subjected to a multivariate analysis using a Cox regression model. The results formed the basis of our nomogram construction. Internal validation was done first by discrimination using the concordance index. Second, the calibration was assessed graphically. And finally, for external validation, the nomogram was used to predict graft survival using the testing group. The predicted probability(s) was compared with the actual survival estimates. RESULTS: Validation of the nomogram yielded a concordance index of 0.77, and the observed correspondence between predicted and actual outcomes suggested a high level of calibration. Nomogram predictions of the testing group revealed no differences in the means of predicted and observed graft survival at 5 years, with a high correlation coefficient and accepted predictive accuracy (concordance index, 0.72). CONCLUSIONS: We developed a well-validated and reasonably precise nomogram for predicting 5-year graft survival.