Comparison of the predictive power of Sokal and ASXL1 and BIM variant genotypes and their combinations. (A) The data set was randomly split into training and test sets 100 times and the concordance index (C-index) was calculated. The C-index is a nonparametric measure to quantify the discriminatory power of a predictive model, in which a C-index of 0.5 indicates no predictive discrimination, and a C-index of 1 indicates perfect predictive accuracy. For each model, during each of the 100 times of random splitting, 80% of the total samples were used to train the model, and the remaining 20% were used as the test set for C-index calculations. Both the Cox multivariate and classification tree multivariate models showed performance superior to that of the Sokal risk score univariate model (one-sided Wilcoxon signed rank test P < .001). The boundaries of the box mark the first and third quartile, with the median in the center and whiskers extending to 1.5 interquartile range from the boundaries. The horizontal black dashed line marks the C-index equivalent to random guess (C-index = 0.5). Cumulative incidence of progression to AP/BC, stratified by (B) Sokal risk group and (C) classification tree risk group.