多变量预测模型在临床效用中的测试与分析

Considering clinical utility, the multivariable prediction model was tested using a combination of 4 parameters in various permutations and combinations. The four indicators mentioned above were divided into prostate-related indicators and sphincter-related indicators, with AUC values of 0.81 and 0.82, respectively. Compared to models using single variables, this did not significantly improve the predictive value.