Establishment and Validation of Coagulation Factor-Based Nomogram for Predicting the Recurrence-Free Survival of Prostate Cancer

Abstract

<b><i>Introduction:</i></b> We aimed to establish and validate a coagulation feature-based nomogram to predict recurrence-free survival in prostate cancer patients. <b><i>Methods:</i></b> The study included 168 prostate cancer patients who had received radical prostatectomy between 2012 and 2018. Kaplan-Meier plot and log-rank analysis were used to screen recurrence-free survival-related features. The nomogram was established by combining the significant coagulation features with clinicopathological characteristics by using Cox regression analysis. The accuracy and clinical significance of the nomogram model were assessed by the receiver operating characteristic curve, Kaplan-Meier plot, and calibration plot. We explored the correlation between coagulation pathway activity and patient prognosis in public datasets by using gene set variation analysis (GSVA). <b><i>Results:</i></b> The results suggested that patients classified by the nomogram into the high-risk subgroup showed unfavorable prognoses compared with those in the low-risk subgroup in both the training (log-rank <i>p</i> &#x3c; 0.0001) and validation (log-rank <i>p</i> = 0.0004) cohorts. The nomogram model exhibited high discriminative accuracy in the training cohort (1-year area under the curve [AUC] of 0.74 and 3-year AUC of 0.69), which was confirmed in the internal validation cohort (C-index = 0.651). The calibration plots confirmed good concordance for the prediction of recurrence-free survival at 1 and 3 years. Subgroup analyses confirmed the utility of this model in different clinicopathological subgroups. Finally, GSVA suggested that patients with higher coagulation pathway scores mostly had unfavorable prognoses compared to those with lower scores, a result consistent with the findings above. <b><i>Conclusions:</i></b> We developed a practical nomogram model for predicting recurrence-free survival in prostate cancer patients. This model may offer clinicians prognostic assessments and facilitate personalized treatment.