Reappraisal of the prognostic significance of vaginal invasion, parametrial invasion and lymph node metastasis in locally advanced cervical cancer patients receiving concurrent chemoradiotherapy: a multi-center large cohort study

OBJECTIVE: To construct and validate a predictive model to better differentiate the survival outcomes for locally advanced cervical cancer patients (LACC) receiving concurrent chemoradiotherapy (CCRT). METHODS: Patients were separately classified as vaginal invasion (VI), parametrial invasion (PI) or lymph node metastasis (LNM) based on disease extension. The survival outcomes were compared. Cox regression were performed to identify risk factors. Accordingly, the nomogram was constructed and validated. Moreover, the predictive performance between model and 2018 FIGO staging system was compared. RESULTS: LNM could significantly impair the survival for patients with VI (5-yr OS:77.7% VS 95.6%) (P = 0.018). Patients with bilateral extension to pelvic sidewall (BEP) and bilateral parametrial invasion (BPI) showed worse 5-yr OS (42.4% and 60.9%) compared to those with unilateral extension to pelvic sidewall (UEP) and unilateral parametrial invasion (UPI) (81.7% and 89.1%) (P < 0.001). Additionally, LNM didn’t affect the prognosis of patients with PI. The nomogram incorporating pathology, PI and pelvic LNM was constructed and showed decent predictive performance both in the training and test cohort. Besides, nomogram possessed better predictive ability than 2018 FIGO staging system both in the training (5-yr AUC: 0.79 VS 0.59) and validation cohort (5-yr AUC: 0.77 VS 0.64). CONCLUSIONS: LNM would impair the survival for patients with VI not those with PI and a nomogram was constructed and validated to better differentiate the prognosis for LACC patients than FIGO staging system.