Abstract
BACKGROUND: In this study, we sought to identify a criterion for the intermediate-risk grouping of patients with cervical cancer who exhibit any intermediate-risk factor after radical hysterectomy. METHODS: In total, 2158 patients with pathologically proven stage IB-IIA cervical cancer with any intermediate-risk factor after radical hysterectomy were randomly assigned to two groups, a development group and a validation group, at a ratio of 3 : 1 (1620 patients:538 patients). To predict recurrence, multivariate models were developed using the development group. The ability of the models to discriminate between groups was validated using the log-rank test and receiver operating characteristic (ROC) analysis. RESULTS: Four factors (histology, tumour size, deep stromal invasion (DSI), and lymphovascular space involvement (LVSI)) were significantly associated with disease recurrence and included in the models. Among the nine possible combinations of the four variables, models consisting of any two of the four intermediate-risk factors (tumour size ≥3 cm, DSI of the outer third of the cervix, LVSI, and adenocarcinoma or adenosquamous carcinoma histology) demonstrated the best performance for predicting recurrence. CONCLUSION: This study identified a 'four-factor model' in which the presence of any two factors may be useful for predicting recurrence in patients with cervical cancer treated with radical hysterectomy.