Risk prediction model for disease-free survival in women with early-stage cervical cancers following postoperative (chemo)radiotherapy

Abstract

Purpose: To investigate disease-free survival (DFS) and prognostic factors following the administration of postoperative (chemo)radiotherapy in patients with early-stage cervical cancers. Methods: The medical records of 1,069 patients from 10 participating institutions were reviewed. Statistically and clinically established factors were considered as candidates for constructing the prediction model. This model was validated, using bootstrapping to correct for optimistic bias. Results: The 5-year DFS rate was 81.1%, with a median follow-up period of 59.6 months. The statistically significant prognostic factors were as follows: pelvic lymph node metastasis, histologic type, parametrial invasion, lymphovascular space invasion, and tumor size. The nomogram for DFS was constructed, and it demonstrated a good discrimination performance, with an internally validated concordance index of 0.72. Conclusions: Our predictive model exhibited accurate predictions and may be useful in designing clinical trials to study if further chemotherapy can reduce the recurrence of disease in high-risk patients.