Nomogram prediction of overall survival after curative irradiation for uterine cervical cancer.

Abstract

The purpose of this study was to develop a nomogram capable of predicting the probability of 5-year survival after radical radiotherapy (RT) without chemotherapy for uterine cervical cancer. We retrospectively analyzed 549 patients that underwent radical RT for uterine cervical cancer between March 1994 and April 2002 at our institution. Multivariate analysis using Cox proportional hazards regression was performed and this Cox model was used as the basis for the devised nomogram. The model was internally validated for discrimination and calibration by bootstrap resampling. By multivariate regression analysis, the model showed that age, hemoglobin level before RT, Fédération Internationale de Gynécologie Obstétrique (FIGO) stage, maximal tumor diameter, lymph node status, and RT dose at Point A significantly predicted overall survival. The survival prediction model demonstrated good calibration and discrimination. The bootstrap-corrected concordance index was 0.67. The predictive ability of the nomogram proved to be superior to FIGO stage (p=0.01). The devised nomogram offers a significantly better level of discrimination than the FIGO staging system. In particular, it improves predictions of survival probability and could be useful for counseling patients, choosing treatment modalities and schedules, and designing clinical trials. However, before this nomogram is used clinically, it should be externally validated.