Abstract

Hospital records of 4490 patients treated for International Federation of Gynecology and Obstetrics (FIGO) stage IB, IIA, or IIB carcinoma of the cervix between 1960 and 2001 at 1 institution were reviewed. Outcomes were estimated using the Kaplan-Meier method and compared using the log-rank method. A proportional hazards regression model was used to evaluate the relative importance of predictive factors. The rates of disease-specific survival and pelvic disease control were strongly correlated with tumor diameter, FIGO stage, histological subtype, and clinical node status. Regression analysis demonstrated that a diameter of greater than 4 cm, a diameter of greater than 6 cm, FIGO stage II (vs IB), the presence and level of lymph node involvement, and histological subtype were all highly significant independent predictors of poor disease-specific survival. Intermediate tumor-diameter categories (>5 cm or >7 cm) and FIGO stage IIB (vs IB or IIA) did not contribute significant additional information to the model. Only a tumor diameter of greater than 4 cm, a diameter of greater than 6 cm, the presence of lymph node involvement, and histological subtype were independent predictors of pelvic disease control. On the basis of these results, we propose dividing each of the FIGO categories IB, IIA, and IIB into 3 groups according to clinical tumor diameter: (1) less than or equal to 4 cm, (2) 4.1 to 6 cm, and (3) greater than 6 cm. The proposed modified system would provide more accurate prognostic information, facilitate comparisons, and maintain continuity with the current staging system.