RESUMO
OBJECTIVE: The objective of this study was to develop a predictive model for parametrial involvement (PMI) and to identify low-risk group of PMI in early stages of cervical cancer based on preoperative magnetic resonance imaging (MRI) parameters. METHODS: We retrospectively analyzed patients with stages IB1 to IIA2 cervical cancer (N = 1347) who underwent type C radical hysterectomy between 2005 and 2012. Clinical records, preoperative MRI, and its association with pathological data were reviewed. A predictive model for PMI was developed using preoperative MRI parameters for the estimation of its performance. RESULTS: Of 1347 patients, 138 (10.2%) had pathological PMI (p-PMI). Multivariate analysis identified the maximal tumor diameter (odds ratio, 2.0; 95% confidence interval, 1.23-3.40; P < 0.001) and PMI (odds ratio, 7.0; 95% confidence interval, 4.49-11.02; P < 0.001) on preoperative MRI (m-PMI) as independent predictive factors for p-PMI. The rate of p-PMI was 1.3% for low-risk patients identified by the current model (maximal tumor diameter ≤2.5 cm and no indication of PMI, n = 448). The 5-year progression-free survival rate was significantly greater (96.7%) in low-risk patients than in those with a maximal tumor diameter greater than 2.5 cm and/or indication of m-PMI (90.8%, P = 0.004). CONCLUSIONS: A predictive model for p-PMI was developed in which p-PMI exclusion was set as a maximal tumor diameter less than or equal to 2.5 cm and no indication of m-PMI. Patients with a low risk of m-PMI could be identified so that less radical surgery for stages IB1 to IIA2 cervical cancer could be offered.
