RESUMO
OBJECTIVES: To evaluate the performances of systematic posttreatment pelvic magnetic resonance imaging (PPMRI) in predicting prognosis of patients treated with chemoradiation therapy (CRT) for locally advanced cervical cancer (LACC). MATERIALS AND METHODS: Multi-institutional data from 216 patients presenting FIGO IB2-IIB cervical cancer for which PPMRI was performed following CRT were retrospectively reviewed. Incomplete response was defined as the identification of persistent lesion on PPMRI. Primary endpoints were patients' 5-year recurrence free (RFS) and overall (OS) survivals. Secondary endpoint was the identification of residual histologic disease on hysterectomy specimens when completion surgery was performed. RESULTS: PPMRI identified an incomplete response in 102 (47.2%) cases. A 70% or more reduction in tumor size on PPMRI was identified as the best predictive cut-off for recurrence (37.7% sensitivity and 78.7% specificity) and death (50% sensitivity and 77.9% specificity) with significant impact on those risks (HRa: 0.42; 95%CI: 0.23-0.77 and HRa: 0.18; 95%CI: 0.06-0.50, respectively). Completion hysterectomy was performed in 117 (54.4%) cases, with histologic residual disease in 55 (47.4%). PPMRI demonstrated 74.5% sensitivity and 50.8% specificity in predicting residual disease. Although survival of patients with complete response at PPMRI was not impacted by completion hysterectomy, it significantly increased 5-year RFS and OS of those with incomplete response: 38.7% vs. 65.3% (p < 0.001) and 63% vs. 82.9% (p = 0.038), respectively. CONCLUSION: A 70% or more reduction of in tumor size on PPMRI following CRT in patients with LACC is predictive of RFS and OS. PPMRI could help triaging patients who could benefit from completion hysterectomy.