RESUMO
PURPOSE: To identify if baseline patient or magnetic resonance imaging (MRI) features can predict which women are at risk for inadequate tumor coverage with only intracavitary tandem and ovoid (T + O) brachytherapy and to correlate tumor coverage with clinical outcomes. METHODS AND MATERIALS: We performed a retrospective study of 50 women with cervical cancer treated with chemoradiation at a single institution between January 2014 and December 2015. All patients had a 3T-MRI performed at baseline (MRI1) and at the completion of external beam radiation therapy (MRI2). Gross tumor volume initial (GTV-Tinit) was measured on MRI1 and high-risk clinical tissue volume (CTVHR) on MRI2. CTVHR extending beyond point A was classified as too large for adequate coverage with T + O and requiring interstitial needles. Multivariate analysis was performed to determine predictive factors of inadequate coverage. Kaplan-Meier and Cox Regression were performed to correlate inadequate coverage with outcomes. RESULTS: Mean patient age was 49.2 ± 13.2 years, and 84% had Federation of Gynecology and Obstetrics IIB/IIIB disease. Forty-two percent of women were estimated to have inadequate tumor coverage with T + O brachytherapy. The GTV-Tinit volume and dimensions (superior-inferior, left-right, anterior-posterior) on MRI1 were all important predictive factors of inadequate coverage on multivariate analysis. Receiver operating characteristics curves identified optimal thresholds of superior-inferior ≥ 4.5 cm (area under the curve [AUC] = 0.718), left-right ≥ 4.5 cm (AUC = 0.745), anterior-posterior ≥ 5.0 cm (AUC = 0.767), and GTV-Tinit ≥ 85 cm3 (AUC = 0.842). Patients with inadequate coverage had worse clinical outcomes. CONCLUSIONS: Baseline MRI tumor size may predict inadequate CTVHR coverage at the time of brachytherapy (i.e., the need for interstitial needles). This may help identify a subset of women requiring early referral to adequately resourced centers to improve clinical outcomes.