Distant Metastases in Locally Advanced Cervical Cancer Pattern of Relapse and Prognostic Factors: Early Results from the Embrace Study

The aim of this study is to characterize distant metastases and prognostic factors in locally advanced cervical cancer (LACC) treated with definitive chemoradiation therapy and image guided adaptive brachytherapy within the EMBRACE study. All patients enrolled in the EMBRACE study and diagnosed with LACC before January 2013 were included in this analysis. Distant metastases were defined as any extrapelvic relapse in lymph nodes, organs, or both. Uni- and multivariate analysis of prognostic factors for distant metastases was performed with actuarial probabilities based on Kaplan Meier statistics and the Cox regression model with hazard ratios (HRs). A total of 753 patients (out of 793 enrolled) were included and had a median follow-up of 29 months. One hundred thirty-three patients had distant metastases as the first event, and 80% within the first 2 years. Sixty-three percent had nodal metastases (including 44% with para-aortic involvement, 18% with para-aortic nodes only), 48% had organ metastases (37% lungs, 10% bones, 9% liver, 2% brain, and 6% other organs), and 9% had peritoneal carcinomatosis. For 9 patients the localization of metastases was not documented, for 51 patients more than one metastatic site at the time of recurrence was documented, and for 73 patients only 1 metastatic site at the time of recurrence was documented. At 3 and 5 years, the distant metastases-free interval was 79% and 77%, respectively. In univariate analysis, nodal status at diagnosis (P ≤ .001), tumor width (P ≤ .001), hemoglobin level <100 g/dL at diagnosis (P ≤ .003), uterine infiltration (P ≤ .023), histology (P ≤ .038), and degree of differentiation (P ≤ .038) were significant prognostic factors. In multivariate analysis, regional node involvement (P ≤ .001, HR = 2.6), adenocarcinoma (P ≤ .005, HR = 2.1), tumor volume >60 cm3 (P ≤ .006, HR = 1.8), and tumor differentiation (low/medium, P ≤ .005/.014, HR = 2.6/2.4) remained significant prognostic factors. Patients with ≥3 of these factors present with a 3-year metastases-free interval of 52% compared to 81% in patients with <3 factors (P ≤ .001). The distribution of distant metastases with comparable numbers for nodal and organ sites and the identification of prognostic factors allows for a better understanding of the course of distant metastases in LACC. This information provides the basis for adapting treatment strategies to minimize distant metastases, that is, for designing future trials aiming at better control of nodal and organ metastases.