Change in patterns of failure after image-guided brachytherapy for cervical cancer: analysis from the RetroEMBRACE study

Purpose Image guided adaptive brachytherapy (IGABT) for cervical cancer improves pelvic control and survival across all stages. Improvement in pelvic control is larger in advanced stages, but improvement in survival is similar across stages. This paper analyzes the patterns of failure in the RetroEMBRACE cohort to investigate this discrepancy. Methods and Materials 731 patients from 12 institutions treated with chemoradiation therapy and magnetic resonance imaging or computed tomography–based IGABT were evaluated. The pattern of failure at time of first relapse was analyzed. Results Three hundred twenty-five failures (single and synchronous) occurred in 222 of 731 patients (30%). Among the 325 failures, 9% were local and 6% regional. Pelvic (local or regional) failures made up 13%, paraaortic node (PAN) 9%, systemic 21%, and distant (systemic + PAN) 24%. Of the 222 patients with treatment failure, 21% had pelvic failure alone, 57% had distant failure alone, and 23% had both pelvic and distant failure. Of all failures that occurred, 40% to 50% occurred in the first year, with a further 20% to 30% occurring in the second year. Although local, regional, and PAN failure tended to plateau after year 3, systemic failure continued to occur up to year 10. Conclusions Implementation of IGABT has changed the patterns of relapse after chemoradiation therapy for cervical cancer. The predominant failure after IGABT is systemic, whereas the predominant failure with conventional brachytherapy is pelvic. Effective treatments to eradicate micrometastases in PAN and distant organs are needed in addition to IGABT and chemoradiation therapy to maximize local, regional, PAN, and systemic control and improve survival.