Importance of radiographic tumor regression during radiotherapy in squamous cell versus adenocarcinoma of the uterine cervix as assessed by MRI and cone beam CT

Abstract

The purpose of this study was to evaluate the kinetics of tumor regression in cervical squamous cell carcinoma (SCC) and adenocarcinoma (AC) using computed tomography (CT) and magnetic resonance imaging (MRI), and to correlate rate of regression and residual tumor with progression and survival. Thirty-two patients with stage IB2-IVA cervical cancer were randomly selected from an institutional database with a 2:1 ratio of SCC to AC. All available on-treatment weekly cone beam CT (CBCT) and pre- and post-external beam MRIs were utilized to generate largest two-dimensional area of tumor and tumor gross tumor volumes (GTV), respectively. Tumor volume regression velocity and percent residual tumor were correlated to 5-year progression-free survival (PFS) and disease-specific survival (DSS) using threshold regression modeling. Kaplan-Meier and Fine-Grey estimators were used for survival and cumulative incidence analysis, respectively. With a median follow-up of 2.9 years, 32 patients were included, 22 (69%) with SCC and 10 (31%) with AC. All received concurrent chemoradiation followed by brachytherapy. The 2-year cumulative incidence (CI) of local progression for both SCC and AC was 10%, and 2-year CI of distant progression was 9% vs. 57% (p = 0.02). Deaths occurred in 3/22 (14%) with SCC and 7/10 (70%) with AC, with a 2-year DSS of 90% versus 60%, respectively. Extent and rate of regression on CBCTs were not correlated with progression or survival; however, consistent rates of tumor regression for both SCC and AC. Thresholds of ≥20% residual disease on post-external beam pre-brachytherapy MRI and regression velocity ≤1.8%/day were associated with worse PFS and DSS. This study showed cervical AC is associated with higher rates of distant progression and worse overall survival than SCC. Cervical AC tends to have a higher initial and residual tumor burden. Our identified thresholds of ≥20% residual tumor and tumor regression of ≤1.8%/day may help identify cases warranting dose escalation.