Clinical Outcomes for 3D High-Dose-Rate Intracavitary Brachytherapy With MRI-Based Planning for the Treatment of Cervical Cancer

Image-guided brachytherapy (IGBT) is increasingly being used for gynecologic cancers due to the ability to improve dose volume parameters and, in turn, decrease toxicities and improve clinical outcomes. MRI-based planning allows better delineation of soft tissue and residual disease. Our results here demonstrate the feasibility and outcomes of such image-guided brachytherapy. A retrospective review of patients with cervical cancer, FIGO stage IB1 to IIIB, was conducted. All patients were treated with external beam radiation therapy followed by intracavitary brachytherapy with a ring and tandem technique, consisting of 25-30 Gy in 5 fractions based on response. Three-dimensional planning (CT- or MR-based) was completed with MRI utilized for at least one fraction. The high-risk clinical target volume (HRCTV) and organs-at-risk were delineated following GEC-ESTRO guidelines. Point A dose was calculated for recording purposes but all optimization was done based on HRCTV volume coverage and critical organ dosimetry. Total doses were calculated into biologically equivalent doses of 2 Gy per fraction using the linear quadratic model. Planned total dose to the HRCTV was 75-85 Gy with constraints to the rectum, bladder and sigmoid consisting of D2 cc ≤70, ≤85, and ≤70 Gy, respectively. Eighty-two patients, treated from 2007-2012, were included for analysis. A majority of patients had FIGO stage IIB disease (67%) with a median age of 51.5 years and mean tumor size of 4.9 cm. The median HRCTV D90 was 82.8 Gy (range, 78.5 - 93.3 Gy). Point A dose showed greater variability with a median dose of 77.5 Gy (range, 61.8 - 93.5 Gy). Median equivalent 2 cc dose to bladder, rectum and sigmoid were 78.2, 56.7, and 66.4 Gy, respectively. With a median follow-up of 16 months (range, 2-65 months), the 2-year actuarial local control and Kaplan-Meier overall survival estimates were 90% and 82%, respectively. Thirteen patients developed recurrence: 5 local failures only, 6 distant failures only, and 2 simultaneous local and distant failures. Median time to local failure was 8.4 months. Local failure was significantly associated with adenocarcinoma histology (p = 0.003). The median HRCTV D90 and point A dose for patients who failed versus controlled was 84.4 vs 82.8 Gy and 77.4 vs 77.5 Gy, respectively. RTOG grade 3 or greater morbidities were seen in two patients, consisting of fistula formation at 12 and 36 months with a 2-year actuarial rate of 2%. No dosimetric correlate was established for these patients. This constitutes one of the largest reported single-institution series on MRI-based IGBT in North America, demonstrating excellent local control with acceptable morbidity. Continued follow-up with a larger cohort is planned to further corroborate these results.