Can Volume-Based MRI-Guided Planning for Interstitial Gynecologic Brachytherapy Overcome Clinicopathologic Risk Factors for Fistula Formation?

Abstract

Tumor and patient factors requiring an interstitial (IS) approach for gynecologic tract brachytherapy (BT) may be associated with high morbidity despite the conformality maximizing effect of ISBT. We reviewed outcomes from a diverse, high-volume center to evaluate the association of clinical features and planning parameters with toxicity, with the hypothesis that MRI-guided volume-based planning mitigates clinicopathologic risk factors associated with treatment related fistula (TRF) after ISBT. We retrospectively reviewed patients treated with Syed-based ISBT at a single institution from 2014-2019. Fisher's exact was used for group comparison with p<0.05. Kaplan-Meier method was used to estimate local control (LC), overall survival (OS), and fistula-free survival (FFS). Cox regression was used for univariate (UVA) and multivariate analysis (MVA) to estimate hazard ratios (HR). Collinearity was assessed using variable inflation factor and Pearson's correlation. A total of 54 patients (median age 58 [IQR 46-67], 54% Caucasian, 39% African American, 15% Hispanic) treated with ISBT (89% MRI-guided) with median follow-up 32 months had initial (89%) or recurrent (11%) disease from cervical (70%), vaginal (15%), urethral (6%), and other cancers. At presentation 5.5% had clinically evident fistula. There was radiographic rectal and bladder involvement in 22.0% and 19.2%, respectively. 89% received concurrent chemotherapy and 98% received EBRT prior to ISBT with a cumulative median high-risk CTV (HRCTV) dose 80.8 Gy (IQR 76.8-84.3), bladder D2cc 84.3 Gy (IQR 75.8-89.0), and rectal D2cc 73.8 Gy (IQR 67.7-80.0 Gy). All 7 patients who developed a TRF (rectovaginal in 6/7 and vesicovaginal in 5/7) were former or current smokers, had primary cervical cancer, and received chemotherapy. Management included surgical intervention in 6 (85.7%). LC, OS, and FFS at 2-years was 89.3% (95% CI 75.9-95.4%), 87.8% (95% CI 66.9-89.8%), and 85.2% (95% CI 69.9-93.1%), respectively. On UVA, current smoking (HR 4.60, 95% CI 1.02-20.74), BT bladder D2cc (HR 1.63, 95% CI 1.07-2.47) and BT rectal Dmax (HR 1.30, 95% CI 1.07-1.58) predicted for increased risk of TRF. Most factors were non-significant including Charlson Comorbidity Index, age, BMI, surgical history, extent of vaginal involvement, race, HRCTV volume and number of needles used. No patients with radiographic bladder or rectal invasion developed TRF. On MVA, only smoking status (HR 14.05, 95% CI 1.48-133.1) remained significant. 0% of never (0/26), 20% of former (3/15) and 31% of current (4/13) smokers developed fistulas from toxicities (p<0.05). In patients with locally advanced cancers of the gynecologic tract treated with MRI-guided volume-based ISBT, smoking was the only factor predictive of TRF formation, occurring primarily in cervical cancer patients. This highlights the importance of smoking cessation during treatment and appropriately counseling patients at high risk for this morbidity.