Parametrial Boost (PMB) in the Era of Image-Guided Brachytherapy (IGBT) for Cervical Cancer

Abstract

Historically, PMB was standard practice in locally advanced cervical cancers to provide sufficient dose to involved parametria. Modern 3D imaging, as well as high-dose rate (HDR) and interstitial (IS) brachytherapy techniques have improved the delineation and coverage of a high-risk clinical target volume (HRCTV) while allowing for higher dose delivery. PMB has become increasingly controversial in the era of IGBT. Due to the lack of data on safety of omitting PMB, patterns of failure with and without use of PMB were explored. Women with locally advanced (FIGO Stage IIB-IVA or with PM invasion identified on imaging) cervical cancer diagnosed between 2001-2016 were identified. All patients received external beam radiotherapy (EBRT) and brachytherapy boost (BT) using either low-dose rate (LDR) or HDR sources (prescribed to point A and HRCTV, respectively). HRCTV was contoured using MRI guidance (either diagnostic or with apparatus in place). PMB was delivered to achieve a cumulative dose of 56-60 Gy to the lateral parametria. Demographic, treatment era (pre and post 2009), pre-treatment imaging modality, FIGO stage, lymph node status, PMB, BT modality (IS vs. tandem & ovoid, T&O), dose rate, and patterns of failure were collected. Cramer’s V and chi-squared tests were used to assess strength of association between variables. The log-rank test was used for univariate analysis (UVA) and a Cox proportional hazards model was created for multivariable analysis (MVA). Overall survival (OS) was estimated using Kaplan-Meier analysis A total of 71 women were identified with median follow up of 25 months. 36 women received PMB. Median doses were as follows: EBRT 45 Gy, PMB 4.5 Gy, HDR BT 29.5 Gy, and LDR 40 Gy. 37 women received HDR and 34 received LDR BT. Six patients who did not receive PMB had PM involvement only on imaging. PMB patients were less likely to have received diagnostic MRI, PET, or IS and more likely to receive LDR. The use of PET, dose rate, and treatment era were highly correlated with one another (Cramer’s V 0.7-0.9, p<0.001). 2-year OS was 74%. On UVA, treatment year, use of PET, use of MRI, and dose rate were associated with OS (all p<0.05). In an MVA including PMB, age, race, nodal status, FIGO stage, and PET, para-aortic nodal involvement (HR= 4.33, 95% CI 1.01-10.7, p=0.049) and use of PET (HR=0.22, 95% CI 0.1-0.7, p=0.007) were associated with OS. PMB was associated with higher crude rates of local (22% vs. 6%, p=0.049), locoregional (30% vs. 6%, p=0.01), and distant failure (28% vs. 9%, p=0.04). In this retrospective study, the omission of PMB did not compromise local control or impact OS in the setting of IGBT. At our institution, PMB treatment has gradually been eliminated in cervical cancer treatment. Adaptive IGBT provides accurate identification and dose shaping of HRCTV previously unachievable in the 2D era, allowing for targeting of gross parametrial involvement.