3746 - High-Risk CTV Geometric Features and Spatial Relationships to Organs at Risk as Determinants of Dosimetric Benefit with Combined Intracavitary/Interstitial Brachytherapy in Cervical Cancer

Purpose/Objective(s):

Aside from volume of high-risk CTV (HRCTV), there is little objective data to guide clinicians in selection of patients that will benefit from use of combined intracavitary/interstitial (ICIS) versus intracavitary (IC) applicators during cervical cancer brachytherapy. We hypothesized that geometric features and spatial relationships between HRCTV and surrounding organs at risk (OARs) may better predict dosimetric benefit from ICIS applicators than the HRCTV volume =30 cc criterion derived from retroEMBRACE data. We further sought to identify an optimal geometric parameter associated with clinically meaningful reductions in OAR D2cc without compromising HRCTV D90%.

Materials/Methods:

Complete post-implant first-fraction DICOM datasets from 41 patients with cervical cancer treated with ICIS brachytherapy at our academic medical center between 2015 and 2025 were extracted from Oncentra Brachy TPS. An in-house Python-based geometric analysis code was developed to compute on a slice-by-slice basis: (1) the Euclidean distance between the uterine tandem axis and the most distant HRCTV boundary (d_HRCTV), and (2) the shortest Euclidean distance between the tandem axis and each organ at risk (d_OAR) on corresponding axial slices for each of bladder, rectum, and sigmoid colon. Maximum organ-specific geometric ratios were calculated as d_HRCTV/d_OAR.

To evaluate the dosimetric implications of d_HRCTV/d_OAR, plans were re-optimized after virtually removing interstitial needles, adjusting dwell times to achieve the lowest feasible OAR D2cc values while maintaining HRCTV D₉₀ coverage, thus creating an ICIS No Needles Optimized (ICIS NNO) cohort. D2cc for rectum, bladder, and sigmoid were plotted versus d_HRCTV/d_OAR and best fit cubic regression lines were generated by SAS V9.4. Mann–Whitney and two-sample t tests were applied to compare dosimetric differences and to identify the organ-specific geometric ratio (d_HRCTV/d_OAR) threshold above which virtual needle removal was associated with statistically significant increases in OAR D2cc.

Results:

Threshold cutoffs were 0.7 for rectum and 1.6 for bladder. Mean D2cc increase above vs. below these cutoffs was 26.8 cGy vs. 45.2 cGy (p = 0.001) for rectum, and 53.1 cGy vs. 152.3 cGy (p = 0.0002) for bladder. No significant cutoff was observed for sigmoid colon. No significant association was identified between the HRCTV volume =30 cc criterion and the geometric ratio cutoffs. Overall, 29% of patients who demonstrated greater D2cc reduction with ICIS had HRCTV volumes <30 cc. Additionally, no significant correlation was observed between HRCTV volume and any organ-specific geometric ratio.

Conclusion:

Quantitative HRCTV geometric and spatial relationships to OARs independently predict dosimetric benefit from combined ICIS applicators and may improve patient selection beyond HRCTV volume criteria alone.