PO19: Direction Modulated Brachytherapy Tandem Model Applicators for Treatment Planning of Multi-Institutional Cervical Cancer Cases.

The rise of image guided adaptive brachytherapy (IGABT) in combination with advances in inverse optimization tools has led to improved plan quality via utilization of rich 3D patient-specific anatomical information. The necessity to generate directional "anisotropic" radiation beam for better conformance of dose distributions to target volumes while integrating seamlessly into existing IGABT clinical workflows is apparent. The novel MR-compatible Direction Modulated Brachytherapy (DMBT) tandem applicator accomplishes this by using a six-peripheral-source-channel design. In this study, the authors sought to validate the implementation of 9 unique DMBT tandem models of varying physical dimensions integrated for the first time into the BrachyVision® (v16.1) treatment planning system (BV-TPS). Additionally, the authors quantify the expected improvement in plan quality via organs at risk (OAR) dose reductions from a large cohort of patient cases collected from three institutions. A multi-institutional cohort of 110 retrospective clinical HDR brachytherapy plans (from 42 patients) were re-planned with BV-TPS, using the latest VEGO® inverse optimization algorithm, with dose heterogeneity accounted for through the AcurosBV® model-based dose calculation algorithm. Plans consisted of both intracavitary (77 plans) and interstitial (33 plans) cases with an average prescription dose of 607±113 cGy. The average high-risk clinical target volume (CTV HR) was 26.96±14.95 [range 6.70-69.58] cm3. The 9 DMBT tandem models were designed with a MR-compatible tungsten alloy with thicknesses comparable to those tandems used clinically and ranged between 4-8 mm. During re-planning, the conventional tandems were replaced by one of the 9 DMBT tandem models while leaving ovoids/rings and needles (if present) in place. The two-step inverse optimization was performed such that the lowest possible OAR D2cc doses could be achieved while 1) keeping equivalent target coverage and, at the same time, 2) maintaining the general pear-shape dose distribution of the original plans. For all plans, this process was repeated using each of the 9 DMBT tandem models for a total re-planning of 990 cases. Noteworthy improvements in plan quality were achieved by all 9 DMBT tandem models, which are presented in Figure 1A. In general, irrespective of the DMBT model, about ∼50 cGy reduction in D2cc across all OARs appear feasible. There is also a general trend of D2cc reductions' magnitude getting smaller as the CTV HR volume increased (Figure 1B), as expected, due to the loss of intensity modulation capacity with increasing distance away from the tandem. Additionally, D2cc reductions in terms of EQD2 [Gy] were calculated assuming each re-plan was delivered throughout the course of treatment (Figure 1C), which includes the external beam radiotherapy dose of 45 Gy, and showed significant reductions of -6.29±4.38 Gy, -3.80±2.06 Gy, and -4.86±3.02 Gy for bladder, rectum, and sigmoid, respectively, for the DMBT model #9. These reductions were achieved with an average net increase in total dwell times of 87.99±71.41 seconds, i.e., typically adding 1-2 minutes of treatment time per fraction. We have successfully incorporated 9 DMBT tandem models into a commercial TPS and re-planned 110 cases, to a total of 990 plans. All 9 DMBT tandem models were each able to generate notable D2cc reductions to OARs (∼50 cGy), without compromising target coverage, across plans from multiple institutions with various clinical/optimization practices. The results indicate both a promising impact and smooth integration of DMBT tandem technology into modern clinical IGABT workflow. [ABSTRACT FROM AUTHOR]