Author: Drexell Hunter Boggs, Carlos E. Cardenas, Allison Dalton, John B Fiveash, Joel A. Pogue, Richard A. Popple, Farnaz Rahim Li 👨🔬
Affiliation: The University of Alabama at Birmingham, University of Alabama at Birmingham 🌍
Purpose: External-beam Accelerated Partial Breast Irradiation (APBI) using stereotactic-body radiotherapy (SBRT) is increasingly adopted as an alternative to whole-breast radiation, offering targeted treatment with a favorable side effect profile compared to whole-breast radiation. However, widespread adoption is limited by SBRT's complexity and the need for specialized planning expertise. Automation through knowledge-based planning (KBP) can standardize and improve plan quality, supporting broader clinical use. This project aims to develop and evaluate automated planning workflows for APBI to enhance treatment efficiency and consistency.
Methods: Sixty-two APBI patients treated with SBRT at our institution were divided into two cohorts. The first cohort included two ipsilateral training templates (24-cases each) and two test cohorts (7-cases each) for left and right-sided targets. The second cohort combined left and right-sided cases into a 48-patient training template with a 14-patient test-set. All plans using the KBP model maintained clinical field geometry and structure sets for dose optimization. Structures included ipsilateral/contralateral breast, lungs, heart, ribs, and skin. RapidPlan models were designed using institutional practice guidelines. Subsequently, if heart and/or lung DVH goals were not met after initial RapidPlan optimization, a 2nd optimization step (initialized at MR3) is employed to improve OAR sparing. Plans were normalized (PTV 98% receives 100% Rx) for 30Gy in 5-fractions.
Results: KBP templates successfully designed APBI RapidPlans without intervention, achieving comparable dose coverage and OAR sparing, and in some cases, outperforming clinically-optimized plans. No significant differences were observed in CI, GI, or dose-volume metrics for the heart (V1.5Gy), lungs (V9Gy), and ribs/skin (Dmax). There was statistical improvement for ipsilateral-breast dose (V30Gy and V15Gy, favoring KBP-plans) but these were not clinically-meaningful.
Conclusion: Automation with KBP models ensures consistent, high-quality APBI SBRT plans, eliminating the need for manual optimization. Its ability to achieve clinically comparable dose-volume metrics highlights its potential for widespread adoption in routine practice.