Temporospatial Feathering of Hot Spots for Computed Tomography-Guided Stereotactic Adaptive Radiotherapy (CT-STAR) for the Ultra-Central Thorax

Abstract

SBRT to the ultra-central thorax is limited by potential toxicity. It has been demonstrated that exposing proximal bronchial tree or pulmonary arteries to high dose per fraction (fx) treatment may induce bronchopulmonary hemorrhage, amongst other serious complications. Online adaptive radiotherapy is a technique that adjusts a treatment plan to the anatomy-of-the-day and benefits have been demonstrated in ultra-central thoracic disease. In addition, feathering is a treatment planning technique that generates several plans to avoid consistent organ-at-risk (OAR) doses throughout treatment. With daily adaptation, it may be possible to adjust the position of a hot spot (>120% prescription (Rx)) within the tumor each fx (temporospatial feathering) while respecting hard OAR constraints. We investigated the feasibility and plan quality of using CBCT-guided stereotactic adaptive radiotherapy (CT-STAR) for ultra-central lung tumors with hotspot temporospatial feathering. Seven patients with ultra-central thoracic disease (6 patients with parenchymal tumors in contact with the trachea, proximal bronchial tree, great vessels, esophagus, or heart; 1 patient with a subcarinal lymph node) receiving standard of care radiotherapy were enrolled on an imaging study. An in-silico planning study first generated an SBRT plan (in silico Rx: 55 Gy in 5fx) that used a GTV_OPT (GTV minus OAR plus a safety margin) to optimize the location of the plan hotspot. Five spherical boost structures were manually created inside of the GTV_OPT structure. The same planning template was used except the boost structures were iteratively used in plan optimization instead of the GTV_OPT structure, to simulate the five CT-STAR fx hotspot temporospatially feathering. The five-plan composite was compared to the initial plan. All plans generated met strict OAR constraints. Table 1 shows the mean difference in PTV, GTV, and OARs percent coverage by various isodose levels. Feathering the hotspot had negligible impact on target coverage by 50 Gy and 55 Gy isodose lines as well as OAR doses compared to the base SBRT plan. The feathered plan sum resulted in 14.7% increase in V66 Gy of the GTV. One patient saw a decrease in V66 Gy coverage to all target structures, though V50 Gy and V55 Gy were not affected. We demonstrated the feasibility and utility of temporospatially adapting the hotspot for central lung SBRT, which safely increases the amount of tumor receiving more than 120% Rx.