Simulation-Free Hippocampal-Avoidance Whole Brain Radiotherapy Using Diagnostic MRI-Based and Cone Beam Computed Tomography-Guided On-Table Adaptive Planning in a Novel Ring Gantry Radiotherapy Device

Overview

Hippocampal-avoidance whole brain radiation therapy (HA-WBRT) limits radiation dose to the hippocampal-avoidance region while still delivering therapeutic doses of radiation to the whole brain. When used in addition to prophylactic memantine, this technique has been shown to better preserve cognitive function in patients with brain metastases outside of the hippocampal-avoidance region with no difference in intracranial progression-free and overall survival. However, HA-WBRT requires considerably longer planning time when compared to conventional WBRT (5-10 business days, compared to next-day), and studies have shown that brain metastases can grow in as rapidly as one week. A proposed solution for quicker initiation of HA-WBRT is the use of simulation-free radiation treatment planning, in which pre-existing diagnostic images are used to generate the radiation treatment plan (as opposed to acquiring planning-specific image sets). This will be paired with the use of artificial intelligence (AI)-assisted semi-automated planning using the FDA-approved treatment planning system called Ethos Therapy. The investigators have developed an institutional HA-WBRT auto-planning template, which has been retrospectively validated for the creation of plans that are compliant with the gold standard NRG Oncology CC001 clinical trial and are dosimetrically comparable to traditional HA-WBRT plans. Semi-automated plans will be constructed using diagnostic imaging, which will be refined as needed (adjustments for difference in gross head positioning between diagnostic imaging and radiation treatment positioning, etc.) while the patient is on the treatment table at fraction one using adaptive radiation planning. Adaptive radiotherapy is standard-of-care practice for other disease sites. The purpose of this study is to demonstrate the feasibility and safety of a simulation-free workflow for HA-WBRT that is AI-assisted and semi-automated.