2044 - A Pilot Study for CT-Based Online Adaptive Radiotherapy in Patients with Newly Diagnosed High-Grade Glioma

Purpose/Objective(s):

Postoperative high-grade glioma (HGG) target volumes are defined on MRI/CT at simulation and remain fixed during 6 weeks of chemoradiation (CRT) despite morphologic changes in the resection cavity and surrounding brain parenchyma. MRI-guided adaptive radiotherapy (ART) enables daily adaptation but increases cost, treatment time, and immobilization duration. Our prior work shows that most anatomic change occurs within 30 days and that strategically timed MRIs during radiation therapy (RT) capture these changes. Thus, we piloted a more widely deployable approach using interval MRI with cone-beam CT (CBCT) on a Varian Ethos LINAC, prospectively evaluating feasibility and anatomic/dosimetric variability of weekly MRI-informed ART. We sought to calculate the difference in the volume of brain parenchyma receiving 20 Gray (Gy) and 60 Gy, known as V20 and V60, respectively, between the delivered plan and the simulated weekly adapted RT plan. We hypothesize that ART will reduce RT dose to nearby brain tissue and the hippocampi, known to be associated with dose-dependent white matter changes and neurocognitive decline.

Materials/Methods:

In this single institution prospective study, 20 patients with resected HGG received CRT on Varian Ethos with daily CBCT. Weekly MRIs were obtained during CRT. Targets were retrospectively contoured on weekly MRI, and new weekly plans were generated for dosimetric comparison to the delivered plan. We report simulated weekly ART for 4 consecutive patients - weekly plans were summed and compared with the delivered plan.

Results:

All patients completed their MRIs as scheduled, with the exception of 1 patient who missed a single MRI. Median follow-up was 15.8 months. One of the four patients had disease progression distantly in the contralateral hemisphere. The median absolute difference in the volume receiving 20 Gray (Gy), known as the V20, was -10% (range: -0.2 – -15.0%) of brain volume between the delivered and adapted plans, corresponding with a median reduction of 125.6 cm³ (range: 2.8 – 205.8 cm³). The median absolute difference in the V60, was -4.1% (range: -1.7– -5.5%) of brain volume between the delivered and adapted plans, corresponding with a median reduction of 52.3 cm³ (range: 20.9 – 73.8 cm³). The median dose reduced to the ipsilateral and contralateral hippocampus between the delivered and adapted plans was 4.70 Gy (range: -1.18 Gy – 22.72 Gy) and 0.9 Gy (range: -0.65 Gy – 0.46 Gy), respectively.

Conclusion:

Simulated ART resulted in median reduction in V20 and V60 of 10% and 4.1%, respectively, and permitted a 4.70 Gy median dose reduction (range: -1.18 Gy – 22.72 Gy) to the ipsilateral hippocampus. Adequate follow-up will clarify oncologic outcomes. This cohort underscores the need for prospective studies exploring strategic adaptation of RT plans for patients with HGG to mitigate neurocognitive decline whilst implementing workflow that is resource-conscious and widely feasible, which will pave the way for a multi-institutional randomized study.