LBA05 - MR-Linac On-Line Weekly Adaptive Radiotherapy for High Grade Glioma (HGG): Results from the UNITED Single Arm Phase II Trial

J. Detsky¹, A. W. Chan¹, D. M. Palhares¹, J. M. Hudson¹, J. Stewart¹, H. Chen¹, S. Das², N. Lipsman³, M. J. Lim-Fat⁴, J. Perry⁴, M. E. Ruschin¹, S. D. Myrehaug¹, H. Soliman¹, C. L. Tseng⁵, and A. Sahgal¹; ¹Department of Radiation Oncology, Odette Cancer Centre, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada, ²Division of Neurosurgery, St. Michael’s Hospital, Unity Health Toronto, University of Toronto, Toronto, ON, Canada, ³Division of Neurosurgery, University of Toronto, Toronto, ON, Canada, ⁴Department of Neurooncology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada, ⁵Department of Radiation Oncology, Sunnybrook Health Sciences Centre, University of Toronto, Toronto, ON, Canada

Purpose/Objective(s): To assess the feasibility and safety of weekly on-line MR-Linac (MRL) adaptive radiotherapy with concurrent temozolomide (chemoRT), with a reduced 5 mm clinical target volume (CTV) margin, for patients with HGG (grade 4 astrocytoma or IDH-wildtype glioblastoma). Materials/

Methods: In this single arm Phase 2 trial (NCT04726397), patients with newly diagnosed HGG planned for chemoRT with either 60 Gy/30 fractions (long course) or 40 Gy/15 fractions (short course) were eligible. Gross tumor volume (GTV) was defined as the surgical cavity and residual tumor; the CTV was a 5 mm uniform expansion plus involved T2-hyperintense FLAIR signal at the discretion of the oncologist; PTV was 3 mm. All patients were treated using a 1.5T integrated MRL. At fraction 1, and each subsequent fifth fraction, an on-line contrast-enhanced MR was acquired, volumes were re-contoured and treatment was re-planned. For the remaining fractions an on-line workflow used non-enhanced MR images to account for positional shifts. The primary endpoint was the risk of a marginal failure (MF) defined as 20-80% of the recurrent GTV (at the time of failure) within the 95% treatment isodose line and/or within a standard 15 mm CTV envelope. Using a historical 11.1% MF event risk from previously reported series using non-adaptive chemoRT with standard (15-20 mm) CTV margins, non-inferiority would be demonstrated if 13 or fewer MF events were observed within a sample size of 98, assuming a 11.9% non-inferiority margin (95% upper boundary of historical outcomes). Secondary endpoints included progression-free survival (PFS) and overall survival (OS) according to treatment schedule (long versus short course).
Results: A total of 108 patients were consented between April 2021 and May 2023 of which 98 completed the treatment protocol (59 long course and 39 short course). All tumors were IDH-wt, and 52/98 (53%) were MGMT methylated (MGMT-m), 41/98 (42%) unmethylated (MGMT-um) and 5/98 (5%) indeterminate. Median follow up was 23.3 months (mo). MF events were observed in 4/98 (4.1%, 95% CI: 1.6-10%) patients, establishing non-inferiority (p<0.001); the most common pattern of failure was central (52%). Median PFS was 11.6 mo for a long course (14.1 mo for MGMT-m and 8.5 mo for MGMT-um), and 6.8 mo for those treated with a short course (9.4 mo for MGMT-m and 5.1 mo for MGMT-um). Median OS was 18.5 mo after a long course (31.9 mo for MGMT-m and 13.0 mo for MGMT-um) and 10.6 mo after short course (15.3 mo for MGMT-m and 8.9 mo for MGMT-um).
Conclusion: We present the first trial evaluating on-line MRL weekly adaptive chemoRT for HGG with a limited CTV. Safety and feasibility was demonstrated with a low risk of MF (4%) without compromising PFS or OS. Further trials are required to test whether the reduction in irradiated normal brain tissue using this approach results in neurocognitive and quality of life benefits.