Initial report of memory avoidance whole brain radiotherapy to treat brain metastases: A prospective phase 2 trial.

Abstract

2039 Background: A common approach for patients with extensive brain metastases requiring radiation is hippocampal avoidance whole brain radiotherapy (HA-WBRT) prescribed with memantine; this was proven to be efficacious based on NRG CC001. However, a subset of patients who receive HA-WBRT with memantine still experience cognitive decline. Other brain structures with important roles in memory and cognition include the corpus callosum, fornix, amygdala, hypothalamus, and pituitary; these structures all have a low propensity for brain metastases and therefore can be safely spared in a radiotherapy plan without increasing the risk of relapse. A subset of patients enrolled on a Phase 2 Randomized Controlled Trial ( NCT05503251 ) received an advanced “memory-avoidance WBRT (MA-WBRT) approach that spared these substructures in addition to the hippocampus, with a primary endpoint of improved cognition compared to a historical control (NRG CC001). Methods: All patients with > 15 brain metastases on a prospective clinical trial, which randomized patients to either neuropsychology evaluation and intervention plus brain radiotherapy or brain radiotherapy alone, received MA-WBRT. Exclusion criteria included prior WBRT, pre-existing mental disability, and metastases within the avoidance neurocognitive substructures. All patients received 30 Gy in 10 fractions of MA-WBRT and were prescribed memantine. Cognition was measured by Hopkins Verbal Learning Test-Revised, Controlled Oral Word Association Test, and Trail Making Test A/B, with cognitive decline defined as decline on at least one assessment using reliable change index (same tests and definition as NRG CC001). Results: Between August 2022 and May 2024, 29 patients received MA-WBRT. Baseline characteristics included a median KPS of 80 (IQR 70, 90), median age of 64 (IQR 54, 69), 62% female patients, and a plurality of patients with a primary lung cancer (48%), The median overall survival or time to last follow up was 7.9 months. The three-month decline in neurocognitive function comparing the control and intervention groups for patients receiving MA-WBRT was 15.4% and 18.8%, respectively (p = 0.39). There was one failure in the right fornix 10 months after enrollment, but this was associated with concurrent distant intracranial failure outside the memory avoidance zone. Conclusions: The cognitive decline rate of approximately 17% at three months for patients receiving MA-WBRT compares favorably to a 3-month cognitive decline rate of 50% seen on NRG CC001. Additionally, MA-WBRT does not appear to significantly increase the risk of intracranial failure. Further evaluation of the delayed impact ( > 6 months) of MA-WBRT on cognitive function will be reported when data are available. A direct comparison of MA-WBRT plus memantine vs. HA-WBRT plus memantine is forthcoming with a randomized phase 3 trial. Clinical trial information: NCT05503251 .