Significant Preservation of Neurocognitive Function (NCF) and Patient-Reported Symptoms with Hippocampal Avoidance (HA) during Whole-Brain Radiotherapy (WBRT) for Brain Metastases: Final Results of Nrg Oncology CC001

Abstract

NRG Oncology CC001, a phase III trial of WBRT plus memantine (WBRT+M) with or without HA, sought to evaluate the neuro-protective effects of avoiding the hippocampus using intensity-modulated radiotherapy. Patients (pts) with brain metastasis were stratified by RPA class and prior radiosurgery/surgery and randomized to WBRT+M or HA-WBRT+M (30Gy in 10 fractions). Standardized NCF tests were performed at baseline, 2, 4, 6, and 12 months (mos). The primary endpoint was NCF failure, defined as decline using the reliable change index on Hopkins Verbal Learning Test-Revised, Trail Making Test, or Controlled Oral Word Association. Patient-reported symptoms were assessed using the MD Anderson Symptom Inventory Brain Tumor module. Cumulative incidence estimated time to NCF failure (death without NCF failure was competing risk); between-arms differences tested using Gray’s test. Deterioration at each collection time point was tested using a chi-square test. Patient-reported symptoms were assessed using the MD Anderson Symptom Inventory Brain Tumor module and analyzed using mixed effects models and t-tests. From 7/2016 to 3/2018, 518 pts were randomized. Median age was 61.5 years. Median follow-up for alive patients was 7.9 mos. HA-WBRT+M was associated with lower risk of NCF failure (adjusted HR=0.74, 95% confidence interval (CI): 0.58-0.95, p=0.02). The difference was first seen at 4 months (62.7% HA-WBRT+M vs. 54.5% WBRT+M) and was attributable to improvements in executive function at 4 mos (p=0.01) and encoding (p=0.049) and consolidation (p=0.02) at 6 mos. Age≤61 predicted lower risk of NCF failure (HR=0.60, 95%CI: 0.46-0.79, p=0.0002); non-significant test for interaction indicated independent effects of HA and age on NCF. Patient-reported fatigue (p=0.036), difficulty speaking (p=0.049) and problems remembering things (p=0.013) at 6 mos favored the HA-WBRT+M arm. Imputation models accounting for missing data confirmed NCF results and also favored the HA-WBRT+M arm for patient-reported cognition (p=0.011) and symptom interference (p=0.008) at 6 mos. Toxicity, overall survival, and intracranial progression were not significantly different between the treatment arms. Hippocampal avoidance during WBRT better preserves neurocognitive function and patient-reported symptoms, while achieving similar intracranial control and survival, and should be considered standard of care for patients eligible to receive WBRT and anticipated to live 4 months or longer. Data analyses for longer-term (12-mo) follow-up are planned for June 2019 and will be presented for the first time at the meeting. Supported by grants UG1CA189867 (NCORP), U10CA180868 (NRG Oncology Operations), DCP from the National Cancer Institute.