Hippocampal Avoidance Confers Protection against Accelerated Brain Aging when Compared to Conventional Whole Brain Radiotherapy

Abstract

<h3>Purpose/Objective(s)</h3> Conventional whole brain radiotherapy (c-WBRT) has been a mainstay in the treatment of brain metastases, but is associated with significant long-term neurocognitive deficits in memory, problem-solving and executive function. WBRT with hippocampal avoidance (HA-WBRT) has been shown to better preserve cognitive function with no detriment in oncologic outcomes. Our group previously established using a magnetic resonance imaging (MRI) deep learning model that c-WBRT is associated with rapid, global brain aging based on anatomic surface morphology. Herein, we compare post-treatment brain aging between c-WBRT and HA-WBRT. <h3>Materials/Methods</h3> We studied patients with brain metastases who underwent either c-WBRT or HA-WBRT (per NRG-CC001 guidelines) and examined MRI anatomic surface data before and after treatment. We implemented an established graph convolutional neural network model to estimate patient's brain age. We further developed a fixed-effects linear model to compare age of whole brain and substructures after WBRT, written as follows: Δ^Age = φ + (1|patient); where for each patient, Δ^Age = [model_predicted_age_post_treatment – actual_age_post_treatment], and φ = fixed-effect coefficient for type of WBRT (c-WBRT versus HA-WBRT respectively). φ > 0 corresponded to HA-WBRT being associated with less aging compared to c-WBRT, φ < 0 corresponded to HA-WBRT being associated with greater aging, and non-significant φ implied that the interventions were not different in terms of aging. Mann-Whitney (MW) test and Fisher exact (FE) test were used to compare variables between the groups. <h3>Results</h3> 54 patients were analyzed (37 patients with c-WBRT and 17 patients with HA-WBRT). The median age was 61.5 years versus 55 years for c-WBRT and HA-WBRT patients respectively (<i>p</i>=0.2001 MW test). For c-WBRT and HA-WBRT patients respectively, 75% versus 82% had five or more metastases (<i>p</i>=0.7251 FE test), and 78% versus 88% had prior systemic therapy (<i>p</i>=0.4668 FE test). Median dose was 30 Gy for both groups. 71% of HA-WBRT patients received concurrent memantine versus 31% of c-WBRT patients (<i>p</i>=0.0149 FE test). Results of aging analysis are presented in Table 1<b>.</b> Between c-WBRT and HA-WBRT, there was no difference in Δ^Age for the whole brain, the cortex or subcortical structures; however, there was a statistically significant decrease in Δ^Age of the hippocampus for HA-WBRT group (φ = 9.99, <i>p</i>=0.002). Based on median follow up of approximately 6 months, the hippocampus aged an additional 9.99 years in c-WBRT patients compared to HA-WBRT patients. <h3>Conclusion</h3> Our results suggest that HA-WBRT with concurrent memantine confers focal brain aging protection that mirrors the long-term neurocognitive benefits seen when compared to c-WBRT.