PURPOSE: In primary central nervous system lymphoma (PCNSL), the extent to which post-methotrexate consolidation contributes to neurotoxicity is unclear. Concerns for neurotoxicity from standard dose whole-brain radiation therapy (WBRT) have led to declining use. Cerebral atrophy is an established surrogate for neurotoxicity
however, the relative extent to which modern consolidation (ie, reduced-dose [RD-]WBRT ≤24 Gy, autologous hematopoietic cell transplant) contributes to cerebral atrophy is unclear. METHODS AND MATERIALS: Patients with PCNSL from 2000-2020 who achieved a complete response to consolidation following methotrexate-based induction were included. Inclusion criteria were preconsolidation magnetic resonance imaging (baseline) and ≥1 magnetic resonance imaging showing sustained remission at 1, 3, 5, or 10 years. An expert neuroradiologist longitudinally measured parenchymal volume loss via ventricular volumetric change. Linear mixed-effects models were performed to estimate absolute and annual volumetric change rates. RESULTS: Of 139 patients (median follow-up, 4.5 years), most were Memorial Sloan Kettering Cancer Center (MSK) recursive partitioning analysis (RPA) class 2 (age ≤50 years, Karnofsky performance score (KPS) ≥70). Consolidation therapies included nonmyeloablative chemotherapy (n = 57
41%), high-dose myeloablative chemotherapy with autologous hematopoietic cell transplant (n = 50
36%), and RD-WBRT (n = 28
20%). A higher MSK RPA class was associated with greater baseline ventricular volume (P <
.001). Overall adjusted annual ventricular volume change rates were greater than those published in healthy controls (4.3% vs 1.8%) and generally increased by age/decade at diagnosis: 40 to 49-year-olds 1.8% (95% CI, -1.4% to 5.0%), 50 to 59-year-olds 3.1% (95% CI, 0.7%-5.5%), 60 to 69-year-olds 4.8% (95% CI, 2.4%-7.3%), 70 to 79-year-olds 7.2% (95% CI, 4.3%-10.2%), and 80 to 89-year-olds 4.2% (95% CI, -1.1% to 9.6%). There were no significant associations between consolidation strategy and ventricular volume change rates accounting for age, KPS, gender, baseline ventricular volume, or interaction between age and consolidation. CONCLUSIONS: These findings demonstrate accelerated cerebral atrophy in PCNSL after consolidation compared with healthy adults. However, atrophy did not differ by consolidation strategy. These long-term results suggest acceptable neurotoxicity following RD-WBRT.