Consolidation Regimen and Cerebral Atrophy in Patients With Primary Central Nervous System Lymphoma

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.

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.

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.

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.