Real World (RW) Outcomes and Prognostication of Older Patients with Primary Central Nervous System Lymphoma (PCNSL) in the Contemporary Era

Introduction: Treatment of older patients (pts) with PCNSL is challenging due to the prevalence of comorbidities, frailty, and complexities with delivery of chemotherapy (CT). The optimal induction CT and consolidation regimens for older PCNSL pts is unknown. Moreover, there are few large scale prognostication studies available, including analysis of geriatric assessments (GA). We analyzed detailed characteristics, treatment patterns and outcomes with prognostication across 17 academic centers.

Methods: We conducted a large, RW retrospective study of newly diagnosed PCNSL pts (1/2008-1/2019) ages ≥ 60 years (yrs). Survival rates were estimated by Kaplan-Meier with differences assessed by log rank test. We detailed Cumulative Index Rating Scale-Geriatric (CIRS-G) scores & other GAs. Univariate associations were derived via Cox model with variables p<0.05 entered stepwise into a multivariate (MVA) model.

Results: Among 491 initial cases, n=450 cases were verified for diagnosis & follow-up. Clinical features included: median age 71 yrs (60-88); male 47%; elevated LDH 30%; creatinine clearance <60 ml/min in 20% (median 81 ml/min); hemoglobin <10 g/dL in 8%; and albumin <3.5 g/dL in 35%. 21% of pts had prior or concurrent malignancy and 7% had history of solid organ transplant or autoimmune disease. Histology was DLBCL in 96% (COO non-GC in 76%) with CD20 expression seen in 98%. Immunohistochemistry showed MYC & BCL2 double expression in 40%; EBER was noted in 10% of pts (3/4 being PTLD pts). For disease location, brain parenchyma was involved in 94% of pts with 46% having a single site (54% >1 site). Cerebral involvement predominated in 75% with deep structure involvement in 20% & cerebellum in 5%. CSF involvement was documented in 13% of pts (unchecked in 26%). For GA at diagnosis, the median CIRS-G score was 6 (range 0-27) and impaired self-care activities of daily living (ADLs) were noted in 36%. Furthermore, geriatric syndrome (ie, dementia, delirium, depression, and/or falls) was present in 45% of pts.

Induction therapy included CT in 91% of pts (of whom 82% had rituximab (Rtx)) and radiation therapy (RT) in 8%. The most common chemotherapy regimens were: high-dose methotrexate (HD MTX) or HD MTX with Rtx (MR) in 38%; HD MTX/procarbazine/vincristine (MPV) +/- Rtx 30%; HD MTX/temozolomide/Rtx (MTR) 22%; Rtx alone 2%; and HD MTX/cytarabine/thiotepa/Rtx (MATRIX) in 2%. Median MTX dosing for all pts was 3.5 g/m² (range 1-8 g/m²), and by 3 most common regimens (all g/m²): MTR 5.1; MR 5.4; MPV 3.1 (P<.0001). For response to induction, 60% had complete response (CR), 18% partial response (PR), 6% stable disease & 16% had primary refractory disease. Induction CT was stopped due to toxicity in 21% of pts and the treatment-related mortality was 7%. Among 321 pts with CR or PR, 14% had autologous stem cell transplant (ASCT); 25% received consolidative CT; and 5% had RT. The most common CT maintenance regimens were temozolomide (n=22), lenalidomide (n=20) and HD MTX (n=15). Among pts experiencing relapse or progression, the most common 2^nd line therapies were RT (n=40), MTX (n=39), temozolomide (n=14) and MTR (n=10); 2 pts ultimately went on to receive ASCT.

With 42 month median follow-up (1-125), 3-yr PFS & OS for all pts were 38% & 52%, respectively (Fig 1A/1B). On MVA, factors associated with inferior PFS were: advancing age (continuous HR 1.05, P<.001); anemia (HR 1.14, P=.0035); high CIRS-G (HR 1.038, P=0.017) and geriatric syndrome (HR 1.537, P=.0098) (Fig 1C); and for inferior OS: advancing age (continuous HR 1.04, P=.01); low albumin (HR 2.203, P<.001); high CIRS-G (HR 1.053, P=.011); and geriatric syndrome (HR 1.851, P=.005) (Fig 1D). Among all pts, increasing HD MTX dosing in 500 mg/m² increments was associated with improved PFS (HR 0.958, P=.0002) & OS (HR 0.954, P=.001); and pts treated with MTR had improved PFS & OS vs MPV or MR (Fig 1E /1F). The favorable effect of MTR vs MR persisted when controlling for age, CIRS-G & geriatric syndrome. Additionally, use of Rtx was associated with improved outcomes (PFS HR 0.592, P<.0001; OS HR 0.528, P<0.0001). Finally, pts achieving CR had significantly improved survival (Fig 1G/1H).

Conclusions: Older pts with PCNSL have suboptimal outcomes, with 2/3 progressing in the first several years. GA is an important prognostic tool, and could be used to stratify pts in future investigations. In addition, use of Rtx, increasing MTX dose, and the MTR regimen were associated with improved outcomes.