Comparison of Clinical Features and Outcomes in Patients with Primary Central Nervous System Lymphoma Treated within and Outside a Clinical Trial: A Retrospective Study

Background: More than 25 years ago, it was already pointed out that clinical trials for older patients are urgently needed (Kennedy, JCO 1991). Today, clinical trials in patients with primary central nervous system lymphoma (PCNSL) include mainly a highly selected patient population with regard to age and comorbidity. Recent studies report improved survival rates, yet it is questionable whether these encouraging results from clinical trials in PCNSL mirror the "real-world" outcome.

Patients and Methods: This retrospective single-center study was performed in newly diagnosed consecutive patients with PCNSL treated in our tertiary care center between November 2000 and May 2017. The aim of this study was to compare the characteristics and survival data of patients treated within a clinical trial (German PCNSL Study Group 1 trial [NCT00153530]; TRIAL group) and patients treated outside this clinical trial ("real-world" setting; R-WORLD group). This analysis includes only patients who had received chemotherapy. Thus, patients treated with palliative dexamethasone or radiation therapy only as first-line treatment were excluded. In the TRIAL group, high-dose methotrexate (HD-MTX) as monotherapy was considered standard of care at that time and therefore was used as first-line treatment in this clinical trial. Most of the R-WORLD patients were treated with HD-MTX, but were "off-protocol". However, in January 2016, rituximab was added to HD-MTX in R-WORLD patients as first-line combination treatment.

Results: Altogether, consecutive 89 patients with newly diagnosed PCNSL were identified; 47 (52.8%) were men. Histology revealed aggressive lymphoma in 97.8% (87/89). Median age at diagnosis was 68 years (range 23-81 years), whereby 42.7% (38/89) of patients were ≥70 years. 69 patients were in the R-WORLD group and 20 patients in the TRIAL group. The R-WORLD patients were older (median age 70 years vs. 62 years; p=0.003), had a worse performance status (ECOG 2-4: 58.0% vs. 20.0%; p=0.005; median Karnofsky index 70% vs. 80%; p=0.002) and had more frequently high-risk disease according to Memorial Sloan Kettering Cancer Center prognostic model (class 3: 42.0% vs. 20.0%; p=0.12). No differences for comorbidity (according to Charlson Comorbidity Index) were found. The majority of the patients in the R-WORLD group received HD-MTX as first-line treatment (56/69 [81.2%]); all patients in the TRIAL group received HD-MTX as first-line treatment. As of 07/13/2017, median follow-up was 113.9 months (range 0.5-172.3 months). In this period, 66 deaths had occurred. In the entire cohort, the median overall survival (OS) was 12.6 months (95% CI 2.0-23.2). Median OS in the R-WORLD group was shorter comparing to the TRIAL group (9.5 months [95% CI 4.0-15.0] vs. 33.8 months [95% CI 17.6-50.0], hazard ratio [HR] 0.67 [95% CI 0.38-1.19]; p=0.17). Interestingly, we found no differences in the number of patients with early death (i.e., death within 4 months after histological diagnosis) comparing R-WORLD group and TRIAL group (20/69 [29.0%] vs. 4/20 [25.0%]; p=0.62). For the entire cohort, a median progression-free survival (PFS) of 5.0 months (95% CI 0-10.2) was estimated. In the R-WORLD group, PFS was 3.9 months (95% CI 2.2-5.6) vs. 25.1 months (95% CI 4.7-45.5) in the TRIAL group (HR 0.67 [95% CI 0.38-1.17]; p=0.16).

Interpretation: Our single-center experience in PCNSL patients with a long follow-up of 10 years showed that improved OS was restricted to patients treated within the PCNSL clinical trial, only. However, older and frail patients (≥70 years and ECOG ≥2, respectively) were underrepresented in the PCNSL clinical trial, thereby potentially producing misleading results with respect to the "real-world" situation. Thus, in PCNSL, a careful review of trial eligibility criteria is required in order to better include the "real-world" patient population, tailor appropriately therapy regimens and provide better clinical care.