Revisiting the Concept of “Local” Recurrence in Primary Central Nervous System Lymphomas (PCNSL) after Complete Response to MTX-based Therapy: Are “Local” Failures Truly Just Periventricular?

Abstract

Despite favorable initial responses of PCNSL to 1st line methotrexate (MTX) based chemotherapy, its subsequent management remains plagued by high recurrence rates, toxicity, and unknown optimal consolidation strategies. Understanding patterns of relapse may help inform underlying disease biology, mechanisms of recurrence, ideal treatment strategies, and optimal consolidative radiation therapy (RT) design. In this study we report patterns of relapse in PCNSL patients following 1st line MTX with or without consolidation RT. Retrospective, single-institution analysis of patients with histologically confirmed PCNSL following 1st line MTX-based chemotherapy between March 2000 and May 2019. Exclusion criteria included primary refractory disease and an absence of MRI-based parenchymal brain involvement (i.e. intraocular or CSF only). The location of T1 contrast and T2 FLAIR lesions were compared between initial diagnosis and at first relapse. Lesions were categorized with respect to periventricular location (within ≤10 mm of ventricles). Relapses were further characterized as in-field (within T1 contrast enhancing lesion), marginal (within T2 FLAIR, but not T1 enhancing region), or distant (within brain, but non-overlapping with initial T2 FLAIR). Failure within CSF, extra-axial, and vitreous compartments was also recorded. PFS and OS were estimated with Kaplan-Meier methods and compared with the log-rank test. To test differences between cohorts, the Kruskal–Wallis, Pearson’s Chi-square and Fisher’s exact tests were used when appropriate. 83 patients with PCNSL met inclusion criteria and received MTX-based chemotherapy, of which 31 (37%) underwent consolidation RT. The median RT dose including the boost was 23.4 Gy (range, 18-45.0 Gy). PFS and OS were 58.1 and 67.8 months, respectively. After a median follow-up of 38.8 months a total of 34 (41%) patients recurred. Most patients (N = 54, 65%) had initial disease with some component involving the periventricular region. Periventricular vs. non-periventricular recurrence location, PFS, and OS rates did not significantly differ by receipt of consolidation XRT (p>0.05). Surprisingly, local recurrences (N = 11) were exclusively observed within periventricular lesions, whereas distant recurrences (N = 22) were seen in both periventricular and non-periventricular locations (p = 0.002). Median time to progression was strikingly shorter for locally recurrent lesions as compared to distant recurrences (13.8 vs. 28.7 months, p = 0.02). Following 1st line MTX, few lesions appear to recur within the initial T1 contrast volume and many local failures may be alternatively classified as additional periventricular failures. These observations argue against the use of focal-only RT consolidation in patients who achieve CR following MTX-based chemotherapy, and suggest that periventricular spread, potentially via occult CSF involvement, may play a central role in PCNSL recurrence.