Lessons from the IELSG45 trial: The impact of patient fitness on dropout rate in primary central nervous system lymphoma (PCNSL) studies

Abstract Background: Primary central nervous system lymphomas (PCNSL) are rare, aggressive B-cell malignancies with distinct biological and clinical characteristics [1]. Standard PCNSL treatment in fit patients is based on high-dose methotrexate (HDMTX)-containing induction chemotherapy followed by autologous stem cell transplant consolidation. In elderly patients, however, this approach is usually not feasible, and consolidation strategies are associated with unacceptable risk of neurological and hematological toxicity. Maintenance may be a suitable strategy for older responders. Herein, we share insights from the IELSG45 FIORELLA randomized phase II trial on fitness- and comorbidity-tailored treatment in elderly patients with newly diagnosed PCNSL. Methods: The design and sample size of the trial were based on the 37% 2-year progression-free survival (PFS) rate and 40% dropout rate among elderly patients in the PRIMAIN study [2], the largest available prospective trial in this setting. For the IELSG45 FIORELLA trial, all patients aged ≥70 years who were treated at the participating sites were evaluated for trial participation. Patients were screened for suitability to receive HDMTX-containing chemotherapy (based on organ function and physician judgment) and assigned to one of two treatment strategies. Patients deemed adequately fit were assigned to receive the standard combination of HDMTX, procarbazine, and rituximab as induction therapy. Responding patients were then randomized to receive either procarbazine or lenalidomide as maintenance therapy (Part A). Patients ineligible for HDMTX were assigned to receive concomitant whole-brain radiotherapy, temozolomide, and rituximab as induction therapy, followed by maintenance temozolomide (Part B). Results: Seventy-two evaluable patients were enrolled across 24 clinical sites in four countries between June 2019 and June 2022. Preliminary analysis of the FIORELLA trial revealed an excessive (i.e., >40%) dropout rate at the end of induction chemo-immunotherapy, suggesting that the enrolled population was less fit than expected. This discrepancy prompted a protocol amendment introducing a predefined threshold in Part A of the study: if a subsequent interim analysis showed a dropout rate exceeding 50%, the study would be terminated. Although no unexpected toxicity was observed, the interim analysis confirmed an overall dropout rate of 60% (95% CI, 46%-73%). Therefore, according to the amended study protocol, the trial was closed. At a median follow-up of 36 months (range 25–55 months), the PFS rate was 21% (±5%) and the overall survival (OS) rate was 25% (±6%) in the entire study population (N=72). Fifty-five patients were registered in Part A; 40 of them experienced a PFS-defining event: 33 had progressive disease and 7 died from treatment-related complications. The 3-year PFS and OS were 23 (±6%) and 25 (±7%), respectively. Twenty-two patients with responsive disease and no major complications were randomized to receive either procarbazine (n=9) or lenalidomide (n=13) as maintenance therapy. The 3-year PFS was 33 (±16%) in the procarbazine arm and 46 (±14%) in the lenalidomide arm. The 3-year OS was 44 (±16%) for procarbazine and 54 (±14%) for lenalidomide. Seventeen patients who were not eligible for HDMTX-containing chemotherapy were enrolled in part B. Fourteen of them experienced a PFS-defining event: 9 had progressive lymphoma and 5 died (3 due to infection, 1 for subarachnoid bleeding and 1 for unknown cause). The 3-year PFS and OS in this group were 18 (±9%) and 24 (±10%), respectively. Conclusion: Risk of treatment failure (in terms of dropout and survival rates) appeared higher compared to the PRIMAIN trial. Discrepancies in feasibility and efficacy between PRIMAIN and FIORELLA trials generate some concerns about generalization of results of PCNSL trials among different geographical regions. These unexpected outcomes offer several valuable lessons on the incorporation of comorbidity indexes into trial eligibility/stratification and the relevance of interim analyses to inform ethical decisions early. References Ferreri, A.J.M., et al. Primary central nervous system lymphoma. 2023. Nat Rev Dis Primers. 2023; 9: 29. Fritsch, K., et al. High-dose methotrexate-based immuno-chemotherapy for elderly primary CNS lymphoma patients (PRIMAIN study). 2017. Leukemia. 2017; 31: 846-852.