Results of the IELSG32 trial at a median follow-up of 88 months demonstrate that matrix followed by autologous transplant is associated with excellent survival and neurotolerability in patients with primary CNS lymphoma

Abstract

Introduction: The MATRix regimen (methotrexate, cytarabine, thiotepa, rituximab) significantly improved outcome of pts with primary CNS lymphoma (PCNSL) enrolled in the IELSG32 trial. At a median follow-up of 40 months, both whole-brain irradiation (WBRT) and autologous transplantation (ASCT) were safe and equally effective. However, sound assessment of OS, late complications, incidence of secondary tumors, and cognitive impairment requires longer follow-up. Herein, we report the results of IELSG32 trial at a median follow-up of 88 (IQR 77- 99) months. Methods: pts with untreated PCNSL (18-70 years) were randomly as signed to methotrexate-cytarabine (arm A), or arm A + rituximab (arm B), or arm B + thiotepa (MATRix; arm C). A second randomization assigned pts with responsive/stable disease after induction to WBRT (arm D) or BCNU-Thiotepa-conditioning ASCT (arm E). Treatment effect on cognitive functions and quality of life (QoL) were addressed by IPCG tests panel and EORTC-QLQ. Results: 219 pts were randomized (arm A 75; B 69; C 75). After induction, 167 had responsive/stable disease: 118 were assigned to WBRT (59) or ASCT (59) while 49 were excluded from 2nd randomization. Fifteen pts died of iatrogenic toxicity; 87 (40%) pts remain relapse-free (A 17; B 28; C 42), 14 of them died of unrelated causes (Table 1). Among 117 relapsing pts, 96 died of PCNSL, 7 of salvage therapy complications. Eight pts developed second cancers at 48-96 months from WBRT (5) or ASCT (3). Second tumors and deaths in relapse-free pts or during salvage were not significantly related to treatments (Table 1). Neuropsychological tests showed a significant impairment in attentiveness and executive functions in the WBRT arm, while transplanted pts had a significant improvement in these functions as well as memory and QoL. Pts treated with MATRix showed significantly better PFS (7-year: 20% arm A; 29% arm B; 52% arm C) and OS (7-year: 26% arm A; 37% arm B; 56% arm C). No significant differences were seen between the consolidation arms for either PFS (7-yr: 55% arm D; 50% arm E) or OS (7- yr: 63% vs 57%). Pts treated with MATRix and consolidation had a 7-yr OS of 70%, without a difference between WBRT and ASCT. Conclusions: MATRix was linked to excellent long-lasting outcome. WBRT and ASCT have comparable efficacy. MATRix and ASCT did not result in higher non-relapse mortality or second tumors onset. WBRT led to impairment of specific cognitive functions.