Incidence of Central Nervous System (CNS) Relapse in De Novo Adult Acute Lymphoblastic Leukemia (ALL)

In the era of improved outcomes with systemic chemotherapy for de novo adult ALL, prevention of CNS relapse becomes even more paramount. Isolated CNS relapse may herald eventual marrow relapse in the absence of definitive CNS-directed and systemic chemotherapy strategies. CNS prophylaxis can be successfully achieved with intrathecal chemotherapy treatments (IT) in conjunction with systemic chemotherapy inclusive of high-dose methotrexate (MTX) and/or high-dose cytarabine (ara-C) without use of radiation therapy (XRT). The hyper-CVAD regimen (cycles of fractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone alternating with cycles of high dose MTX (1 g/m²) and cytarabine (3 gm/m²; 1 gm/m² for age > 60 yrs) incorporates IT MTX alternating with IT ara-C during the intensive phase of induction-consolidation. The number of IT treatments was determined by CNS risk (4 IT for low risk = LDH < 1400 U/L and low proliferative index; 16 IT for high risk = LDH > 1400 U/L or high proliferative index; 8 IT for indeterminate risk = one of parameters unknown). An analysis performed in 2000 (Kantarjian et al, J Clin Oncol 18: 547, 2000) identified CNS relapse rates of 6%, 2%, and 0% for the low, high, and indeterminate CNS risk groups, respectively. The number of IT administered was modified from 4 to 6, from 16 to 8, and maintained at 8 for these CNS risk groups, respectively, until 2010, at which point all pts received 8 IT. An analysis was conducted to examine the incidence of CNS relapse for 565 pts with de novo ALL treated with hyper-CVAD-based regimens (n=453, inclusive of nelarabine, monoclonal antibodies such as rituximab, ofatumumab or inotuzumab, and/or tyrosine kinase inhibitors such as imatinib, dasatinib or ponatinib) from 2001 to 2014 or with the augmented BFM regimen (one IT ara-C, 15 IT MTX, Capizzi methotrexate) (n=112) from 2006 to 2014. T-lymphoblastic lymphoma (designated for 8 IT) and Burkitt leukemia/lymphoma (designated for 16 IT) cases were excluded from the analysis. The incidence of CNS relapse (n=42) was 7% overall; 59% were isolated CNS relapses (n=25, 5%) without concurrent marrow relapse. Median time to isolated CNS relapse was 19 mos versus 10 mos for concurrent relapse. The incidence of CNS relapse for cases with CNS disease at initial presentation (n=70) was 19%. Factors predictive of higher incidence of CNS relapse included younger age (12% for 30 yrs or younger, 3.5% for 60 yrs or older, p=.01), elevated LDH > 1400 U/L (13% vs 4%, p<.001), and Philadelphia chromosome positivity (12% vs 6%, p=.05). The overall CNS relapse rates were 4% for the hyper-CVAD and monoclonal antibody regimens; 11% for the augmented BFM regimen, and 14%-15% for the hyper-CVAD and imatinib/dasatinib regimens. The CNS relapse rate declined from 10% for pts treated prior to 2010 to 8% thereafter (8 IT for all risk groups). Overall, the 3-yr survival rates for CNS relapse was 47% vs 63% without CNS relapse (p=0.006). Modifications to the hyper-CVAD and tyrosine kinase inhibitor regimens for Philadelphia positive ALL have been implemented to increase the number of IT from 8 to 12. Similar modifications will be implemented for the augmented BFM regimen with consideration for incorporation of high dose MTX. Additional analyses will further refine the CNS risk model in order to guide CNS prophylaxis. Alternative agents such as IT rituximab may improve outcomes for CD20 positive cases with CNS disease at presentation; a clinical trial in the setting of active CNS disease is underway.