Characterization of Extramedullary Disease (EMD) Response to CD19 Targeted Chimeric Antigen Receptor T-Cells (CART) in Patients with Relapsed/Refractory B-Cell Acute Lymphoblastic Leukemia (B-ALL): A Multi-Site, Retrospective Cohort Review

Background CD19 CART are highly effective in eradicating bone marrow (BM) B-ALL, but efficacy in EMD is less clear. Recent data suggests that response in central nervous system (CNS) disease may be comparable to BM disease without increased toxicity. Data using CART in non-CNS EMD is limited to small retrospective analyses with demonstration of mixed BM/EMD responses. An improved understanding of CART efficacy in EMD is needed to help inform management decisions. Methods We conducted a multi-site, retrospective review of children and young adults who received CD19 CART between 2012 and 2019. Objectives included determining event free survival (EFS) and overall survival (OS), using the Kaplan-Meier method, of patients with variants of CNS disease (defined as CNS2/3 disease) and/or non-CNS EMD with either low (<5%, measurable residual disease (MRD) positive or negative) or high (≥5%) BM disease burden and describing locations of non-CNS EMD relapse off modern therapy. Results 290 patients were included in this cohort with disease involvement shown in Figure 1. Sites of non-CNS EMD are included in Figure 2. For those with and without any CNS/non-CNS EMD involvement, 2-year (yr) OS was 63.5% (50.3-74.1%) vs. 71.8% (65.1-77.4%) [p=0.056] and EFS was 34.1% (22.4-46.1%) vs. 54.3% (47.4-60.8%) [p=0.0015]. For patients with low BM burden + CNS disease (LD-CNS) vs. those with high BM burden + CNS disease (HD-CNS), 2-yr OS was 100% vs. 33.3% (14.9-53.1%) [p= 0.0028] and EFS was 57.7% (25.6-80.1%) vs. 14.3% (3.6-32.1%) [p=0.0006] (Figure 3). For patients with low BM burden + non-CNS EMD (LD-EMD) vs. those with high BM burden + non-CNS EMD (HD-EMD), 2-yr OS was 100% vs. 50.0% (22.9-72.2%) [p=0.046] and EFS was 58.4% (22.7-82.3%) vs. 21.4% (5.2-44.8%) [p=0.017]. There was no statistical difference in 2-yr OS/EFS in the LD-CNS group comparing patients who were MRD negative (100%/52.5% [12.2-82.1%]) vs. MRD positive (100%/60.0% [3.6-32.1%]) or in the LD-EMD group comparing patients who were MRD negative (100%/66.7% [19.5-90.4%]) vs. MRD positive (100%/33.3% [0.9-77.4%]). Conclusions Sites and presentations of non-CNS EMD involvement were diverse, and oncologists must be aware of this to promptly investigate unexpected sites of relapse prior to administration of CART. The presence of CNS or non-CNS EMD involvement was associated with inferior EFS compared to those without any CNS or non-CNS EMD involvement. High disease burden and CNS or non-CNS EMD involvement was associated with inferior EFS and OS compared to low disease burden and EMD or CNS involvement. Durable responses to CART can be achieved in patients with EMD or CNS disease but given that high burden BM disease with CNS/non-CNS EMD involvement represents a high-risk group with poor outcomes, new pre-CART staging evaluations and therapeutic interventions may need to be studied to improve outcomes.