Comprehensive Characteristics and Outcomes of Myeloid Neoplasms Following Chimeric Antigen Receptor T-Cell Therapy for Other Hematologic Malignancies

Introduction: Secondary myeloid neoplasms (MN) are a recognized risk following chimeric antigen receptor (CAR) T-cell therapy. Prior studies are limited to low-resolution adverse event reporting or small case series. We provide a comprehensive account of MN characteristics and outcomes following CAR T.

Methods: We reviewed our institutional experience with consecutive MN cases diagnosed after CAR T therapy for other hematologic malignancies since 2015. All preceding lines of therapy (LoT), including those administered before histologic transformation, were documented. Autologous stem cell transplant (autoSCT) was counted as a unique LoT. We compared available bone marrow samples from the immediate pre-CAR T period with those from MN diagnosis to identify clonal expansion.

Results: Of 1,484 patients treated with immune effector cells at our center, 37 (2.5%) received a subsequent MN diagnosis. Patients primarily received CD19- (81%) or BCMA- (16%) targeting CAR T products, all with fludarabine and cyclophosphamide lymphodepletion, at a median of 65 years of age (IQR 60-69) after a median of 5 (range 1-11) prior LoT for non-Hodgkin B-cell lymphoma or multiple myeloma (MM). Forty-three percent had undergone prior autoSCT with BEAM or melphalan conditioning. All patients had prior exposure to alkylating agents, and 41% had previous exposure to lenalidomide (len). Complete response rate post initial CAR T was 87%. Of the 11 patients who required additional therapy for the primary malignancy, 2 received a second CAR T infusion.

Among the patients (28 with MDS, 1 with CMML, 8 with AML) who developed a MN, 70% were male. At the time of MN diagnosis, 95% of patients had their primary malignancy in remission. Median baseline bone marrow blast percentage was 6% (IQR 2-11) and 46% of cases had ≥ 2% ring sideroblasts. Cytogenetic risk was chiefly adverse: 46% complex karyotype, 62% deletion 7/7q, 14% MECOM rearranged and 3% KMT2A rearranged. While no patients had pure erythroid leukemia, 18 had a multi-hit TP53 status per International Consensus Classification 2022 criteria. The MN was diagnosed a median of 16.1 months (IQR 8.8-24.4) following CAR T infusion, and a median of 5.5 years, 5.1 years, and 5.7 years after first autoSCT, alkylating agent, and len exposures, respectively. One third of patients received only supportive care and 65% were treated with hypomethylating agents. Most (80%) patients had no response to initial therapy. Of the 6 patients who underwent allogeneic stem cell transplant (alloSCT), 3 died from infection or graft-versus-host disease. Of the 3 patients who experienced a durable remission post alloSCT, 2 had a wild-type TP53 status.

Median overall survival (mOS) after MN diagnosis was 9.0 months (95% CI 6.1, 16.2). The majority (89%) of deaths were from MN-related complications. No statistically significant associations were identified between post MN survival and sex, age ≥ 65 years at CAR T infusion, prior autoSCT, ≥ 3 LoT pre-CAR T or TP53 status. There was an observed survival benefit with len exposure: mOS 14.3 months (95% CI 9.0, NE) with versus (vs) 5.1 months (95% CI 2.3, 16.1) without prior len exposure (p = 0.029). This association persisted when MM cases were excluded (p = 0.005). A trend towards improved survival was observed in alloSCT recipients: median OS 14.3 months with vs 7.5 months without transplant, although it did not reach statistical significance.

All 4 patients with documented TP53 mutation(s) (variant allele frequency [VAF] detection limit 0.3 to 2%) within 12 months prior to CAR T, and without active marrow involvement from the primary malignancy, demonstrated clonal expansion (relative VAF increase ≥ 50%), often accompanied by a second TP53 hit. Retrospective TP53 sequencing and T-cell profiling in available marrow samples are being explored.

Conclusions: Our incidence of MN following CAR T is reassuringly rare and comparable to an incidence of 1.4% from publicly available FDA data. The 5-year latency from pre–CAR T therapies to MN diagnosis aligns with historical therapy-related MN latencies, underscoring their role in malignant transformation. The shorter latency after CAR T therapy is confounded by lead-time bias but raises the question of whether CAR T-induced immune dysregulation amplifies clonal selection. Screening for TP53 mutations before CAR T administration may inform MN risk in a patient population now often cured of their primary malignancy.