Dynamics of Leukocyte Subpopulations Reconstitution Predict Infection Propensity in a Multiethnic Real World Cohort Treated with Anti-CD19 CAR-T Cell Therapy (Axicabtagene-Ciloleucel)

Background: Adoptive immunotherapy using CD19-targeted Chimeric Antigen Receptor T-cells (CAR-T) has revolutionized the treatment of relapsed/refractory diffuse large B-cell lymphoma (DLBCL). We have demonstrated the efficacy of FDA-approved axicabtagene ciloleucel (Yescarta) in a multiethnic New York City underserved population with 80% complete response (CR) rate in the first ten patients treated at our institution (Abbasi et al., 2020). There is limited data on the propensity of infections and lymphohematopoietic reconstitution after Day 30 (D30) following CAR-T cell therapy. In this study, we evaluated the prevalence and nature of infectious complications in an expanded cohort of DLBCL patients treated with CD19 CAR-T therapy and its association with the dynamics of leukocyte subpopulation reconstitution post-CAR-T cell therapy.

Methods: We conducted a retrospective study of patients who received CAR-T therapy at our institution between 2018-2020. Variables collected include patient demographics, absolute neutrophil (ANC), lymphocyte (ALC) and monocyte counts (AMC) at Day 30, hematologic reconstitution (ANC≥ 1500/µL) at Day 90 (D90), presence or absence of infections after D30 by clinical and/or microbiological parameters. Associations between presence of infection and D30 ANC, ALC, AMC, ANC/ALC ratio, AMC/ALC ratio were assessed using Kruskal-Wallis test. Association between infection and hematologic reconstitution at D90 was done using Chi-square test. Kaplan-Meier curves with log-rank test were used to evaluate overall survival (OS) and progression-free survival (PFS).

Results: Nineteen patients were evaluated in our study, consisting of 42% (8) Hispanic, 32% (6) Caucasian, 21% (4) African-American, and 5% (1) Asian subjects. Based on clinical and microbiologic data, 47% (9) developed an infection after D30 (infection group) while 53% (10) of subjects remained infection-free after D30 (non-infection group). The most common infection type observed was viral (11 patients) followed by bacterial (8 patients) and fungal (3 patients) (Table 1). Of 25 total infectious events, 44% (11) were grade 1 or 2 and 48% (12) were grade 3 with 10 being viral in etiology. Two deaths occurred due to an infectious process. Three patients tested SARS-CoV-2 positive and were hospitalized with COVID-19 pneumonia. Median OS and PFS has not been reached in either group. To determine the kinetics of lymphohematopoietic reconstitution and its association with infection risk, we evaluated the relationship between cytopenias and rates of infection after D30. Notably, compared to non-infection group, infection group had a higher median ALC (1000/µL vs 600/µL p=0.04), a lower median ANC/ALC ratio (1.4 vs 4.5 p<0.01) and a lower median AMC/ALC at D30 (0.36 vs 1.33, p=0.01) (Table 2). In addition, patients in the infection group had a lower rate of hematologic reconstitution (ANC >1500/µL) at D90. We observed that only 22% (2) of patients had recovered ANC > 1500/µLin the infection group as opposed to 80% (8) in the non-infection group at D90 (p= 0.038). Rates of cytokine release syndrome (CRS) were comparable between the two groups (55.6% vs 70% p=0.52). Surprisingly, rates of immune-effector cell associated neurotoxicity syndrome (ICANS) was lower (55.6%) in the infection group compared to (90%) non-infection group (p=0.09). Fourteen of 19 patients had follow-up over one year, of which 8 (57%) remained in complete remission (CR).

Conclusions: We demonstrate an infection rate of 47% (9) beyond D30 in patients undergoing CD19 CAR-T. Increased ALC, lower ANC/ALC and AMC/ALC ratios at D30 may be predictive of infectious complications. Median OS has not been reached in our cohort. Given the potential clinical impact, our observations should be corroborated using larger datasets.

View largeDownload slide

View largeDownload slide

Close modal