Background
CD19-targeted chimeric antigen receptor T (CAR-T) cell therapy stands out as a revolutionary intervention, exhibiting remarkable remission rates in patients with refractory/relapsed (R/R) B-cell malignancies. However, the potential side effects of therapy, particularly cytokine release syndrome (CRS) and infections, pose significant challenges due to their overlapping clinical features. Promptly distinguishing between CRS and infection post CD19 target CAR-T cell infusion (CTI) remains a clinical dilemma. Our study aimed to analyze the incidence of infections and identify key indicators for early infection detection in febrile patients within 30 days post-CTI for B-cell malignancies.
Methods
In this retrospective cohort study, a cohort of 104 consecutive patients with R/R B-cell malignancies who underwent CAR-T therapy was reviewed. Clinical data including age, gender, CRS, ICANS, treatment history, infection incidence, and treatment responses were collected. Serum biomarkers procalcitonin (PCT), interleukin-6 (IL-6), and C-reactive protein (CRP) levels were analyzed using chemiluminescent assays. Statistical analyses employed Pearson’s Chi-square test, t-test, Mann–Whitney U-test, Kaplan–Meier survival analysis, Cox proportional hazards regression model, Spearman rank correlation, and receiver operating characteristic (ROC) curve analysis to evaluate diagnostic accuracy and develop predictive models through multivariate logistic regression.
Results
In this study, 38 patients (36.5%) experienced infections (30 bacterial, 5 fungal, and 3 viral) within the first 30 days of CAR T-cell infusion. In general, bacterial, fungal, and viral infections were detected at a median of 7, 8, and 9 days, respectively, after CAR T-cell infusion. Prior allogeneic hematopoietic cell transplantation (HCT) was an independent risk factor for infection (Hazard Ratio [HR]: 4.432 [1.262–15.565], P = 0.020). Furthermore, CRS was an independent risk factor for both infection ((HR: 2.903 [1.577–5.345], P < 0.001) and severe infection (9.040 [2.256–36.232], P < 0.001). Serum PCT, IL-6, and CRP were valuable in early infection prediction post-CAR-T therapy, particularly PCT with the highest area under the ROC curve (AUC) of 0.897. A diagnostic model incorporating PCT and CRP demonstrated an AUC of 0.903 with sensitivity and specificity above 83%. For severe infections, a model including CRS severity and PCT showed an exceptional AUC of 0.991 with perfect sensitivity and high specificity. Based on the aforementioned analysis, we proposed a workflow for the rapid identification of early infection during CAR-T cell therapy.
Conclusions
CRS and prior allogeneic HCT are independent infection risk factors post-CTI in febrile B-cell malignancy patients. Our identification of novel models using PCT and CRP for predicting infection, and PCT and CRS for predicting severe infection, offers potential to guide therapeutic decisions and enhance the efficacy of CAR-T cell therapy in the future.
