Predictive Factors of Response to Immunotherapy in Lymphomas: A Multicentre Clinical Data Warehouse Study (PRONOSTIM)

SIMPLE SUMMARY: Immunotherapy is increasingly used in lymphoma strategy. Risk-adapted therapeutical management and set-up scores to -stratify the most vulnerable patients by risk are becoming major concerns. With the continuing upward trend of real-world data usage in addition to clinical trial data, it is possible to test the feasibility of using data from clinical data warehouses (CDWs) to identify new predictive factors for response or toxicity to immunotherapy. Based on a large set of biological and clinical factors, our results confirm already known predictors factors of CAR T (chimeric antigen receptor T) cells: age, elevated lactate dehydrogenase, and C-Reactive Protein at the time of infusion. Additionally male gender, low hemoglobin, and hypo- or hyperkalemia are demonstrated to be predictive factors for progression after CAR T cell therapy. Thus, the attractiveness of CDW for generating data by building ever larger cohorts is proven, enabling significant results to be obtained in line with those previously described in the literature. ABSTRACT: Immunotherapy (IT) is a major therapeutic strategy for lymphoma, significantly improving patient prognosis. IT remains ineffective for a significant number of patients, however, and exposes them to specific toxicities. The identification predictive factors around efficacy and toxicity would allow better targeting of patients with a higher ratio of benefit to risk. PRONOSTIM is a multicenter and retrospective study using the Clinical Data Warehouse (CDW) of the Greater Paris University Hospitals network. Adult patients with Hodgkin lymphoma or diffuse large-cell B lymphoma treated with immune checkpoint inhibitors or CAR T (Chimeric antigen receptor T) cells between 2017 and 2022 were included. Analysis of covariates influencing progression-free survival (PFS) or the occurrence of grade ≥3 toxicity was performed. In total, 249 patients were included. From this study, already known predictors for response or toxicity of CAR T cells such as age, elevated lactate dehydrogenase, and elevated C-Reactive Protein at the time of infusion were confirmed. In addition, male gender, low hemoglobin, and hypo- or hyperkalemia were demonstrated to be potential predictive factors for progression after CAR T cell therapy. These findings prove the attractiveness of CDW in generating real-world data, and show its essential contribution to identifying new predictors for decision support before starting IT.