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.