The hidden cardiovascular burden of CAR-T candidates: insights from a prospective multimodal imaging study

Abstract Background Candidates for CAR-T cell therapy usually present with a complex cardiovascular (CV) background due to prior chemotherapies and comorbidities. While guidelines recommend baseline risk assessment, the actual diagnostic yield of a comprehensive multimodal screening—and its impact on clinical management prior to infusion—remains poorly defined. Objective To characterise the cardiovascular phenotype of CAR-T candidates and evaluate the clinical utility of a systematic screening protocol including cardiac magnetic resonance (CMR). Methods prospective, single centre cohort of 236 consecutive candidates referred for CAR-T therapy (2019-2024). We conducted a baseline assessment, which comprised clinical evaluation (including SCORE2/SCORE2 OP), biomarkers (NT-proBNP, high sensitivity troponin I/T), transthoracic echocardiography, and cardiac magnetic resonance (CMR) with tissue characterisation (T1/T2 mapping, extracellular volume). We analysed the prevalence of undiagnosed CV disease and the rate of therapeutic optimisation. Results The cohort (n=236, mean age 59.3±11.4 years, 61.4% male) had a substantial burden of risk: 86.4% had at least one CV risk factor, yet 44.5% had suboptimal control. The systematic screening was high-yield, leading to a new diagnosis of left ventricular systolic disfunction in 23 patients (9.7%), previously unknown ischaemic heart disease in 9 patients (3.8%) and oedema in baseline CMR in 33 patients (14.2%). Consequently, medical therapy was optimised in 30.2% of the cohort. ECG documented sinus rhythm in 93.5% and repolarisation abnormalities in 16.2%. Regarding imaging studies, left ventricular ejection fraction (LVEF) was comparable between echocardiography and CMR (59.5% vs 59.4%, p=0.19). However, CMR identified subclinical abnormalities missed by echocardiography: 25.6% had elevated extracellular volume (>30%), and pathological mapping values (Z-score >2) were observed in 9.7% for native T1 and 17.3% for T2. Furthermore, late gadolinium enhancement was identified in 28 patients (14.3%), predominantly displaying non-ischaemic patterns (intramyocardial n=11, subepicardial n=4) rather than subendocardial scars. Conclusion Systematic multimodal CV screening in CAR-T candidates is clinically actionable: even when TTE shows preserved left ventricular ejection fraction, CMR unmasks occult inflammation/fibrosis and refines pre-infusion risk stratification. These findings support routine multimodal screening to facilitate targeted optimisation prior to CAR-T.Table. Baseline characteristics.