Time-dynamic perspectives on cancer therapy-related cardiotoxicity: a systematic review and time-course evidence synthesis for multibiomarker monitoring.

OBJECTIVE: Current monitoring of cancer therapy-related cardiac dysfunction relies largely on threshold-based interpretation at isolated time points, yet its temporal evolution remains insufficiently characterized. This review aims to map longitudinal monitoring evidence across treatments, biomarkers, and time windows; reconstruct trajectories of key biochemical and imaging markers in evidence-rich settings; and explore temporal patterns between biochemical injury signals and imaging-detected functional changes. METHODS: PubMed and Embase were searched from inception to 1 October 2025. A three-dimensional evidence map was constructed across treatment modality, monitoring marker, and time window using a custom four-level evidence grading system based on data completeness and extractability. Quantitative synthesis was restricted to cohorts providing extractable longitudinal absolute values at mappable follow-up times in non-outcome-driven designs. Standardized trajectories of hs-cTnI/T, NT-proBNP, left ventricular ejection fraction (LVEF), and global longitudinal strain (GLS) were aggregated within predefined time windows. RESULTS: Forty-six independent cohorts yielded 387 raw longitudinal observations; after harmonization and representative-value selection, 173 effect rows from 31 analytic study IDs (29 primary reports) contributed to the trajectory synthesis. Evidence was concentrated in anthracycline-based and anthracycline plus concurrent/sequential HER2 inhibitor settings, whereas data on immune checkpoint inhibitor (ICI) therapy, vascular endothelial growth factor receptor tyrosine kinase inhibitor (VEGF-TKI) therapy, and the biomarkers sST2 and GDF-15 were sparse. In evidence-rich settings, hs-cTnI tended to show early elevations (D0-M3), while more pronounced declines in LVEF and GLS were observed mainly after M3. Because estimates were derived from aggregated study-level data, this ordering is descriptive and hypothesis-generating rather than evidence of within-patient temporal precedence. Under anthracycline plus concurrent/sequential HER2 inhibitor exposure, hs-cTnI elevations and later imaging declines were greater in the available windows; data beyond 12 months were sparse, making the apparent absence of recovery provisional. CONCLUSION: This review provides a unified time-window framework for multimarker longitudinal evidence across anticancer treatment settings, identifies critical evidence gaps, and describes an exploratory temporal pattern that may inform hypotheses about stage-specific monitoring in evidence-rich settings. Patient-level longitudinal studies are required before any monitoring schedule can be validated. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD420261367699, PROSPERO CRD420261367699.