Emerging Biomarkers for Predicting and Mitigating Cardiotoxicity in Breast Cancer Treatment

Circulating microRNA profiling identifies candidates that predict cardiotoxic risk in HER2-positive breast cancer patients receiving anthracyclines and trastuzumab, offering a data-driven route to earlier detection. This update gives clinicians concise, actionable evidence on specific miRNA markers, model performance, implicated pathways, and practical next steps for risk-based surveillance and pilot interventions.

In a baseline cohort of 47 HER2-positive patients, plasma was sampled before therapy with cardiac assessments every 3 months for up to 15 months; differential miRNA expression at baseline tracked with later cardiac toxicity. The principal diagnostic signals were hsa-miR-155-5p and hsa-miR-124-3p, which showed the strongest ROC performance among candidates. Overall, 45 miRNAs were differentially expressed versus non‑toxic controls, supporting a coherent baseline signature. AUCs around 0.75–0.76 for the top miRNAs indicate moderate-to-good discrimination for early risk detection from a single baseline measurement.

Machine-learning classifiers—support vector machines, Random Forest, and k‑nearest neighbors—produced the best prognostic discrimination in this sample, with SVM and Random Forest outperforming a simple decision-tree approach. The top miRNA predictors yielded mid‑0.7 AUCs; as always, sensitivity and specificity depend on the chosen threshold. Still, multivariable, model-based classification improved baseline risk separation compared with univariate biomarkers alone, suggesting a role for model-informed triage to intensified imaging or early cardiology referral.

Pathway enrichment for the top-ranked miRNAs highlighted apoptosis, p53 signaling, MAPK cascades, and focal adhesion pathways—mechanisms congruent with known cardiotoxic effects of anthracyclines and trastuzumab. Anthracycline-related oxidative stress and DNA damage, together with trastuzumab-associated disruption of HER2 survival signaling, make p53 and MAPK engagement biologically plausible. The focal adhesion and apoptosis signals further support hypotheses of cytoskeletal disruption and programmed cell death contributing to functional decline.

Practical next steps for translation include using miRNA-based risk scores to trigger intensified imaging (for example, earlier echocardiography or strain imaging), testing established cardioprotective agents within higher-risk strata, and prioritizing drug candidates that target the enriched pathways. Integrating miRNA-driven risk stratification with clinical risk factors can help select patients for prophylactic cardioprotection rather than apply it universally. Prospective piloting or an adaptive trial randomizing high‑risk patients to intensified surveillance plus cardioprotection would generate stronger evidence of clinical utility.