Artificial intelligence in cardiac tumors: Multimodal diagnosis, surgery, precision oncology, and clinical translation
Year: 2026
Type: article
Abstract: Artificial intelligence (AI) may assist the evaluation of cardiac tumors, but diagnostic performance must be distinguished from clinical usefulness.This Minireview examines a deliberately limited targeted sample of six full-source-checked primary reports, selected to illustrate magnetic resonance, computed tomography, echocardiographic classification, and registry-based prognosis.It is not a systematic review, a representative sample, or a near-comprehensive account of the field.The examples show promising discrimination in selected populations, including institution-external retrospective testing, but also small test sets, mixed reference standards, possible dependence between repeated observations, and restricted tumor spectra.Conflicting source values are identified rather than silently reconciled.Tumor-thrombus discrimination does not establish malignant-subtype recognition, resectability, or treatment benefit.For cardiovascular surgeons, the most relevant research outputs are reproducible lesion boundaries and relationships to valves, coronary arteries, myocardium, and great vessels.Molecular profiling, three-dimensional reconstruction, and magnetic-resonance-guided radiotherapy provide complementary clinical context, not proof of validated cardiac-tumor AI.A compact, author-proposed readiness framework prioritizes patient-disjoint evaluation, clinically meaningful errors, calibrated uncertainty, and prospective clinician-system assessment.These selected reports do not establish autonomous diagnosis or treatment selection.
DOI: 10.35713/aic.126838
Authors: Harendra Kumar, Fnu Rimza, Shashvat Gupta
Institution: Mayo Clinic
Language: English
Topic: Cardiac tumors and thrombi
Subfield: Cardiology and Cardiovascular Medicine
Field: Medicine
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7234122273
