Abstract: Myocardial fibrosis is an important pathological basis for the occurrence and development of various cardiovascular diseases in the elderly, and is closely related to heart failure, arrhythmia and poor prognosis. With the development of imaging technology, non-invasive methods such as echocardiography, computed tomography and cardiac magnetic resonance have played important roles in the assessment of myocardial fibrosis. However, traditional image analysis mostly has the limitations of relying on visual assessment or having limited parameter measurement, and still has deficiencies in the identification of diffuse and microscopic heterogeneous fibrosis. Radiomics, by mining a large number of quantitative features from medical image data and combining statistical and machine learning methods for modeling and analysis, provides a new technical path for non-invasive and precise assessment of myocardial fibrosis. This article reviews the pathological basis of myocardial fibrosis in the elderly, the current status of imaging assessment and the progress of radiomics research. It points out that future efforts should focus on establishing standardized radiomics workflows to improve feature stability and reproducibility, conducting multicenter studies, and integrating multimodal imaging with multi-omics data to comprehensively characterize the disease phenotype of myocardial fibrosis from macroscopic morphology to microscopic molecular level.

Key words: myocardial fibrosis, aged, radiomics, cardiac magnetic resonance, cardiovascular imaging