Background: Abdominal aortic aneurysm (AAA) carries high morbidity and mortality, while adverse events (AE) after endovascular aneurysm repair (EVAR) remain a major challenge. Accurate risk stratification is essential. Radiomics, by extracting quantitative imaging features, offers a noninvasive approach that may improve prediction of AAA outcomes and EVAR-related AE beyond conventional imaging.

Main body: A search was conducted across Medline, Scopus, Embase, and Web of Science databases for articles published up to February 2024 to evaluate the application of radiomics in predicting the outcomes of AAA and diagnosing or predicting the AE related to EVAR. Inclusion criteria concentrated on observational studies that utilized a radiomics model based on the radiomic features extracted from imaging modalities to predict AAA outcomes and predict or diagnose EVAR-related AE. Eight studies involving 1729 observations met the inclusion criteria out of the 371 records yielded from the databases. Radiomics achieved a high area under the receiver operating characteristic (ROC) curve (AUROC) for predicting EVAR-related AE in three studies (up to 0.95), and it showed predictive value for AAA growth in three studies (AUROC 0.79-0.93) and rupture or repair in one study (AUROC 0.75). One study highlighted the diagnostic value of radiomics in post-EVAR endoleak detection using unenhanced computed tomography images (AUROC 0.91).

Conclusion: Radiomics showed promising but preliminary results in predicting AAA outcomes and EVAR-related AE, providing a noninvasive tool for risk assessment. However, further research with larger cohorts and external validation of the radiomics models is essential to confirm its clinical applicability.