Algorithm 1: Selection Process for the NRR

Define search query: ○ “(interventional radiology[Title/Abstract]) AND ((neural network[Title/Abstract]) OR (Artificial Intelligence[Title/Abstract]) OR (deep learning[Title/Abstract]) OR (ANN[Title/Abstract]) OR (GAN[Title/Abstract]))” 2.Conduct searches in PubMed and Scopus using the defined query. 3.Select relevant studies from peer-reviewed journals that focus on the field priority, such as recent reviews that assess prior studies and reviews providing broad analyses, integrating findings from previous works and with a focus on AI applications in line with the journal topic (articles more focused on computer sciences were excluded on the basis that they were not in line the focus of the journal). 4. Evaluate each study based on the following parameters: ○ N1: Clear rationale in the introduction. ○ N2: Appropriate research design. ○ N3: Clearly described methodology. ○ N4: Well-presented results. ○ N5: Conclusions justified by results. ○ N6: Disclosure of conflicts of interest. 5.Assign scores to parameters N1–N5 (scale of 1–5). 6.Assess N6 using a binary Yes/No measure. 7. Preselect studies meeting the following criteria: ○ N6 = “Yes” (conflict of interest disclosed). ○ N1–N5 scores > 3 (ensuring methodological rigor). 8.Include preselected studies in the final synthesis.