To synthesize the applications of computer vision in vascular surgery and evaluate the quality of existing studies, highlighting their significance.
Key Findings:
33% of studies addressed aortic pathologies, 30% carotid stenosis, and 25% foot ulcers, indicating a focus on major vascular issues.
81% of studies were observational with retrospective data, and only one clinical trial was included, raising concerns about the robustness of findings.
Dice coefficient (51%) and accuracy (36%) were the most reported performance metrics, with AUROC used in only 17% of studies, suggesting a need for standardized metrics.
Only 15% of studies had a low risk of bias, and adherence to the TRIPOD+AI checklist was poor at 57%, indicating methodological weaknesses.
Interpretation:
The findings indicate a significant increase in research on computer vision in vascular surgery, but highlight the need for improved study quality and focus on underrepresented areas like peripheral artery disease.
Limitations:
Limited focus on peripheral artery disease, which may affect comprehensive understanding.
Poor adherence to established reporting guidelines, potentially compromising study reliability.
Predominance of observational studies may affect generalizability of results.
Conclusion:
There is a need for greater attention to peripheral artery disease and improved methodological rigor in studies utilizing computer vision technologies in vascular surgery, to enhance clinical applicability.
