Predicting the radiological outcome of cerebral aneurysm treatment with machine learning algorithms; a systematic review and diagnostic meta-analysis

Background: Up now, several model were proposed for predicting the outcome of treating brain aneurysm. This study aims to investigate the performance of Machine learning (ML) in predicting the outcome of intracranial aneurysm after endovascular or microsurgical management. Method: This systematic review and meta-analysis was prepared with adhering to Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guideline. The electronic databases of PubMed/Medline, Embase, Scopus, and Web of Science were systematically reviewed up to 5th June 2023. All statistical analysis was done by STATA/mp V.20. Results: Recent studies have utilized ML models to forecast the results of cerebral aneurysm treatment. The models exhibit a sensitivity of 0.88 and a specificity of 0.80, with the latter revealing substantial variability across the various studies. Meanwhile, the sensitivity remained consistent. The models displayed a positive DLR (PLR) of 4.5 and a negative DLR (NLR) of 0.15. The aggregate diagnostic score determined that heterogeneity was negligible, with a value of 3.41. However, the diagnostic odds ratio 30.29 indicated noteworthy heterogeneity among the studies. The pooled AUC of 0.90 showcased the ML's ability to predict cerebral aneurysm treatment outcomes accurately. Conclusion: Predicting treatment outcomes in patients with intracranial aneurysms non-invasively is a promising approach with good diagnostic performance. However, further studies are required to improve the accuracy of ML algorithms, as the pooled metrics had a wide confidence interval. This will not only enhance the reliability of the predictions but also facilitate the integration of these algorithms into daily clinical practice