Recent advances in sMRI and artificial intelligence for presurgical planning in focal cortical dysplasia: A systematic review

Comprehensive Summary

Through recent advances in Artificial Intelligence (AI), machine learning (ML), and deep learning (DL), there can be significant advances in the detection of Focal Cortical Dysplasia (FCD) and the standard of care. This study was conducted as a systematic review utilizing articles published from 2020 onwards from research databases such as: PubMed, Embase, Scopus, Web of Science, and ScienceDirect. The articles extracted were all original studies investigating the use of AI and structural MRI (sMRI) to detect FCD in human models, with quantitative performance metrics. Inclusion criteria of the analyzed studies were evaluated separately by two reviewers, with a third solving possible discrepancies. Through the analysis of each study, it was found that the model performance of sMRI and AI varied with FCD type and training datasets, posing the challenge of the efficacy of ML and DL in diagnosing and presurgical planning in a variety of FCD types. That being said, even with varying FCD types, AI detection increased the sensitivity of sMRI FCD diagnosis to 97.1% and specificities up to 84.3% in the MRI sequences, including MPRAGE, MP2RAGE, and FLAIR. With AI models, it is possible to detect and enhance presurgical planning for patients suffering from FCD.

Outcomes and Implications

Artificial intelligence (AI), including deep learning (DL) and machine learning (ML), holds significant potential to improve the accuracy of diagnosis as well as preoperative and operative planning for Focal Cortical Dysplasia (FCD).