Development and prospective clinical validation of a convolutional neural network for automated detection and segmentation of focal cortical dysplasias.

PURPOSE: Focal cortical dysplasias (FCDs) are a leading cause of drug-resistant epilepsy. Early detection and resection of FCDs have favorable prognostic implications for postoperative seizure freedom. Despite advancements in imaging methods, FCD detection remains challenging. House et al. (2021) introduced a convolutional neural network (CNN) for automated FCD detection and segmentation, achieving a sensitivity of 77.8%. However, its clinical applicability was limited due to a low specificity of 5.5%. The objective of this study was to improve the CNN's performance through data-driven training and algorithm optimization, followed by a prospective validation on daily-routine MRIs. MATERIAL AND METHODS: A dataset of 300 3 T MRIs from daily clinical practice, including 3D T1 and FLAIR sequences, was prospectively compiled. The MRIs were visually evaluated by two neuroradiologists and underwent morphometric assessment by two epileptologists. The dataset included 30 FCD cases (11 female, mean age: 28.1 ± 10.1 years) and a control group of 150 normal cases (97 female, mean age: 32.8 ± 14.9 years), along with 120 non-FCD pathological cases (64 female, mean age: 38.4 ± 18.4 years). The dataset was divided into three subsets, each analyzed by the CNN. Subsequently, the CNN underwent a two-phase-training process, incorporating subset MRIs and expert-labeled FCD maps. This training employed both classical and continual learning techniques. The CNN's performance was validated by comparing the baseline model with the trained models at two training levels. RESULTS: In prospective validation, the best model trained using continual learning achieved a sensitivity of 90.0%, specificity of 70.0%, and accuracy of 72.0%, with an average of 0.41 false positive clusters detected per MRI. For FCD segmentation, an average Dice coefficient of 0.56 was attained. The model's performance improved in each training phase while maintaining a high level of sensitivity. Continual learning outperformed classical learning in this regard. CONCLUSIONS: Our study presents a promising CNN for FCD detection and segmentation, exhibiting both high sensitivity and specificity. Furthermore, the model demonstrates continuous improvement with the inclusion of more clinical MRI data. We consider our CNN a valuable tool for automated, examiner-independent FCD detection in daily clinical practice, potentially addressing the underutilization of epilepsy surgery in drug-resistant focal epilepsy and thereby improving patient outcomes.

Automated detection and segmentation of focal cortical dysplasias (FCDs) with artificial intelligence: Presentation of a novel convolutional neural network and its prospective clinical validation.

PURPOSE: Focal cortical dysplasias (FCDs) represent one of the most frequent causes of pharmaco-resistant focal epilepsies. Despite improved clinical imaging methods over the past years, FCD detection remains challenging, as FCDs vary in location, size, and shape and commonly blend into surrounding tissues without clear definable boundaries. We developed a novel convolutional neural network for FCD detection and segmentation and validated it prospectively on daily-routine MRIs. MATERIAL AND METHODS: The neural network was trained on 201 T1 and FLAIR 3 T MRI volume sequences of 158 patients with mainly FCDs, regardless of type, and 7 focal PMG. Non-FCD/PMG MRIs, drawn from 100 normal MRIs and 50 MRIs with non-FCD/PMG pathologies, were added to the training. We applied the algorithm prospectively on 100 consecutive MRIs of patients with focal epilepsy from daily clinical practice. The results were compared with corresponding neuroradiological reports and morphometric MRI analyses evaluated by an experienced epileptologist. RESULTS: Best training results reached a sensitivity (recall) of 70.1 % and a precision of 54.3 % for detecting FCDs. Applied on the daily-routine MRIs, 7 out of 9 FCDs were detected and segmented correctly with a sensitivity of 77.8 % and a specificity of 5.5 %. The results of conventional visual analyses were 33.3 % and 94.5 %, respectively (3/9 FCDs detected); the results of morphometric analyses with overall epileptologic evaluation were both 100 % (9/9 FCDs detected) and thus served as reference. CONCLUSION: We developed a 3D convolutional neural network with autoencoder regularization for FCD detection and segmentation. Our algorithm employs the largest FCD training dataset to date with various types of FCDs and some focal PMG. It provided a higher sensitivity in detecting FCDs than conventional visual analyses. Despite its low specificity, the number of false positively predicted lesions per MRI was lower than with morphometric analysis. We consider our algorithm already useful for FCD pre-screening in everyday clinical practice.