5 March 2023 in Epilepsia
Focal cortical dysplasias (FCDs) are malformations of cortical development and one of the most common pathologies causing pharmacoresistant focal epilepsy. Resective neurosurgery yields high success rates, especially if the full extent of the lesion is correctly identified and completely removed. The visual assessment of magnetic resonance imaging does not pinpoint the FCD in 30%–50% of cases, and half of all patients with FCD are not amenable to epilepsy surgery, partly because the FCD could not be sufficiently localized. Computational approaches to FCD detection are an active area of research, benefitting from advancements in computer vision. Automatic FCD detection is a significant challenge and one of the first clinical grounds where the application of artificial intelligence may translate into an advance for patients' health. The emergence of new methods from the combination of health and computer sciences creates novel challenges. Imaging data need to be organized into structured, well‐annotated datasets and combined with other clinical information, such as histopathological subtypes or neuroimaging characteristics. Algorithmic output, that is, model prediction, requires a technically correct evaluation with adequate metrics that are understandable and usable for clinicians. Publication of code and data is necessary to make research accessible and reproducible. This critical review introduces the field of automatic FCD detection, explaining underlying medical and technical concepts, highlighting its challenges and current limitations, and providing a perspective for a novel research environment.
31 December 2024 in Epilepsia
Objective: Focal cortical dysplasia (FCD) is a common cause of drug‐resistant focal epilepsy but can be challenging to detect visually on magnetic resonance imaging. Three artificial intelligence models for automated FCD detection are publicly available (MAP18, deepFCD, MELD) but have only been compared on single‐center data. Our first objective is to compare them on independent multicenter test data. Additionally, we train and compare three new models and make them publicly available.Methods: We retrospectively collected FCD cases from four epilepsy centers. We chose three novel models that take two‐dimensional (2D) slices (2D‐nnUNet), 2.5D slices (FastSurferCNN), and large 3D patches (3D‐nnUNet) as inputs and trained them on a subset of Bonn data. As core evaluation metrics, we used voxel‐level Dice similarity coefficient (DSC), cluster‐level F1 score, subject‐level detection rate, and specificity.Results: We collected 329 subjects, 244 diagnosed with FCD (27.7 ± 14.4 years old, 54% male) and 85 healthy controls (7.1 ± 2.4 years old, 51% female). We used 118 subjects for model training and kept the remaining subjects as an independent test set. 3D‐nnUNet achieved the highest F1 score of .58, the highest DSC of .36 (95% confidence interval [CI] = .30–.41), a detection rate of 55%, and a specificity of 86%. deepFCD showed the highest detection rate (82%) but had the lowest specificity (0%) and cluster‐level precision (.03, 95% CI = .03–.04, F1 score = .07). MELD showed the least performance variation across centers, with detection rates between 46% and 54%.Significance: This study shows the variance in performance for FCD detection models in a multicenter dataset. The two models with 3D input data showed the highest sensitivity. The 2D models performed worse than all other models, suggesting that FCD detection requires 3D data. The greatly improved precision of 3D‐nnUNet may make it a sensible choice to aid FCD detection.
29 October 2024 in Epilepsia
Objective: High‐field magnetic resonance imaging (MRI) is a standard in the diagnosis of epilepsy. However, high costs and technical barriers have limited adoption in low‐ and middle‐income countries. Even in high‐income nations, many individuals with epilepsy face delays in undergoing MRI. Recent advancements in ultra‐low‐field (ULF) MRI technology, particularly the development of portable scanners, offer a promising solution to the limited accessibility of MRI. In this study, we present and evaluate the imaging capability of ULF MRI in detecting structural abnormalities typically associated with epilepsy and compare it to high‐field MRI at 3 T.Methods: Data collection was conducted within 3 consecutive weeks at the University Hospital Bonn. Inclusion criteria were a minimum age of 18 years, diagnosed epilepsy, and clinical high‐field MRI with abnormalities. We used a .064 T Swoop portable MR Imaging System. Both high‐field MRI and ULF MRI scans were evaluated independently by two experienced neuroradiologists as part of their clinical routine, comparing pathology detection and diagnosis completeness.Results: Twenty‐three individuals with epilepsy were recruited. One subject presented with a dual pathology. Across the entire cohort, in 17 of 24 (71%) pathologies, an anomaly colocalizing with the actual lesion was observed on ULF MRI. For 11 of 24 (46%) pathologies, the full diagnosis could be made based on ULF MRI. Tumors and posttraumatic lesions could be diagnosed best on ULF MRI, whereas cortical dysplasia and other focal pathologies were the least well diagnosed.Significance: This single‐center series of individuals with epilepsy demonstrates the feasibility and utility of ULF MRI for the field of epileptology. Its integration into epilepsy care offers transformative potential, particularly in resource‐limited settings. Further research is needed to position ULF MRI within imaging modalities in the diagnosis of epilepsy.
1 April 2025 in Epilepsia Open
Objective: This study aims to report human performance in the detection of Focal Cortical Dysplasias (FCDs) using an openly available dataset. Additionally, it defines a subset of this data as a “difficult” test set to establish a public baseline benchmark against which new methods for automated FCD detection can be evaluated.Methods: The performance of 28 human readers with varying levels of expertise in detecting FCDs was originally analyzed using 146 subjects (not all of which are openly available), we analyzed the openly available subset of 85 cases. Performance was measured based on the overlap between predicted regions of interest (ROIs) and ground‐truth lesion masks, using the Dice‐Soerensen coefficient (DSC). The benchmark test set was chosen to consist of 15 subjects most predictive for human performance and 13 subjects identified by at most 3 of the 28 readers.Results: Expert readers achieved an average detection rate of 68%, compared to 45% for non‐experts and 27% for laypersons. Neuroradiologists detected the highest percentage of lesions (64%), while psychiatrists detected the least (34%). Neurosurgeons had the highest ROI sensitivity (0.70), and psychiatrists had the highest ROI precision (0.78). The benchmark test set revealed an expert detection rate of 49%.Significance: Reporting human performance in FCD detection provides a critical baseline for assessing the effectiveness of automated detection methods in a clinically relevant context. The defined benchmark test set serves as a useful indicator for evaluating advancements in computer‐aided FCD detection approaches.Plain Language Summary: Focal cortical dysplasias (FCDs) are malformations of cortical development and one of the most common causes of drug‐resistant focal epilepsy. Once found, FCDs can be neurosurgically resected, which leads to seizure freedom in many cases. However, FCDs are difficult to detect in the visual assessment of magnetic resonance imaging. A myriad of algorithms for automated FCD detection have been developed, but their true clinical value remains unclear since there is no benchmark dataset for evaluation and comparison to human performance. Here, we use human FCD detection performance to define a benchmark dataset with which new methods for automated detection can be evaluated.