Morphometric analysis on T1-weighted MRI complements visual MRI review in focal cortical dysplasia

Wong‐Kisiel, Lily C.; Quiroga, Diego F. Tovar; Kenney‐Jung, Daniel L.; Witte, Robert J.; Santana-Almansa, Alexandra; Worrell, Gregory A.; Britton, Jeffrey W.; Brinkmann, Benjamin H.

doi:10.1016/j.eplepsyres.2018.01.018

Cited by 35 publications

(27 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Mainly, MRI features of FCD include focal cortical thickening, fuzziness between gray matter (GM) and white matter (WM), cortical/subcortical WM hyperintensity on T2-weighted imaging (T2WI)/fluidattenuated inversion recovery (FLAIR), widened gyri, and abnormal sulci. 2 Currently, three main conventional methods for detecting epileptic foci exist: the voxel-based morphometry (VBM) algorithm; 3,4 surface-based morphometry (SBM) algorithm; [5][6][7] and postprocessing method, [8][9][10] which is based on pixel feature extraction.…”

Section: Introductionmentioning

confidence: 99%

Automatic localization and segmentation of focal cortical dysplasia in FLAIR‐negative patients using a convolutional neural network

Feng

Zhao

et al. 2020

J Applied Clin Med Phys

View full text Add to dashboard Cite

Purpose Focal cortical dysplasia (FCD) is a common cause of epilepsy; the only treatment is surgery. Therefore, detecting FCD using noninvasive imaging technology can help doctors determine whether surgical intervention is required. Since FCD lesions are small and not obvious, diagnosing FCD through visual evaluations of magnetic resonance imaging (MRI) scans is difficult. The purpose of this study is to detect and segment histologically confirmed FCD lesions in images of normal fluid‐attenuated inversion recovery (FLAIR)‐negative lesions using convolutional neural network (CNN) technology. Methods The technique involves training a six‐layer CNN named Net‐Pos, which consists of two convolutional layers (CLs); two pooling layers (PLs); and two fully connected (FC) layers, including 60 943 learning parameters. We employed activation maximization (AM) to optimize a series of pattern image blocks (PIBs) that were most similar to a lesion image block by using the trained Net‐Pos. We developed an AM and convolutional localization (AMCL) algorithm that employs the mean PIBs combined with convolution to locate and segment FCD lesions in FLAIR‐negative patients. Five evaluation indices, namely, recall, specificity, accuracy, precision, and the Dice coefficient, were applied to evaluate the localization and segmentation performance of the algorithm. Results The PIBs most similar to an FCD lesion image block were identified by the trained Net‐Pos as image blocks with brighter central areas and darker surrounding image blocks. The technique was evaluated using 18 FLAIR‐negative lesion images from 12 patients. The subject‐wise recall of the AMCL algorithm was 83.33% (15/18). The Dice coefficient for the segmentation performance was 52.68. Conclusion We developed a novel algorithm referred to as the AMCL algorithm with mean PIBs to effectively and automatically detect and segment FLAIR‐negative FCD lesions. This work is the first study to apply a CNN‐based model to detect and segment FCD lesions in images of FLAIR‐negative lesions.

show abstract

Section: Introductionmentioning

confidence: 99%

Automatic localization and segmentation of focal cortical dysplasia in FLAIR‐negative patients using a convolutional neural network

Feng

Zhao

et al. 2020

J Applied Clin Med Phys

View full text Add to dashboard Cite

show abstract

“…The cortical thickness mean image and the cortical thickness SD image were calculated from the T1 images of 32 normal subjects. Each individual was processed from step (1) to step (4). The average of all individuals was taken to obtain the cortical thickness mean image, and the standard deviation of all individuals was taken to obtain the cortical thickness SD image.…”

Section: Methodsmentioning

confidence: 99%

“…Several conventional methods for the detection of epileptic foci include the voxelbased morphometry algorithm (VBM) [2][3][4][5], the surface-based morphometry algorithm (SBM) [6,7] and the postprocessing method [8,9] based on voxel feature extraction. The VBM technique is mainly based on the image density, compared with the normal template, and the abnormal area found in the image is taken as the lesion area.…”

mentioning

confidence: 99%

Detecting focal cortical dysplasia lesions from FLAIR-negative images based on cortical thickness

et al. 2020

View full text Add to dashboard Cite

Background: Focal cortical dysplasia (FCD) is a neuronal migration disorder and is a major cause of drug-resistant epilepsy. However, many focal abnormalities remain undetected during routine visual inspection, and many patients with histologically confirmed FCD have normal fluid-attenuated inversion recovery (FLAIR-negative) images. The aim of this study was to quantitatively evaluate the changes in cortical thickness with magnetic resonance (MR) imaging of patients to identify FCD lesions from FLAIR-negative images. Methods: We first used the three-dimensional (3D) Laplace method to calculate the cortical thickness for individuals and obtained the cortical thickness mean image and cortical thickness standard deviation (SD) image based on all 32 healthy controls. Then, a cortical thickness extension map was computed by subtracting the cortical thickness mean image from the cortical thickness image of each patient and dividing the result by the cortical thickness SD image. Finally, clusters of voxels larger than three were defined as the FCD lesion area from the cortical thickness extension map. Results: The results showed that three of the four lesions that occurred in non-temporal areas were detected in three patients, but the detection failed in three patients with lesions that occurred in the temporal area. The quantitative analysis of the detected lesions in voxel-wise on images revealed the following: specificity (99.78%), accuracy (99.76%), recall (67.45%), precision (20.42%), Dice coefficient (30.01%), Youden index (67.23%) and area under the curve (AUC) (83.62%). Conclusion: Our studies demonstrate an effective method to localize lesions in nontemporal lobe regions. This novel method automatically detected FCD lesions using only FLAIR-negative images from patients and was based only on cortical thickness feature. The method is noninvasive and more effective than a visual analysis for helping doctors make a diagnosis.

show abstract

“…This hypothesis serves as the basis of many EEG-based applications, e.g., detection of epileptic seizures [ 13 ] or sleep stages [ 14 ]. EEG has already become one of the major technologies to monitor and detect the epilepsy occurrences due to its nature of non-invasiveness and low-cost [ 15 ].…”

Section: Introductionmentioning

confidence: 99%

Integration of 24 Feature Types to Accurately Detect and Predict Seizures Using Scalp EEG Signals

Zhang

Shen

Liu

et al. 2018

Sensors

View full text Add to dashboard Cite

The neurological disorder epilepsy causes substantial problems to the patients with uncontrolled seizures or even sudden deaths. Accurate detection and prediction of epileptic seizures will significantly improve the life quality of epileptic patients. Various feature extraction algorithms were proposed to describe the EEG signals in frequency or time domains. Both invasive intracranial and non-invasive scalp EEG signals have been screened for the epileptic seizure patterns. This study extracted a comprehensive list of 24 feature types from the scalp EEG signals and found 170 out of the 2794 features for an accurate classification of epileptic seizures. An accuracy (Acc) of 99.40% was optimized for detecting epileptic seizures from the scalp EEG signals. A balanced accuracy (bAcc) was calculated as the average of sensitivity and specificity and our seizure detection model achieved 99.61% in bAcc. The same experimental procedure was applied to predict epileptic seizures in advance, and the model achieved Acc = 99.17% for predicting epileptic seizures 10 s before happening.

show abstract

Morphometric analysis on T1-weighted MRI complements visual MRI review in focal cortical dysplasia

Cited by 35 publications

References 38 publications

Automatic localization and segmentation of focal cortical dysplasia in FLAIR‐negative patients using a convolutional neural network

Automatic localization and segmentation of focal cortical dysplasia in FLAIR‐negative patients using a convolutional neural network

Detecting focal cortical dysplasia lesions from FLAIR-negative images based on cortical thickness

Integration of 24 Feature Types to Accurately Detect and Predict Seizures Using Scalp EEG Signals

Contact Info

Product

Resources

About