Diffusion MRI for Pediatric Epilepsy Surgery
This study is testing a new way to use brain magnetic resonance imaging (MRI) to help children and teenagers with focal epilepsy (seizures starting in one area of the brain) who are planning to have surgery. The goal is to make surgery safer by more accurately finding important brain areas to avoid damaging them, which could reduce side effects and improve seizure control. We are looking at how well a special type of MRI, called diffusion-weighted MRI (DWMRI), combined with a deep convolutional neural network (DCNN) — a type of artificial intelligence — can predict these important brain pathways. This new imaging method takes about 10 minutes and doesn't require contrast or sedation. We are enrolling children and teens aged 3 to 19 with drug-resistant focal epilepsy who are planning two-stage epilepsy surgery, as well as healthy controls aged 5 to 19.
- Study design
- This interventional study plans to enroll 60 participants. It compares a new MRI technique to the current standard method for mapping brain areas.
- What's involved
- Participants will undergo brain MRI scans and age-appropriate neuro-psychology testing to assess brain function. The study will track outcomes for up to 1.5 years after surgery.
- Compensation
- Not stated in the trial record.
- Follow-up
- Participants will be followed for 1.5 years after surgery to assess for postoperative deficits.
AI-generated from the public study record. Only the study team can confirm whether you're eligible — confirm details with them before making decisions.
Diffusion MRI Methods to Minimize Postoperative Deficits in Pediatric Epilepsy Surgery
At a glance
Conditions
Where it's being run
1 sites across 1 statesWho to contact
Opens a ready-to-send draft in your own email app — review before sending.
What this trial measures
- Accuracy of DCNN tract classification for detection of ESM-defined eloquent white matter pathways in healthy controlsDuring procedure
Spatial overlap of DCNN tract classification (range: 0-100%, 0 indicating no overlap and 100% indicating complete overlap) will be evaluated between two different DWMRI scans of healthy controls: single-shell and generalized Q-sampling imaging (GQI) that are acquired on the same day. 14 ESM-defined eloquent pathways will be obtained using 14 DCNN tract classifications from the single-shell and GQI data, and the spatial overlap between single shell and GQI data (score: %) will be assessed per each pathway.
- Accuracy of DCNN tract classification for detection of ESM-defined eloquent area that will be acquired a month after the DCNN tract classification in children with drug-resistant epilepsy1 month
Spatial overlap (range: 0-100%, 0 indicating no overlap and 100% indicating complete overlap) will be measured between cortical terminals of DCNN-classified white matter pathways and their ground truth data: ESM-defined eloquent areas that will be acquired a month after the DCNN tract classification.
- Accuracy of DCNN tract classification for prediction of eloquent white matter pathways providing no postoperative deficits that will be assessed at 1.5 years after surgery1.5 years
Preservation (score: 1) vs. no preservation (score: 0) of preoperative DCNN-classified white matter pathways will be compared with presence (score: 1) vs. absence (score: 0) of postoperative deficits in primary motor, language, auditory, and visual functions that will be assessed at 1.5 years after surgery.
- Accuracy of DCNN tract classification combined with Kalman analysis to predict optimal margin balancing maximal seizure freedom and minimal functional deficits that will be assessed at 1.5 years after surgery1.5 years
Preservation (score: 1) vs. no preservation (score: 0) of preoperative DCNN-Kalman filter predicted surgical margin will be compared with presence (score: 1) vs. absence (score: 0) of postoperative deficits and seizure freedom that will be assessed at 1.5 years after surgery.
- Strength of association between local efficiency of preoperative network and functional measure: full-scale IQ, verbal-IQ, non-verbal IQ, expressive language, receptive language, and motor function that will be assessed at 1.5 years after surgery1.5 years
Local efficiency value (range: 0-1, 0 indicating no efficacy and 1 indicating the strongest efficacy) will be evaluated from full-scale IQ network, non-verbal IQ network, verbal IQ network, expressive language network, receptive language network, and motor network of preoperative DWMRI connectome data, respectively. Full-scale IQ (normal mean: 100, standard deviation: 15), verbal IQ (normal mean: 100, standard deviation: 15), non-verbal IQ (normal mean: 100, standard deviation: 15), expressive language score (normal mean: 50, standard deviation: 10), receptive language score (normal mean: 50, standard deviation: 10), and motor score (normal mean: 50, standard deviation: 10) will be also evaluated from neuro-psychology testing at 1.5 years after surgery. The correlation coefficient (range: 0-1, 0 indicating no correlation and 1 indicating complete correlation) will be evaluated between local efficiency and neuro-psychology score measured for each corresponding function.
- Strength of association between local efficiency of preoperative full-scale IQ network and epilepsy duration that will be assessed at the time of preoperative MRI (Hypothesis 2.2)Within 1 day
Full-scale IQ (normal mean: 100, standard deviation: 15) will be assessed at the time of preoperative MRI scan. It will be associated with local efficiency (range: 0-1, 0 indicating no efficacy and 1 indicating the strongest efficacy) of preoperative full-scale IQ network and epilepsy duration (range: 0-19 years) that will be assessed within 1 day of preoperative MRI scan. The correlation coefficient (range: 0-1, 0 indicating no correlation and 1 indicating a perfect correlation) will be evaluated between full-scale IQ and local efficiency of preoperative full-scale IQ network.
- Strength of association between local efficiency change of contralateral verbal-/non-verbal IQ network and verbal-/non-verbal IQ change that will be measured between 1 month before surgery and 1.5 years after surgery1.5 years
Longitudinal change of local efficiency in contralateral verbal-/non-verbal IQ network (range: -1 - +1, -1 indicating a complete loss of local efficiency after surgery and +1 indicating a complete gain of local efficiency after surgery) will be measured from postoperative and preoperative DWMRI connectome data that will be measured between 1 month before surgery and 1.5 years after surgery, respectively. It will be then correlated with the longitudinal change of verbal/non-verbal IQ (range: -100 - +100, -100 indication a complete loss of verbal/non-verbal IQ after surgery and +100 indicating a complete improvement of verbal/non-verbal IQ after surgery) that will be measured between 1 month before surgery and 1.5 years after surgery. The correlation coefficient (range: 0-1, 0 indicating no correlation and 1 indicating a perfect correlation) will be calculated between two longitudinal changes.