EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Fully automated three‐dimensional deep learning‐based magnetic resonance imaging segmentation of brain cavities in epilepsy surgery

Raphael Fernandes Casseb iD, Brunno Machado de Campos iD, Wallace Souza Loos iD, Marcelo Eduardo Ramos Barbosa iD, Marina Koutsodontis Machado Alvim iD, Gabriel Chagas Lutfala Paulino iD, Enrico Ghizoni iD, Francesco Pucci, Samuel Worrell, Clarissa Lin Yassuda iD, Roberto Medeiros de Souza iD, Lara Jehi iD und 1 weitere

Epilepsia · 2026

DOI: 10.1002/epi.70296 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Epilepsia, 2026. Verfasst von Raphael Fernandes Casseb und 12 weiteren Autor:innen. Beteiligte Einrichtung: Neuroimaging Laboratory, Universidade Estadual de Campinas, Campinas, São Paulo, Brazil. Hauptthemen laut MEDLINE: Deep Learning, Magnetic Resonance Imaging, Epilepsy, Imaging, Three-Dimensional. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: NINDS NIH HHS, Fundação de Amparo à Pesquisa do Estado de São Paulo, NIH HHS.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Objective

There are several clinical and research applications for determining the amount of brain tissue resected after epilepsy surgery; however, manual segmentation of postoperative magnetic resonance imaging (MRI) is imprecise and time‐consuming. In this study, we developed and benchmarked ResectVol DL, a freely available deep learning‐based tool that performs this task automatically.

Methods

To create ResectVol DL, we trained a UNet‐like deep learning model using postoperative T1‐weighted MRI from epilepsy surgery patients and evaluated it against manual delineations (ground truth). ResectVol DL was also compared with three automated methods (ResectVol 1.1.2, DeepResection, and Auto3DSeg) using Dice similarity coefficient (DSC), Pearson correlation coefficient, and relative volume difference from manual segmentation. To assess false‐positive detections and generalizability beyond epilepsy, we additionally processed images from healthy controls (no resection) and brain tumor cases.

Results

The final epilepsy cohort comprised 120 patients (57 women, mean age at surgery = 31.5 ± 15.9 [SD] years), split into training ( n = 72) and test ( n = 48) sets. An additional 42 images (22 healthy controls and 20 brain tumor cases) were included to test for false positives and generalizability. Segmentation performance differed across methods (Friedman test, p < .001). ResectVol Dl achieved the highest median DSC (.925), significantly outperforming ResectVol 1.1.2, DeepResection, and Auto3DSeg after Bonferroni correction. Volume‐based metrics were similar for Auto3DSeg and ResectVol DL ( r = .988, relative difference = 8.4% vs. r = .985, 8.1%; no significant difference), yet Auto3DSeg produced three false‐positive cavities in no‐surgery controls (3/22, 95% confidence interval [CI] = 3%–35%), whereas none was observed for ResectVol DL and DeepResection (0/22, 95% CI = 0%–15%). Significance ResectVol DL provides accurate, fully automated segmentation of postoperative resection cavities, offering a robust and reproducible methodological tool for large‐scale postoperative imaging studies in epilepsy surgery. ResectVol DL also provides volumetric information derived from region labeling, which may serve as potential input for predictive models associated with surgery outcome; however, this application has not yet been validated.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Deep LearningMagnetic Resonance ImagingEpilepsyImaging, Three-DimensionalBrainHumansFemaleAdultMaleYoung AdultBrain NeoplasmsMiddle Aged

Stichwörter der Autor:innen

ResectVol DLbrain region labelingnnU‐netpostoperative MRIvolumetry

Autor:innen und Institutionen

  • Neuroimaging Laboratory, Universidade Estadual de Campinas, Campinas, São Paulo, Brazil.
  • Advanced Imaging and Artificial Intelligence Lab, University of Calgary, Calgary, Alberta, Canada.
  • School of Medical Sciences, Pontifical Catholic University of Campinas, Campinas, São Paulo, Brazil.
  • Department of Neurology, Cleveland Clinic Foundation, Cleveland, Ohio, USA.

ORCID-Profile: Raphael Fernandes Casseb, Brunno Machado de Campos, Wallace Souza Loos, Marcelo Eduardo Ramos Barbosa, Marina Koutsodontis Machado Alvim, Gabriel Chagas Lutfala Paulino, Enrico Ghizoni, Clarissa Lin Yassuda, Roberto Medeiros de Souza, Lara Jehi, Fernando Cendes

Förderung

  • NINDS NIH HHS
  • Fundação de Amparo à Pesquisa do Estado de São Paulo
  • NIH HHS
  • National Institutes of Health

Zitieren

APA

Casseb, R. F., de Campos, B. M., Loos, W. S., Barbosa, M. E. R., Alvim, M. K. M., Paulino, G. C. L., Ghizoni, E., Pucci, F., Worrell, S., Yassuda, C. L., de Souza, R. M., Jehi, L., & Cendes, F. (2026). Fully automated three‐dimensional deep learning‐based magnetic resonance imaging segmentation of brain cavities in epilepsy surgery. Epilepsia. https://doi.org/10.1002/epi.70296

Vancouver

Casseb RF, de Campos BM, Loos WS, Barbosa MER, Alvim MKM, Paulino GCL, et al. Fully automated three‐dimensional deep learning‐based magnetic resonance imaging segmentation of brain cavities in epilepsy surgery. Epilepsia. 2026. doi:10.1002/epi.70296

BibTeX
@article{Casseb2026,
  author = {Casseb, Raphael Fernandes and de Campos, Brunno Machado and Loos, Wallace Souza and Barbosa, Marcelo Eduardo Ramos and Alvim, Marina Koutsodontis Machado and Paulino, Gabriel Chagas Lutfala and Ghizoni, Enrico and Pucci, Francesco and Worrell, Samuel and Yassuda, Clarissa Lin and de Souza, Roberto Medeiros and Jehi, Lara and Cendes, Fernando},
  title = {Fully automated three‐dimensional deep learning‐based magnetic resonance imaging segmentation of brain cavities in epilepsy surgery},
  journal = {Epilepsia},
  year = {2026},
  doi = {10.1002/epi.70296}
}
Kostenlos starten →