EUVIMEDEUVIMED entdecken
ÜbersichtsarbeitOpen Access · CC BY-NC 4.0

A Novel Convolutional Neural Network for Automated Multiple Sclerosis Brain Lesion Segmentation

Emma Dereskewicz iD, Francesco La Rosa iD, Jonadab Dos Santos Silva iD, Edward Sizer, Amit Kohli, Maxence Wynen, William A Mullins, Pietro Maggi, Sarah Levy, Kamso Onyemeh, Batuhan Ayci, Andrew J Solomon und 5 weitere

Journal of Neuroimaging · 2025-09 · 35 (5)

DOI: 10.1111/jon.70085 ↗

Kurz gefasst

Übersichtsarbeit in der Zeitschrift Journal of Neuroimaging, online erschienen am 12.09.2025. Verfasst von Emma Dereskewicz und 16 weiteren Autor:innen. Beteiligte Einrichtung: Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, New York, USA. Hauptthemen laut MEDLINE: Multiple Sclerosis, Magnetic Resonance Imaging, Brain, Neural Networks, Computer. Die Arbeit wurde bisher 13-mal zitiert (Crossref) und nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: ORIP NIH HHS, Bristol Myers Squibb Foundation, TRAIL.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background and Purpose Assessment of brain lesions on magnetic resonance imaging (MRI) is crucial for research in multiple sclerosis (MS). Manual segmentation is time‐consuming and inconsistent. We aimed to develop an automated MS lesion segmentation algorithm for T2‐weighted fluid‐attenuated inversion recovery (FLAIR) MRI.

Methods

We developed FLAIR Lesion Analysis in Multiple Sclerosis (FLAMeS), a deep learning‐based MS lesion segmentation algorithm based on the nnU‐Net 3D full‐resolution U‐Net and trained on 668 FLAIR 1.5 and 3 tesla scans from persons with MS. FLAMeS was evaluated on three external datasets: MSSEG‐2 ( n = 14), MSLesSeg ( n = 51), and a clinical cohort ( n = 10), and compared to SAMSEG, LST‐LPA, and LST‐AI. Performance was assessed qualitatively by two blinded experts and quantitatively by comparing automated and ground truth lesion masks using standard segmentation metrics.

Results

In a blinded qualitative review of 20 scans, both raters selected FLAMeS as the most accurate segmentation in 15 cases, with one rater favoring FLAMeS in two additional cases. Across all testing datasets, FLAMeS achieved a mean Dice score of 0.74, a true positive rate of 0.84, and an F1 score of 0.78, consistently outperforming the benchmark methods. For other metrics, including positive predictive value, relative volume difference, and false positive rate, FLAMeS performed similarly to or better than benchmark methods. Most lesions missed by FLAMeS were smaller than 10 mm 3 , whereas the benchmark methods missed larger lesions in addition to smaller ones.

Conclusions

FLAMeS is an accurate, robust method for MS lesion segmentation that outperforms other publicly available methods.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Multiple SclerosisMagnetic Resonance ImagingBrainNeural Networks, ComputerImage Interpretation, Computer-AssistedHumansFemaleMaleAdultMiddle AgedAlgorithmsDeep LearningConvolutional Neural Networks

Stichwörter der Autor:innen

MRIlesion segmentationmachine learningmultiple sclerosis

Autor:innen und Institutionen

  • Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
  • Windreich Department of Artificial Intelligence & Human Health, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
  • ICTeam, Université Catholique De Louvain, Louvain-la-Neuve, Belgium.
  • Neuroinflammation Imaging Lab, Université Catholique, de Louvain, Brussels, Belgium.
  • Translational Neuroradiology Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland, USA.
  • Cliniques Universitaires Saint-Luc, Université Catholique De Louvain, Brussels, Belgium.

ORCID-Profile: Emma Dereskewicz, Francesco La Rosa, Jonadab Dos Santos Silva, Erin S Beck

Förderung

  • ORIP NIH HHS
  • Bristol Myers Squibb Foundation
  • TRAIL
  • NINDS NIH HHS
  • NCATS NIH HHS
  • National Multiple Sclerosis Society
  • NIH HHS
  • Office of the Assistant Secretary of Defense for Health Affairs, Multiple Sclerosis Research Program

Zitieren

APA

Dereskewicz, E., La Rosa, F., Dos Santos Silva, J., Sizer, E., Kohli, A., Wynen, M., Mullins, W. A., Maggi, P., Levy, S., Onyemeh, K., Ayci, B., Solomon, A. J., Assländer, J., Al-Louzi, O., Reich, D. S., Sumowski, J., & Beck, E. S. (2025). A Novel Convolutional Neural Network for Automated Multiple Sclerosis Brain Lesion Segmentation. Journal of Neuroimaging, 35(5). https://doi.org/10.1111/jon.70085

Vancouver

Dereskewicz E, La Rosa F, Dos Santos Silva J, Sizer E, Kohli A, Wynen M, et al. A Novel Convolutional Neural Network for Automated Multiple Sclerosis Brain Lesion Segmentation. Journal of Neuroimaging. 2025;35(5). doi:10.1111/jon.70085

BibTeX
@article{Dereskewicz2025,
  author = {Dereskewicz, Emma and La Rosa, Francesco and Dos Santos Silva, Jonadab and Sizer, Edward and Kohli, Amit and Wynen, Maxence and Mullins, William A and Maggi, Pietro and Levy, Sarah and Onyemeh, Kamso and Ayci, Batuhan and Solomon, Andrew J and Assländer, Jakob and Al-Louzi, Omar and Reich, Daniel S and Sumowski, James and Beck, Erin S},
  title = {A Novel Convolutional Neural Network for Automated Multiple Sclerosis Brain Lesion Segmentation},
  journal = {Journal of Neuroimaging},
  year = {2025},
  volume = {35},
  number = {5},
  doi = {10.1111/jon.70085}
}
Kostenlos starten →