EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Positional contrastive learning for improved thigh muscle segmentation in MR images

Nicola Casali, Elisa Scalco iD, Maria Giovanna Taccogna, Fulvio Lauretani, Simone Porcelli iD, Andrea Ciuni, Alfonso Mastropietro iD, Giovanna Rizzo iD

NMR in Biomedicine · 2024

DOI: 10.1002/nbm.5197 ↗

Kurz gefasst

Fachartikel in der Zeitschrift NMR in Biomedicine, 2024. Verfasst von Nicola Casali und 7 weiteren Autor:innen. Beteiligte Einrichtung: Institute of Intelligent Industrial Technologies and Systems for Advanced Manufacturing, National Research Council, Milan, Italy. Hauptthemen laut MEDLINE: Magnetic Resonance Imaging, Thigh, Muscle, Skeletal. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministero dell'Università e della Ricerca.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The accurate segmentation of individual muscles is essential for quantitative MRI analysis of thigh images. Deep learning methods have achieved state‐of‐the‐art results in segmentation, but they require large numbers of labeled data to perform well. However, labeling individual thigh muscles slice by slice for numerous volumes is a laborious and time‐consuming task, which limits the availability of annotated datasets. To address this challenge, self‐supervised learning (SSL) emerges as a promising technique to enhance model performance by pretraining the model on unlabeled data. A recent approach, called positional contrastive learning, exploits the information given by the axial position of the slices to learn features transferable on the segmentation task. The aim of this work was to propose positional contrastive SSL for the segmentation of individual thigh muscles from MRI acquisitions in a population of elderly healthy subjects and to evaluate it on different levels of limited annotated data. An unlabeled dataset of 72 T1w MRI thigh acquisitions was available for SSL pretraining, while a labeled dataset of 52 volumes was employed for the final segmentation task, split into training and test sets. The effectiveness of SSL pretraining to fine‐tune a U‐Net architecture for thigh muscle segmentation was compared with that of a randomly initialized model (RND), considering an increasing number of annotated volumes ( S = 1, 2, 5, 10, 20, 30, 40). Our results demonstrated that SSL yields substantial improvements in Dice similarity coefficient (DSC) when using a very limited number of labeled volumes (e.g., for = 1, DSC 0.631 versus 0.530 for SSL and RND, respectively). Moreover, enhancements are achievable even when utilizing the full number of labeled subjects, with DSC = 0.927 for SSL and 0.924 for RND. In conclusion, positional contrastive SSL was effective in obtaining more accurate thigh muscle segmentation, even with a very low number of labeled data, with a potential impact of speeding up the annotation process in clinics.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Magnetic Resonance ImagingThighMuscle, SkeletalHumansMaleFemaleAgedImage Processing, Computer-AssistedSupervised Machine LearningMiddle Aged

Stichwörter der Autor:innen

deep learningmusculoskeletal MRIself‐supervised learningthigh muscle segmentation

Autor:innen und Institutionen

  • Institute of Intelligent Industrial Technologies and Systems for Advanced Manufacturing, National Research Council, Milan, Italy.
  • Department of Electronics, Information and Bioengineering, Politecnico di Milano, Milan, Italy.
  • Institute of Biomedical Technologies, National Research Council, Segrate, Italy.
  • Department of Medicine and Surgery, University of Parma, Parma, Italy.
  • Geriatric Clinic Unit, Geriatric-Rehabilitation Department, Parma University Hospital, Parma, Italy.
  • Department of Molecular Medicine, University of Pavia, Pavia, Italy.

ORCID-Profile: Elisa Scalco, Simone Porcelli, Alfonso Mastropietro, Giovanna Rizzo

Förderung

  • Ministero dell'Università e della Ricerca

Zitieren

APA

Casali, N., Scalco, E., Taccogna, M. G., Lauretani, F., Porcelli, S., Ciuni, A., Mastropietro, A., & Rizzo, G. (2024). Positional contrastive learning for improved thigh muscle segmentation in MR images. NMR in Biomedicine. https://doi.org/10.1002/nbm.5197

Vancouver

Casali N, Scalco E, Taccogna MG, Lauretani F, Porcelli S, Ciuni A, et al. Positional contrastive learning for improved thigh muscle segmentation in MR images. NMR in Biomedicine. 2024. doi:10.1002/nbm.5197

BibTeX
@article{Casali2024,
  author = {Casali, Nicola and Scalco, Elisa and Taccogna, Maria Giovanna and Lauretani, Fulvio and Porcelli, Simone and Ciuni, Andrea and Mastropietro, Alfonso and Rizzo, Giovanna},
  title = {Positional contrastive learning for improved thigh muscle segmentation in MR images},
  journal = {NMR in Biomedicine},
  year = {2024},
  doi = {10.1002/nbm.5197}
}
Kostenlos starten →