A deep learning framework deploying segment anything to detect pan-cancer mitotic figures from haematoxylin and eosin-stained slides.

Autor: Shen Z; Department of Medical Physics and Biomedical Engineering, University College London, London, UK. zhuoyan.shen.18@ucl.ac.uk., Simard M; Department of Medical Physics and Biomedical Engineering, University College London, London, UK., Brand D; Department of Medical Physics and Biomedical Engineering, University College London, London, UK.; Department of Radiotherapy, University College London Hospitals NHS Foundation Trust, London, UK., Andrei V; Research Department of Pathology, University College London Cancer Institute, London, UK.; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Al-Khader A; Research Department of Pathology, University College London Cancer Institute, London, UK.; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Oumlil F; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Trevers K; Research Department of Pathology, University College London Cancer Institute, London, UK.; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Butters T; Research Department of Pathology, University College London Cancer Institute, London, UK., Haefliger S; Research Department of Pathology, University College London Cancer Institute, London, UK.; Institute of Medical Genetics and Pathology, University Hospital Basel, University of Basel, Basel, CH, Switzerland., Kara E; Department of Neurology, Rutgers Biomedical and Health Sciences, Rutgers University, NJ, USA., Amary F; Research Department of Pathology, University College London Cancer Institute, London, UK.; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Tirabosco R; Research Department of Pathology, University College London Cancer Institute, London, UK.; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Cool P; Department of Orthopaedics, The Robert Jones and Agnes Hunt Orthopaedic Hospital, Oswestry, UK.; School of Medicine, Keele University, Newcastle, UK., Royle G; Department of Medical Physics and Biomedical Engineering, University College London, London, UK., Hawkins MA; Department of Medical Physics and Biomedical Engineering, University College London, London, UK.; Department of Radiotherapy, University College London Hospitals NHS Foundation Trust, London, UK., Flanagan AM; Research Department of Pathology, University College London Cancer Institute, London, UK.; Cellular and Molecular Pathology, Royal National Orthopaedic Hospital NHS Foundation Trust, Middlesex, UK., Collins-Fekete CA; Department of Medical Physics and Biomedical Engineering, University College London, London, UK.
Jazyk: angličtina
Zdroj: Communications biology [Commun Biol] 2024 Dec 19; Vol. 7 (1), pp. 1674. Date of Electronic Publication: 2024 Dec 19.
DOI: 10.1038/s42003-024-07398-6
Abstrakt: Mitotic activity is an important feature for grading several cancer types. However, counting mitotic figures (cells in division) is a time-consuming and laborious task prone to inter-observer variation. Inaccurate recognition of MFs can lead to incorrect grading and hence potential suboptimal treatment. This study presents an artificial intelligence-based approach to detect mitotic figures in digitised whole-slide images stained with haematoxylin and eosin. Advances in this area are hampered by the small size and variety of datasets available. To address this, we create the largest dataset of mitotic figures (N = 74,620), combining an in-house dataset of soft tissue tumours with five open-source datasets. We then employ a two-stage framework, named the Optimised Mitoses Generator Network (OMG-Net), to identify mitotic figures. This framework first deploys the Segment Anything Model to automatically outline cells, followed by an adapted ResNet18 that distinguishes mitotic figures. OMG-Net achieves an F1 score of 0.84 in detecting pan-cancer mitotic figures, including human breast carcinoma, neuroendocrine tumours, and melanoma. It outperforms previous state-of-the-art models in hold-out test sets. To summarise, our study introduces a generalisable data creation and curation pipeline and a high-performance detection model, which can largely contribute to the field of computer-aided mitotic figure detection.
Competing Interests: Competing interests: The authors declare no competing interests.
(© 2024. The Author(s).)
Databáze: MEDLINE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje