Journal Article (Review Article)

DKFZ-2024-02143

http://join2-wiki.gsi.de/foswiki/pub/Main/Artwork/join2_logo100x88.png Unleashing the strengths of unlabelled data in deep learning-assisted pan-cancer abdominal organ quantification: the FLARE22 challenge.

Ma, J. ; Zhang, Y. ; Gu, S. ; Ge, C. ; Mae, S. ; Young, A. ; Zhu, C. ; Yang, X. ; Meng, K. ; Huang, Z. ; Zhang, F. ; Pan, Y. ; Huang, S. ; Wang, J. ; Sun, M. ; Zhang, R. ; Jia, D. ; Choi, J. W. ; Alves, N. ; de Wilde, B. ; Koehler, G.DKFZ* ; Lai, H. ; Wang, E. ; Wiesenfarth, M.DKFZ* ; Zhu, Q. ; Dong, G. ; He, J. ; Consortium, F. C. (Collaboration Author) ; Wang, B. ; He, J. (Contributor) ; Yang, H. (Contributor) ; Huang, B. (Contributor) ; Lyu, M. (Contributor) ; Ma, Y. (Contributor) ; Guo, H. (Contributor) ; Xu, W. (Contributor) ; Maier-Hein, K.DKFZ* ; Wu, Y. (Contributor)

2024
The Lancet London

This record in other databases:  

Please use a persistent id in citations: doi:10.1016/S2589-7500(24)00154-7

Abstract: Deep learning has shown great potential to automate abdominal organ segmentation and quantification. However, most existing algorithms rely on expert annotations and do not have comprehensive evaluations in real-world multinational settings. To address these limitations, we organised the FLARE 2022 challenge to benchmark fast, low-resource, and accurate abdominal organ segmentation algorithms. We first constructed an intercontinental abdomen CT dataset from more than 50 clinical research groups. We then independently validated that deep learning algorithms achieved a median dice similarity coefficient (DSC) of 90·0% (IQR 87·4-91·3%) by use of 50 labelled images and 2000 unlabelled images, which can substantially reduce manual annotation costs. The best-performing algorithms successfully generalised to holdout external validation sets, achieving a median DSC of 89·4% (85·2-91·3%), 90·0% (84·3-93·0%), and 88·5% (80·9-91·9%) on North American, European, and Asian cohorts, respectively. These algorithms show the potential to use unlabelled data to boost performance and alleviate annotation shortages for modern artificial intelligence models.

Keyword(s): Deep Learning (MeSH) ; Humans (MeSH) ; Tomography, X-Ray Computed (MeSH) ; Algorithms (MeSH) ; Abdomen: diagnostic imaging (MeSH)

---

Contributing Institute(s):

1. E230 Medizinische Bildverarbeitung (E230)
2. C060 Biostatistik (C060)

Research Program(s):

1. 315 - Bildgebung und Radioonkologie (POF4-315) (POF4-315)

Appears in the scientific report 2024

---

Database coverage:
; ; Article Processing Charges ; Clarivate Analytics Master Journal List ; Current Contents - Clinical Medicine ; DOAJ Seal ; Essential Science Indicators ; Fees ; IF >= 30 ; JCR ; SCOPUS ; Science Citation Index Expanded ; Web of Science Core Collection

---