A systematic review of automated segmentation of 3D computed-tomography scans for volumetric body composition analysis

被引:23
|
作者
Mai, Dinh Van Chi [1 ,2 ,7 ]
Drami, Ioanna T. [1 ,3 ]
Pring, Edward E. [1 ,2 ]
Gould, Laura [1 ,4 ]
Lung, Phillip [1 ,2 ]
Popuri, Karteek [5 ]
Chow, Vincent F. [6 ]
Beg, Mirza [6 ]
Athanasiou, Thanos T. [2 ]
Jenkins, John [1 ,2 ]
机构
[1] St Marks Hosp, St Marks Acad Inst, Dept Surg, London, England
[2] Imperial Coll, Dept Surg & Canc, London, England
[3] Imperial Coll, Dept Metab Digest & Reprod, London, England
[4] Univ Glasgow, Coll Med Vet & Life Sci, Sch Canc Sci, Glasgow, Scotland
[5] Mem Univ Newfoundland, Dept Comp Sci, St John, NF, Canada
[6] Simon Fraser Univ, Sch Engn Sci, Burnaby, BC, Canada
[7] Northwick Pk Hosp & Clin Res Ctr, St Marks Hosp, Dept Colorectal Surg, Harrow Rd, London HA1 3UJ, England
关键词
AI; Body composition measurement; Computed tomography; Deep learning; Sarcopenia; Segmentation; SUBCUTANEOUS ADIPOSE-TISSUE; SKELETAL-MUSCLE; ARTIFICIAL-INTELLIGENCE; SARCOPENIC OBESITY; PROGNOSTIC-FACTOR; CANCER; VALIDATION; IMAGES; PREDICTION; MORTALITY;
D O I
10.1002/jcsm.13310
中图分类号
R592 [老年病学]; C [社会科学总论];
学科分类号
03 ; 0303 ; 100203 ;
摘要
Automated computed tomography (CT) scan segmentation (labelling of pixels according to tissue type) is now possible. This technique is being adapted to achieve three-dimensional (3D) segmentation of CT scans, opposed to single L3-slice alone. This systematic review evaluates feasibility and accuracy of automated segmentation of 3D CT scans for volumetric body composition (BC) analysis, as well as current limitations and pitfalls clinicians and researchers should be aware of. OVID Medline, Embase and grey literature databases up to October 2021 were searched. Original studies investigating automated skeletal muscle, visceral and subcutaneous AT segmentation from CT were included. Seven of the 92 studies met inclusion criteria. Variation existed in expertise and numbers of humans performing ground-truth segmentations used to train algorithms. There was heterogeneity in patient characteristics, pathology and CT phases that segmentation algorithms were developed upon. Reporting of anatomical CT coverage varied, with confusing terminology. Six studies covered volumetric regional slabs rather than the whole body. One study stated the use of whole-body CT, but it was not clear whether this truly meant head-to-fingertip-to-toe. Two studies used conventional computer algorithms. The latter five used deep learning (DL), an artificial intelligence technique where algorithms are similarly organized to brain neuronal pathways. Six of seven reported excellent segmentation performance (Dice similarity coefficients > 0.9 per tissue). Internal testing on unseen scans was performed for only four of seven algorithms, whilst only three were tested externally. Trained DL algorithms achieved full CT segmentation in 12 to 75 s versus 25 min for non-DL techniques. DL enables opportunistic, rapid and automated volumetric BC analysis of CT performed for clinical indications. However, most CT scans do not cover head-to-fingertip-to-toe; further research must validate using common CT regions to estimate true whole-body BC, with direct comparison to single lumbar slice. Due to successes of DL, we expect progressive numbers of algorithms to materialize in addition to the seven discussed in this paper. Researchers and clinicians in the field of BC must therefore be aware of pitfalls. High Dice similarity coefficients do not inform the degree to which BC tissues may be under- or overestimated and nor does it inform on algorithm precision. Consensus is needed to define accuracy and precision standards for ground-truth labelling. Creation of a large international, multicentre common CT dataset with BC ground-truth labels from multiple experts could be a robust solution.
引用
收藏
页码:1973 / 1986
页数:14
相关论文
共 50 条
  • [41] A machine learning approach relating 3D body scans to body composition in humans
    James D. Pleuss
    Kevin Talty
    Steven Morse
    Patrick Kuiper
    Michael Scioletti
    Steven B. Heymsfield
    Diana M. Thomas
    European Journal of Clinical Nutrition, 2019, 73 : 200 - 208
  • [42] Validation of an automated segmentation method for body composition analysis in colorectal cancer patients using diagnostic abdominal computed tomography images
    Querido, Nadira R.
    Bours, Martijn J. L.
    Brecheisen, Ralph
    Iersel, Liselot Valkenburg-van
    Breukink, Stephanie O.
    Janssen-Heijnen, Maryska L. G.
    Keulen, Eric T. P.
    Konsten, Joop L. M.
    de Vos-Geelen, Judith
    Weijenberg, Matty P.
    Simons, Colinda C. J. M.
    CLINICAL NUTRITION ESPEN, 2024, 63 : 659 - 667
  • [43] Automatic segmentation of the mandible from computed tomography scans for 3D virtual surgical planning using the convolutional neural network
    Qiu, Bingjiang
    Guo, Jiapan
    Kraeima, Joep
    Glas, Haye H.
    Borra, Ronald J. H.
    Witjes, M. J. H.
    van Ooijen, Peter M. A.
    PHYSICS IN MEDICINE AND BIOLOGY, 2019, 64 (17):
  • [44] Fully automated body composition analysis in routine CT imaging using 3D semantic segmentation convolutional neural networks
    Sven Koitka
    Lennard Kroll
    Eugen Malamutmann
    Arzu Oezcelik
    Felix Nensa
    European Radiology, 2021, 31 : 1795 - 1804
  • [45] Fully automated body composition analysis in routine CT imaging using 3D semantic segmentation convolutional neural networks
    Koitka, Sven
    Kroll, Lennard
    Malamutmann, Eugen
    Oezcelik, Arzu
    Nensa, Felix
    EUROPEAN RADIOLOGY, 2021, 31 (04) : 1795 - 1804
  • [46] Utility of Computed Tomography Body Composition Measures in Chronic Lung Disease: Systematic Review
    Rozenberg, D.
    Orsso, C. E.
    Chohan, K.
    Orchanian-Cheff, A.
    Fidler, L.
    Vagaon, A.
    Singer, L. G.
    Mathur, S.
    AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE, 2019, 199
  • [47] The cranial capacity of the Saudi population measured using 3D computed tomography scans
    Alshamrani, Khalaf A.
    NEUROSCIENCES, 2023, 28 (03) : 184 - 189
  • [48] Automated volume analysis of head and neck lesions on CT scans using 3D level set segmentation
    Street, Ethan
    Hadjiiski, Lubomir
    Sahiner, Berkman
    Gujar, Sachin
    Ibrahim, Mohannad
    Mukherji, Suresh K.
    Chan, Heang-Ping
    MEDICAL PHYSICS, 2007, 34 (11) : 4399 - 4408
  • [49] Automated thoracic and abdominal computed tomography segmentation using the belief functions theory:: Application to 3D internal dosimetry
    Dieudonne, A.
    Zhang, P.
    Vannoorenberghe, P.
    Gardin, I.
    EUROPEAN JOURNAL OF NUCLEAR MEDICINE AND MOLECULAR IMAGING, 2006, 33 : S374 - S374
  • [50] Automated thoracic and abdominal computed tomography segmentation using the belief functions theory:: application to 3D internal dosimetry
    Dieudonne, A.
    Zhang, P.
    Vannoorenberghe, P.
    Gardin, I.
    EUROPEAN JOURNAL OF NUCLEAR MEDICINE AND MOLECULAR IMAGING, 2006, 33 : S179 - S179