Sven Koitka1, Lennard Kroll2, Eugen Malamutmann3, Arzu Oezcelik3, Felix Nensa2. 1. Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Essen, Germany. sven.koitka@uk-essen.de. 2. Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Essen, Germany. 3. Department of General, Visceral and Transplantation Surgery, University Hospital Essen, Essen, Germany.
Abstract
OBJECTIVES: Body tissue composition is a long-known biomarker with high diagnostic and prognostic value not only in cardiovascular, oncological, and orthopedic diseases but also in rehabilitation medicine or drug dosage. In this study, the aim was to develop a fully automated, reproducible, and quantitative 3D volumetry of body tissue composition from standard CT examinations of the abdomen in order to be able to offer such valuable biomarkers as part of routine clinical imaging. METHODS: Therefore, an in-house dataset of 40 CTs for training and 10 CTs for testing were fully annotated on every fifth axial slice with five different semantic body regions: abdominal cavity, bones, muscle, subcutaneous tissue, and thoracic cavity. Multi-resolution U-Net 3D neural networks were employed for segmenting these body regions, followed by subclassifying adipose tissue and muscle using known Hounsfield unit limits. RESULTS: The Sørensen Dice scores averaged over all semantic regions was 0.9553 and the intra-class correlation coefficients for subclassified tissues were above 0.99. CONCLUSIONS: Our results show that fully automated body composition analysis on routine CT imaging can provide stable biomarkers across the whole abdomen and not just on L3 slices, which is historically the reference location for analyzing body composition in the clinical routine. KEY POINTS: • Our study enables fully automated body composition analysis on routine abdomen CT scans. • The best segmentation models for semantic body region segmentation achieved an averaged Sørensen Dice score of 0.9553. • Subclassified tissue volumes achieved intra-class correlation coefficients over 0.99.
OBJECTIVES: Body tissue composition is a long-known biomarker with high diagnostic and prognostic value not only in cardiovascular, oncological, and orthopedic diseases but also in rehabilitation medicine or drug dosage. In this study, the aim was to develop a fully automated, reproducible, and quantitative 3D volumetry of body tissue composition from standard CT examinations of the abdomen in order to be able to offer such valuable biomarkers as part of routine clinical imaging. METHODS: Therefore, an in-house dataset of 40 CTs for training and 10 CTs for testing were fully annotated on every fifth axial slice with five different semantic body regions: abdominal cavity, bones, muscle, subcutaneous tissue, and thoracic cavity. Multi-resolution U-Net 3D neural networks were employed for segmenting these body regions, followed by subclassifying adipose tissue and muscle using known Hounsfield unit limits. RESULTS: The Sørensen Dice scores averaged over all semantic regions was 0.9553 and the intra-class correlation coefficients for subclassified tissues were above 0.99. CONCLUSIONS: Our results show that fully automated body composition analysis on routine CT imaging can provide stable biomarkers across the whole abdomen and not just on L3 slices, which is historically the reference location for analyzing body composition in the clinical routine. KEY POINTS: • Our study enables fully automated body composition analysis on routine abdomen CT scans. • The best segmentation models for semantic body region segmentation achieved an averaged Sørensen Dice score of 0.9553. • Subclassified tissue volumes achieved intra-class correlation coefficients over 0.99.
Entities:
Keywords:
Abdomen; Body composition; Computer-assisted image analysis; Deep learning
Authors: Alexander D Weston; Panagiotis Korfiatis; Timothy L Kline; Kenneth A Philbrick; Petro Kostandy; Tomas Sakinis; Motokazu Sugimoto; Naoki Takahashi; Bradley J Erickson Journal: Radiology Date: 2018-12-11 Impact factor: 11.105
Authors: David Zopfs; Sebastian Theurich; Nils Große Hokamp; Jana Knuever; Lukas Gerecht; Jan Borggrefe; Max Schlaak; Daniel Pinto Dos Santos Journal: Eur Radiol Date: 2019-11-27 Impact factor: 5.315
Authors: J Aubrey; N Esfandiari; V E Baracos; F A Buteau; J Frenette; C T Putman; V C Mazurak Journal: Acta Physiol (Oxf) Date: 2014-03 Impact factor: 6.311
Authors: Joel Kullberg; Anders Hedström; John Brandberg; Robin Strand; Lars Johansson; Göran Bergström; Håkan Ahlström Journal: Sci Rep Date: 2017-09-05 Impact factor: 4.379
Authors: Federico Greco; Rodrigo Salgado; Wim Van Hecke; Romualdo Del Buono; Paul M Parizel; Carlo Augusto Mallio Journal: Quant Imaging Med Surg Date: 2022-03
Authors: Robin F Gohmann; Batuhan Temiz; Patrick Seitz; Sebastian Gottschling; Christian Lücke; Christian Krieghoff; Christian Blume; Matthias Horn; Matthias Gutberlet Journal: Quant Imaging Med Surg Date: 2021-10
Authors: Thomas Ying; Pablo Borrelli; Lars Edenbrandt; Olof Enqvist; Reza Kaboteh; Elin Trägårdh; Johannes Ulén; Henrik Kjölhede Journal: Eur Radiol Exp Date: 2021-11-19
Authors: David Zopfs; Daniel Pinto Dos Santos; Jonathan Kottlors; Robert P Reimer; Simon Lennartz; Roman Kloeckner; Max Schlaak; Sebastian Theurich; Christoph Kabbasch; Marc Schlamann; Nils Große Hokamp Journal: Eur Radiol Date: 2022-04-07 Impact factor: 7.034