Deep-learning-based Segmentation of Skeletal Muscle Mass in Routine Abdominal CT Scans.

Kreher, Robert; Hinnerichs, Mattes; Preim, Bernhard; Saalfeld, Sylvia; Surov, Alexey

doi:10.21873/invivo.12896

by Robert Kreher, Mattes Hinnerichs, Bernhard Preim, Sylvia Saalfeld, Alexey Surov

Abstract:

BACKGROUND: For prediction of many types of clinical outcome, the skeletal muscle mass can be used as an independent biomarker. Manual segmentation of the skeletal muscles is time-consuming, therefore we present a deeplearning-based approach for the identification of muscle mass at the L3 level in clinical routine computed tomographic (CT) data. PATIENTS AND METHODS: We conducted a retrospective study of 130 patient datasets. Individual CT slice analysis at the L3 level was fed into a U-Net architecture. As a result, we obtained segmentations of the musculus rectus abdominis, abdominal wall muscles, musculus psoas major, musculus quadratus lumborum and musculus erector spinae in the CT-slice at the L3 level. RESULTS: The Dice score was 0.95±0.02, 0.86±0.12, 0.93±0.05, 0.92±0.05, 0.86±0.08 for the erector spine, rectus, abdominal wall, psoas and quadratus lumborum muscles, respectively. For the overall skeletal muscle mass, the test data achieved a Dice score of 0.95±0.03. CONCLUSION: Our network achieved Dice scores larger than 0.86 for each of the five different muscle types and 0.95 for the overall skeletal muscle mass. The subdivision of muscle types can serve as a basis for obtaining future biomarkers. Our network is publicly available so that it might be beneficial for others to improve the clinical workflow within examination of routine CT scans.

Reference:

Deep-learning-based Segmentation of Skeletal Muscle Mass in Routine Abdominal CT Scans. (Robert Kreher, Mattes Hinnerichs, Bernhard Preim, Sylvia Saalfeld, Alexey Surov), In , International Journal of Experimental and Clinical Pathophysiology and Drug Research - In Vivo, volume 36, 2022.

Bibtex Entry:

@article{kreher_deep-learning-based_2022,
	title = {Deep-learning-based {Segmentation} of {Skeletal} {Muscle} {Mass} in {Routine} {Abdominal} {CT}  {Scans}.},
	volume = {36},
	copyright = {Copyright © 2022, International Institute of Anticancer Research (Dr. George J.  Delinasios), All rights reserved.},
	issn = {1791-7549 0258-851X 0258-851X},
	doi = {10.21873/invivo.12896},
	abstract = {BACKGROUND: For prediction of many types of clinical outcome, the skeletal muscle  mass can be used as an independent biomarker. Manual segmentation of the skeletal  muscles is time-consuming, therefore we present a deeplearning-based approach for  the identification of muscle mass at the L3 level in clinical routine computed  tomographic (CT) data. PATIENTS AND METHODS: We conducted a retrospective study  of 130 patient datasets. Individual CT slice analysis at the L3 level was fed  into a U-Net architecture. As a result, we obtained segmentations of the musculus  rectus abdominis, abdominal wall muscles, musculus psoas major, musculus  quadratus lumborum and musculus erector spinae in the CT-slice at the L3 level.  RESULTS: The Dice score was 0.95±0.02, 0.86±0.12, 0.93±0.05, 0.92±0.05, 0.86±0.08  for the erector spine, rectus, abdominal wall, psoas and quadratus lumborum  muscles, respectively. For the overall skeletal muscle mass, the test data  achieved a Dice score of 0.95±0.03. CONCLUSION: Our network achieved Dice scores  larger than 0.86 for each of the five different muscle types and 0.95 for the  overall skeletal muscle mass. The subdivision of muscle types can serve as a  basis for obtaining future biomarkers. Our network is publicly available so that  it might be beneficial for others to improve the clinical workflow within  examination of routine CT scans.},
	language = {eng},
	number = {4},
	journal = {, International Journal of Experimental and Clinical Pathophysiology and Drug Research - In Vivo},
	author = {Kreher, Robert and Hinnerichs, Mattes and Preim, Bernhard and Saalfeld, Sylvia and Surov, Alexey},
	month = aug,
	year = {2022},
	pmid = {35738592},
	pmcid = {PMC9301401},
	keywords = {*Deep Learning, Abdomen, deep-learning segmentation, Humans, Muscle, Skeletal/diagnostic imaging, Psoas Muscles/diagnostic imaging, Retrospective Studies, sarcopenia, Skeletal muscle mass, Tomography, X-Ray Computed},
	pages = {1807--1811}
}