Machine Learning for Automatic Paraspinous Muscle Area and Attenuation Measures on Low-Dose Chest CT Scans

Ryan Barnard, Josh Tan, Brandon Roller, Caroline Chiles, Ashley A. Weaver, Robert D Boutin, Stephen B. Kritchevsky, Leon Lenchik

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Rationale and Objectives: To develop and evaluate an automated machine learning (ML) algorithm for segmenting the paraspinous muscles on chest computed tomography (CT) scans to evaluate for presence of sarcopenia. Materials and Methods: A convolutional neural network based on the U-Net architecture was trained to perform muscle segmentation on a dataset of 1875 single slice CT images and was tested on 209 CT images of participants in the National Lung Screening Trial. Low-dose, noncontrast CT examinations were obtained at 33 clinical sites, using scanners from four manufacturers. The study participants had a mean age of 71.6 years (range, 70–74 years). Ground truth was obtained by manually segmenting the left paraspinous muscle at the level of the T12 vertebra. Muscle cross-sectional area (CSA) and muscle attenuation (MA) were recorded. Comparison between the ML algorithm and ground truth measures of muscle CSA and MA were obtained using Dice similarity coefficients and Pearson correlations. Results: Compared to ground truth segmentation, the ML algorithm achieved median (standard deviation) Dice scores of 0.94 (0.04) in the test set. Mean (SD) muscle CSA was 14.3 (3.6) cm2 for ground truth and 13.7 (3.5) cm2 for ML segmentation. Mean (SD) MA was 41.6 (7.6) Hounsfield units (HU) for ground truth and 43.5 (7.9) HU for ML segmentation. There was high correlation between ML algorithm and ground truth for muscle CSA (r2 = 0.86; p < 0.0001) and MA (r2 = 0.95; p < 0.0001). Conclusion: The ML algorithm for measurement of paraspinous muscles compared favorably to manual ground truth measurements in the NLST. The algorithm generalized well to a heterogeneous set of low-dose CT images and may be capable of automated quantification of muscle metrics to screen for sarcopenia on routine chest CT examinations.

Original languageEnglish (US)
JournalAcademic radiology
DOIs
StatePublished - Jan 1 2019

Fingerprint

Thorax
Tomography
Muscles
Sarcopenia
Machine Learning
Spine
Lung

Keywords

  • Chest CT
  • Machine learning
  • Muscle
  • Myosteatosis
  • Sarcopenia

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

Machine Learning for Automatic Paraspinous Muscle Area and Attenuation Measures on Low-Dose Chest CT Scans. / Barnard, Ryan; Tan, Josh; Roller, Brandon; Chiles, Caroline; Weaver, Ashley A.; Boutin, Robert D; Kritchevsky, Stephen B.; Lenchik, Leon.

In: Academic radiology, 01.01.2019.

Research output: Contribution to journalArticle

Barnard, Ryan ; Tan, Josh ; Roller, Brandon ; Chiles, Caroline ; Weaver, Ashley A. ; Boutin, Robert D ; Kritchevsky, Stephen B. ; Lenchik, Leon. / Machine Learning for Automatic Paraspinous Muscle Area and Attenuation Measures on Low-Dose Chest CT Scans. In: Academic radiology. 2019.
@article{c12fdfca35894513b2480bb33625e451,
title = "Machine Learning for Automatic Paraspinous Muscle Area and Attenuation Measures on Low-Dose Chest CT Scans",
abstract = "Rationale and Objectives: To develop and evaluate an automated machine learning (ML) algorithm for segmenting the paraspinous muscles on chest computed tomography (CT) scans to evaluate for presence of sarcopenia. Materials and Methods: A convolutional neural network based on the U-Net architecture was trained to perform muscle segmentation on a dataset of 1875 single slice CT images and was tested on 209 CT images of participants in the National Lung Screening Trial. Low-dose, noncontrast CT examinations were obtained at 33 clinical sites, using scanners from four manufacturers. The study participants had a mean age of 71.6 years (range, 70–74 years). Ground truth was obtained by manually segmenting the left paraspinous muscle at the level of the T12 vertebra. Muscle cross-sectional area (CSA) and muscle attenuation (MA) were recorded. Comparison between the ML algorithm and ground truth measures of muscle CSA and MA were obtained using Dice similarity coefficients and Pearson correlations. Results: Compared to ground truth segmentation, the ML algorithm achieved median (standard deviation) Dice scores of 0.94 (0.04) in the test set. Mean (SD) muscle CSA was 14.3 (3.6) cm2 for ground truth and 13.7 (3.5) cm2 for ML segmentation. Mean (SD) MA was 41.6 (7.6) Hounsfield units (HU) for ground truth and 43.5 (7.9) HU for ML segmentation. There was high correlation between ML algorithm and ground truth for muscle CSA (r2 = 0.86; p < 0.0001) and MA (r2 = 0.95; p < 0.0001). Conclusion: The ML algorithm for measurement of paraspinous muscles compared favorably to manual ground truth measurements in the NLST. The algorithm generalized well to a heterogeneous set of low-dose CT images and may be capable of automated quantification of muscle metrics to screen for sarcopenia on routine chest CT examinations.",
keywords = "Chest CT, Machine learning, Muscle, Myosteatosis, Sarcopenia",
author = "Ryan Barnard and Josh Tan and Brandon Roller and Caroline Chiles and Weaver, {Ashley A.} and Boutin, {Robert D} and Kritchevsky, {Stephen B.} and Leon Lenchik",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.acra.2019.06.017",
language = "English (US)",
journal = "Academic Radiology",
issn = "1076-6332",
publisher = "Elsevier USA",

}

TY - JOUR

T1 - Machine Learning for Automatic Paraspinous Muscle Area and Attenuation Measures on Low-Dose Chest CT Scans

AU - Barnard, Ryan

AU - Tan, Josh

AU - Roller, Brandon

AU - Chiles, Caroline

AU - Weaver, Ashley A.

AU - Boutin, Robert D

AU - Kritchevsky, Stephen B.

AU - Lenchik, Leon

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Rationale and Objectives: To develop and evaluate an automated machine learning (ML) algorithm for segmenting the paraspinous muscles on chest computed tomography (CT) scans to evaluate for presence of sarcopenia. Materials and Methods: A convolutional neural network based on the U-Net architecture was trained to perform muscle segmentation on a dataset of 1875 single slice CT images and was tested on 209 CT images of participants in the National Lung Screening Trial. Low-dose, noncontrast CT examinations were obtained at 33 clinical sites, using scanners from four manufacturers. The study participants had a mean age of 71.6 years (range, 70–74 years). Ground truth was obtained by manually segmenting the left paraspinous muscle at the level of the T12 vertebra. Muscle cross-sectional area (CSA) and muscle attenuation (MA) were recorded. Comparison between the ML algorithm and ground truth measures of muscle CSA and MA were obtained using Dice similarity coefficients and Pearson correlations. Results: Compared to ground truth segmentation, the ML algorithm achieved median (standard deviation) Dice scores of 0.94 (0.04) in the test set. Mean (SD) muscle CSA was 14.3 (3.6) cm2 for ground truth and 13.7 (3.5) cm2 for ML segmentation. Mean (SD) MA was 41.6 (7.6) Hounsfield units (HU) for ground truth and 43.5 (7.9) HU for ML segmentation. There was high correlation between ML algorithm and ground truth for muscle CSA (r2 = 0.86; p < 0.0001) and MA (r2 = 0.95; p < 0.0001). Conclusion: The ML algorithm for measurement of paraspinous muscles compared favorably to manual ground truth measurements in the NLST. The algorithm generalized well to a heterogeneous set of low-dose CT images and may be capable of automated quantification of muscle metrics to screen for sarcopenia on routine chest CT examinations.

AB - Rationale and Objectives: To develop and evaluate an automated machine learning (ML) algorithm for segmenting the paraspinous muscles on chest computed tomography (CT) scans to evaluate for presence of sarcopenia. Materials and Methods: A convolutional neural network based on the U-Net architecture was trained to perform muscle segmentation on a dataset of 1875 single slice CT images and was tested on 209 CT images of participants in the National Lung Screening Trial. Low-dose, noncontrast CT examinations were obtained at 33 clinical sites, using scanners from four manufacturers. The study participants had a mean age of 71.6 years (range, 70–74 years). Ground truth was obtained by manually segmenting the left paraspinous muscle at the level of the T12 vertebra. Muscle cross-sectional area (CSA) and muscle attenuation (MA) were recorded. Comparison between the ML algorithm and ground truth measures of muscle CSA and MA were obtained using Dice similarity coefficients and Pearson correlations. Results: Compared to ground truth segmentation, the ML algorithm achieved median (standard deviation) Dice scores of 0.94 (0.04) in the test set. Mean (SD) muscle CSA was 14.3 (3.6) cm2 for ground truth and 13.7 (3.5) cm2 for ML segmentation. Mean (SD) MA was 41.6 (7.6) Hounsfield units (HU) for ground truth and 43.5 (7.9) HU for ML segmentation. There was high correlation between ML algorithm and ground truth for muscle CSA (r2 = 0.86; p < 0.0001) and MA (r2 = 0.95; p < 0.0001). Conclusion: The ML algorithm for measurement of paraspinous muscles compared favorably to manual ground truth measurements in the NLST. The algorithm generalized well to a heterogeneous set of low-dose CT images and may be capable of automated quantification of muscle metrics to screen for sarcopenia on routine chest CT examinations.

KW - Chest CT

KW - Machine learning

KW - Muscle

KW - Myosteatosis

KW - Sarcopenia

UR - http://www.scopus.com/inward/record.url?scp=85068990895&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068990895&partnerID=8YFLogxK

U2 - 10.1016/j.acra.2019.06.017

DO - 10.1016/j.acra.2019.06.017

M3 - Article

AN - SCOPUS:85068990895

JO - Academic Radiology

JF - Academic Radiology

SN - 1076-6332

ER -