1
|
Dabiri S, Popuri K, Cespedes Feliciano EM, Caan BJ, Baracos VE, Beg MF. Muscle segmentation in axial computed tomography (CT) images at the lumbar (L3) and thoracic (T4) levels for body composition analysis. Comput Med Imaging Graph 2019; 75:47-55. [PMID: 31132616 PMCID: PMC6620151 DOI: 10.1016/j.compmedimag.2019.04.007] [Citation(s) in RCA: 51] [Impact Index Per Article: 10.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/27/2018] [Revised: 04/07/2019] [Accepted: 04/29/2019] [Indexed: 01/06/2023]
Abstract
In diseases such as cancer, patients suffer from degenerative loss of skeletal muscle (cachexia). Muscle wasting and loss of muscle function/performance (sarcopenia) can also occur during advanced aging. Assessing skeletal muscle mass in sarcopenia and cachexia is therefore of clinical interest for risk stratification. In comparison with fat, body fluids and bone, quantifying the skeletal muscle mass is more challenging. Computed tomography (CT) is one of the gold standard techniques for cancer diagnostics and analysis of progression, and therefore a valuable source of imaging for in vivo quantification of skeletal muscle mass. In this paper, we design a novel deep neural network-based algorithm for the automated segmentation of skeletal muscle in axial CT images at the third lumbar (L3) and the fourth thoracic (T4) levels. A two-branch network with two training steps is investigated. The network's performance is evaluated for three trained models on separate datasets. These datasets were generated by different CT devices and data acquisition settings. To ensure the model's robustness, each trained model was tested on all three available test sets. Errors and the effect of labeling protocol in these cases were analyzed and reported. The best performance of the proposed algorithm was achieved on 1327 L3 test samples with an overlap Jaccard score of 98% and sensitivity and specificity greater than 99%.
Collapse
Affiliation(s)
- Setareh Dabiri
- School of Engineering Science, Simon Fraser University, Canada.
| | - Karteek Popuri
- School of Engineering Science, Simon Fraser University, Canada
| | | | - Bette J Caan
- Division of Research, Kaiser Permanente Northern California, USA
| | | | | |
Collapse
|
2
|
Abstract
BACKGROUND Analytic morphomics, or more simply, "morphomics," refers to the measurement of specific biomarkers of body composition from medical imaging, most commonly computed tomography (CT) images. An emerging body of literature supports the use of morphomic markers measured on single-slice CT images for risk prediction in a range of clinical populations. However, uptake by healthcare providers been limited due to the lack of clinician-friendly software to facilitate measurements. The objectives of this study were to describe the interface and functionality of CoreSlicer- a free and open-source web-based interface aiming to facilitate measurement of analytic morphomics by clinicians - and to validate muscle and fat measurements performed in CoreSlicer against reference software. RESULTS Measurements of muscle and fat obtained in CoreSlicer show high agreement with established reference software. CoreSlicer features a full set of DICOM viewing tools and extensible plugin interface to facilitate rapid prototyping and validation of new morphomic markers by researchers. We present published studies illustrating the use of CoreSlicer by clinicians with no prior knowledge of medical image segmentation techniques and no formal training in radiology, where CoreSlicer was successfully used to predict operative risk in three distinct populations of cardiovascular patients. CONCLUSIONS CoreSlicer enables extraction of morphomic markers from CT images by non-technically skilled clinicians. Measurements were reproducible and accurate in relation to reference software.
Collapse
Affiliation(s)
- Louis Mullie
- Department of Medicine, McGill University, 3755 Cote Ste Catherine Rd, E-222, Montreal, QC, H3T 1E2, Canada
- Division of Cardiology, McGill University, Montreal, QC, Canada
| | - Jonathan Afilalo
- Department of Medicine, McGill University, 3755 Cote Ste Catherine Rd, E-222, Montreal, QC, H3T 1E2, Canada.
- Division of Cardiology, McGill University, Montreal, QC, Canada.
- Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, McGill University, Montreal, QC, Canada.
| |
Collapse
|
3
|
Kamiya N, Li J, Kume M, Fujita H, Shen D, Zheng G. Fully automatic segmentation of paraspinal muscles from 3D torso CT images via multi-scale iterative random forest classifications. Int J Comput Assist Radiol Surg 2018; 13:1697-1706. [PMID: 30173335 DOI: 10.1007/s11548-018-1852-1] [Citation(s) in RCA: 16] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/12/2018] [Accepted: 08/27/2018] [Indexed: 01/10/2023]
Abstract
PURPOSE To develop and validate a fully automatic method for segmentation of paraspinal muscles from 3D torso CT images. METHODS We propose a novel learning-based method to address this challenging problem. Multi-scale iterative random forest classifications with multi-source information are employed in this study to speed up the segmentation and to improve the accuracy. Here, multi-source images include the original torso CT images and later also the iteratively estimated and refined probability maps of the paraspinal muscles. We validated our method on 20 torso CT data with associated manual segmentation. We randomly partitioned the 20 CT data into two evenly distributed groups and took one group as the training data and the other group as the test data. RESULTS The proposed method achieved a mean Dice coefficient of 93.0%. It took on average 46.5 s to segment a 3D torso CT image with the size ranging from [Formula: see text] voxels to [Formula: see text] voxels. CONCLUSIONS Our fully automatic, learning-based method can accurately segment paraspinal muscles from 3D torso CT images. It generates segmentation results that are better than those achieved by the state-of-the-art methods.
Collapse
Affiliation(s)
- Naoki Kamiya
- School of Information Science and Technology, Aichi Prefectural University, Nagakute, Japan
| | - Jing Li
- Institute for Surgical Technology and Biomechanics, University of Bern, Bern, Switzerland
| | - Masanori Kume
- Department of Electrical, Electronic and Computer Engineering, Faculty of Engineering, Gifu University, Gifu, Japan
| | - Hiroshi Fujita
- Department of Electrical, Electronic and Computer Engineering, Faculty of Engineering, Gifu University, Gifu, Japan
| | - Dinggang Shen
- Department of Brain and Cognitive Engineering, Korea University, Seoul, 02841, Republic of Korea.
| | - Guoyan Zheng
- Institute for Surgical Technology and Biomechanics, University of Bern, Bern, Switzerland.
| |
Collapse
|
4
|
Lee H, Troschel FM, Tajmir S, Fuchs G, Mario J, Fintelmann FJ, Do S. Pixel-Level Deep Segmentation: Artificial Intelligence Quantifies Muscle on Computed Tomography for Body Morphometric Analysis. J Digit Imaging 2018; 30:487-498. [PMID: 28653123 PMCID: PMC5537099 DOI: 10.1007/s10278-017-9988-z] [Citation(s) in RCA: 108] [Impact Index Per Article: 18.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/06/2023] Open
Abstract
Pretreatment risk stratification is key for personalized medicine. While many physicians rely on an “eyeball test” to assess whether patients will tolerate major surgery or chemotherapy, “eyeballing” is inherently subjective and difficult to quantify. The concept of morphometric age derived from cross-sectional imaging has been found to correlate well with outcomes such as length of stay, morbidity, and mortality. However, the determination of the morphometric age is time intensive and requires highly trained experts. In this study, we propose a fully automated deep learning system for the segmentation of skeletal muscle cross-sectional area (CSA) on an axial computed tomography image taken at the third lumbar vertebra. We utilized a fully automated deep segmentation model derived from an extended implementation of a fully convolutional network with weight initialization of an ImageNet pre-trained model, followed by post processing to eliminate intramuscular fat for a more accurate analysis. This experiment was conducted by varying window level (WL), window width (WW), and bit resolutions in order to better understand the effects of the parameters on the model performance. Our best model, fine-tuned on 250 training images and ground truth labels, achieves 0.93 ± 0.02 Dice similarity coefficient (DSC) and 3.68 ± 2.29% difference between predicted and ground truth muscle CSA on 150 held-out test cases. Ultimately, the fully automated segmentation system can be embedded into the clinical environment to accelerate the quantification of muscle and expanded to volume analysis of 3D datasets.
Collapse
Affiliation(s)
- Hyunkwang Lee
- Department of Radiology, Massachusetts General Hospital, 25 New Chardon Street, Suite 400B, Boston, MA 02114 USA
| | - Fabian M. Troschel
- Department of Radiology, Massachusetts General Hospital, 25 New Chardon Street, Suite 400B, Boston, MA 02114 USA
| | - Shahein Tajmir
- Department of Radiology, Massachusetts General Hospital, 25 New Chardon Street, Suite 400B, Boston, MA 02114 USA
| | - Georg Fuchs
- Department of Radiology, Charite - Universitaetsmedizin Berlin, Chariteplatz 1, 10117 Berlin, Germany
| | - Julia Mario
- Department of Radiology, Massachusetts General Hospital, 25 New Chardon Street, Suite 400B, Boston, MA 02114 USA
| | - Florian J. Fintelmann
- Department of Radiology, Massachusetts General Hospital, 25 New Chardon Street, Suite 400B, Boston, MA 02114 USA
| | - Synho Do
- Department of Radiology, Massachusetts General Hospital, 25 New Chardon Street, Suite 400B, Boston, MA 02114 USA
| |
Collapse
|
5
|
Popuri K, Cobzas D, Esfandiari N, Baracos V, Jägersand M. Body Composition Assessment in Axial CT Images Using FEM-Based Automatic Segmentation of Skeletal Muscle. IEEE TRANSACTIONS ON MEDICAL IMAGING 2016; 35:512-520. [PMID: 26415164 DOI: 10.1109/tmi.2015.2479252] [Citation(s) in RCA: 89] [Impact Index Per Article: 11.1] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 06/05/2023]
Abstract
The proportions of muscle and fat tissues in the human body, referred to as body composition is a vital measurement for cancer patients. Body composition has been recently linked to patient survival and the onset/recurrence of several types of cancers in numerous cancer research studies. This paper introduces a fully automatic framework for the segmentation of muscle and fat tissues from CT images to estimate body composition. We developed a novel finite element method (FEM) deformable model that incorporates a priori shape information via a statistical deformation model (SDM) within the template-based segmentation framework. The proposed method was validated on 1000 abdominal and 530 thoracic CT images and we obtained very good segmentation results with Jaccard scores in excess of 90% for both the muscle and fat regions.
Collapse
|
6
|
Kamiya N, Zhou X, Chen H, Muramatsu C, Hara T, Yokoyama R, Kanematsu M, Hoshi H, Fujita H. Automated segmentation of recuts abdominis muscle using shape model in X-ray CT images. ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY. ANNUAL INTERNATIONAL CONFERENCE 2012; 2011:7993-6. [PMID: 22256195 DOI: 10.1109/iembs.2011.6091971] [Citation(s) in RCA: 9] [Impact Index Per Article: 0.8] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/07/2022]
Abstract
Our purpose in this study is to segment the rectus abdominis muscle region in X-ray CT images, and we propose a novel recognition method based on the shape model. In this method, three steps are included in the segmentation process. The first is to generate a shape model for the rectus abdominis muscle. The second is to recognize anatomical feature points corresponding to the origin and insertion of the muscle, and the third is to segment the rectus abdominis muscles based on the shape model. We generated the shape model from 20 CT cases and tested the model to recognize the muscle in 20 other CT cases. The average values for the Jaccard similarity coefficient (JSC) and true segmentation coefficient (TSC) were 0.841 and 0.863, respectively. The results suggest the validity of the model-based segmentation for the rectus abdominis muscle.
Collapse
Affiliation(s)
- N Kamiya
- Department of Intelligent Image Information, Division of Regeneration and Advanced Medical Sciences, Graduate School of Medicine, Gifu University, Yanagido 1-1, Gifu 501-1194, Japan.
| | | | | | | | | | | | | | | | | |
Collapse
|