Automatic Paraspinal Muscle Segmentation in Patients with Lumbar Pathology Using Deep Convolutional Neural Network

Xia, Wenyao; Fortin, Maryse; Ahn, Joshua; Rivaz, Hassan; Battié, Michele C.; Peters, Terry M.; Xiao, Yiming

doi:10.1007/978-3-030-32245-8_36

Cited by 11 publications

(5 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the future, validation of this technique in paraspinal muscle morphometry with respect to larger patient populations, various lumbar pathologies, and additional spinal levels of pathology will be important. For this, automatic image segmentation algorithms based on deep learning techniques 20 will be highly instrumental. Furthermore, intrinsic side-to-side anatomical differences may exist beyond the effects of sex and age and contribute to the identified shape asymmetry.…”

Section: Discussionmentioning

confidence: 99%

“…Cross-sectional areas were computed using manual segmentation of the left and right multifidus and erector spinae muscles. By using manually segmented muscle CSA to constrain the regions of interest, fatty cross-sectional areas were segmented with the k-means algorithm 20 . The algorithm can automatically separate fat and lean muscle based on their intensity difference.…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Statistical morphological analysis reveals characteristic paraspinal muscle asymmetry in unilateral lumbar disc herniation

Xiao

Fortin

Ahn

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

Growing evidence suggests an association of lumbar paraspinal muscle morphology with low back pain (LBP) and lumbar pathologies. Unilateral spinal disorders provide unique models to study this association, with implications for diagnosis, prognosis, and management. Statistical shape analysis is a technique that can identify signature shape variations related to phenotypes but has never been employed in studying paraspinal muscle morphology. We present the first investigation using this technique to reveal disease-related paraspinal muscle asymmetry, using MRIs of patients with a single posterolateral disc herniation at the L5-S1 spinal level and unilateral leg pain. Statistical shape analysis was conducted to reveal disease- and phenotype-related morphological variations in the multifidus and erector spinae muscles at the level of herniation and the one below. With the analysis, shape variations associated with disc herniation were identified in the multifidus on the painful side at the level below the pathology while no pathology-related asymmetry in cross-sectional area (CSA) and fatty infiltration was found in either muscle. The results demonstrate higher sensitivity and spatial specificity for the technique than typical CSA and fatty infiltration measures. Statistical shape analysis holds promise in studying paraspinal muscle morphology to improve our understanding of LBP and various lumbar pathologies.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Statistical morphological analysis reveals characteristic paraspinal muscle asymmetry in unilateral lumbar disc herniation

Xiao

Fortin

Ahn

et al. 2021

Sci Rep

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, it should be noted that in both studies psoas major and quadratus lumborum were not segmented, and that an internal validation on a test set randomly selected from the original database rather than an external validation of a distinct dataset was performed. Xia and colleagues [11] tested several network architectures, including the U-Net as well as novel solutions, to segment psoas major, erector spinae, and multifidus, obtaining excellent Dice similarity scores between 0.913 and 0.95 for the best-performing neural network. Other studies achieved relatively lower performances [21], or focused on different imaging modalities or fields of view [22, 23].…”

Section: Discussionmentioning

confidence: 99%

“…the identification of the contours of the muscle mass or of the individual muscles in each slice. Although semiautomated and fully automated methods to segment MRI scans have been presented [9][10][11][12], in most clinical papers about sarcopenia this step has been conducted by human operators since such methods are not publicly available and their performance is generally lower than that of expert operators, not generalizing well to pathological cases [11]. Besides, manual segmentation requires a substantial amount of work and practically poses a limit to the sample size, since processing high numbers of subjects such as several thousands requires a time not compatible with the length of most research projects.…”

Section: Introductionmentioning

confidence: 99%

An externally validated deep learning model for the accurate segmentation of the lumbar paravertebral muscles

Niemeyer

Zanker

Jonas

et al. 2021

Preprint

View full text Add to dashboard Cite

Purpose. Imaging studies about the relevance of muscles in spinal disorders, and sarcopenia in general, require the segmentation of the muscles in the images which is very labour-intensive if performed manually and poses a practical limit to the number of investigated subjects. This study aimed at developing a deep learning-based tool able to fully automatically perform an accurate segmentation of the lumbar muscles in axial MRI scans, and at validating the new tool on an external dataset. Methods. A set of 60 axial MRI images of the lumbar spine was retrospectively collected from a clinical database. Psoas major, quadratus lumborum, erector spinae, and multifidus were manually segmented in all available slices. The dataset was used to train and validate a deep neural network able to segment muscles automatically. Subsequently, the network was externally validated on images purposely acquired from 22 healthy volunteers. Results. The Jaccard index for the individual muscles calculated for the 22 subjects of the external validation set ranged between 0.862 and 0.935, demonstrating a generally excellent performance of the network. Cross-sectional area and fat fraction of the muscles were in agreement with published data. Conclusions. The externally validated deep neural network was able to perform the segmentation of the paravertebral muscles in axial MRI scans in an accurate and fully automated manner, and is therefore a suitable tool to perform large-scale studies in the field of spinal disorders and sarcopenia, overcoming the limitations of non-automated methods.

show abstract

“…Not only will consistency in segmentation protocols aid in comparing studies and clinical populations, but researchers utilizing the same software to perform segmentations will also allow for easier comparisons between studies. With more researchers using the same segmentation software and protocols, a database can be built to eventually lead towards automated segmentation [23,24]. Compared to manual segmentation, automated segmentation is faster, easier, and may allow for easier access to results, especially if it can be integrated to clinical settings.…”

Section: Future Directionsmentioning

confidence: 99%

PILLAR: ParaspInaL muscLe segmentAtion pRoject - a comprehensive online resource to guide manual segmentation of paraspinal muscles from Magnetic Resonance Imaging

Anstruther¹,

Rossini²,

Zhang³

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Background – There is an increasing interest in assessing paraspinal morphology and composition in relation to low back pain (LBP). However, variations in methods and segmentation protocols contribute to the inconsistent findings in the literature. We present an on-line resource, the ParaspInaL muscLe segmentAtion pRoject (PILLAR , https://projectpillar.github.io/), to provide a detailed description and visual guide of a segmentation protocol by using the publicly available ITK-SNAP software and discuss related challenges when performing paraspinal lumbar muscles segmentations from magnetic resonance imaging (MRI). Methods – T2-weighted and corresponding fat-water IDEAL axial MRI from 3 males and 3 females (2 chronic LBP and 1 control for each sex) were used to demonstrate our segmentation protocol for each lumbar paraspinal muscle (erector spinae, lumbar multifidus, quadratus lumborum and psoas) and lumbar spinal level (L1-L5). Results – Proper segmentation requires an understanding of the anatomy of paraspinal lumbar muscles and the variations in paraspinal muscle morphology and composition due to age, sex, and the presence of LBP or related spinal pathologies. Other challenges in segmentation includes the presence and variations of intramuscular and epimuscular fat, and side-to-side asymmetry. Conclusion – The growing interest to assess the lumbar musculature and its role in the development and recurrence of LBP prompted the need for comprehensive and easy-to-follow resources, such as the PILLAR project to reduce inconsistencies in segmentation protocols. Standardizing manual muscle measurements from MRI will facilitate comparisons between studies while the field is progressively moving towards the automatization of paraspinal muscle measurements for large cohort studies.

show abstract

Automatic Paraspinal Muscle Segmentation in Patients with Lumbar Pathology Using Deep Convolutional Neural Network

Cited by 11 publications

References 11 publications

Statistical morphological analysis reveals characteristic paraspinal muscle asymmetry in unilateral lumbar disc herniation

Statistical morphological analysis reveals characteristic paraspinal muscle asymmetry in unilateral lumbar disc herniation

An externally validated deep learning model for the accurate segmentation of the lumbar paravertebral muscles

PILLAR: ParaspInaL muscLe segmentAtion pRoject - a comprehensive online resource to guide manual segmentation of paraspinal muscles from Magnetic Resonance Imaging

Contact Info

Product

Resources

About