Deep Learning Convolutional Neural Networks for the Automatic Quantification of Muscle Fat Infiltration Following Whiplash Injury

MRI-based assessment paraspinal extensor muscle fatty infiltration in acute cervical spinal cord injury patients - a retrospective study

BACKGROUND

The effect of fat infiltration in the paraspinal muscles on cervical degenerative disease has been confirmed by multiple studies. However, little is known about fat infiltration in the paraspinal extensors in patients with acute cervical spinal cord injury (SCI). This study aimed to investigate the difference in paraspinal extensor fatty infiltration between patients with acute cervical SCI and healthy controls, and to further explore the protective role of the paravertebral extensor muscles in patients with cervical SCI.

METHODS

A total of 50 patients with acute cervical SCI admitted to the emergency department from January 2019 to November 2023 were retrospectively analyzed, including 26 males and 24 females, with an average age of 59.60 ± 10.81 years. A control group of 50 healthy middle-aged and elderly individuals was also included, comprising 28 males and 22 females, with an average age of 55.00 ± 8.21 years. Cervical spine magnetic resonance imaging (MRI) was used to measure the cross-sectional areas of the superficial and deep cervical extensor muscles, the corresponding vertebral body cross-sectional areas, and the fat area within the superficial and deep extensor muscle groups using Image J software. Differences between the two groups were compared, and the cervical SCI patients were further analyzed based on the severity of the spinal cord injury and gender differences.

RESULTS

The deep fatty infiltration ratio (DFIR) and superficial fatty infiltration ratio (SFIR) at C4-C7 in the cervical SCI group were significantly higher than those in the control group (P < 0.001). The cross-sectional area of the functional deep extensor area (FDEA) relative to the vertebral body area (VBA) and the cross-sectional area of the functional superficial extensor area (FSEA) relative to the VBA at the C5 and C6 levels in the cervical SCI group were significantly lower than those in the control group (P < 0.001, P < 0.001, P = 0.034, P = 0.004 respectively). Among the cervical SCI patients, the cross-sectional areas of the deep extensor area (DEA) and the superficial extensor area (SEA) in males were significantly higher than those in females (P < 0.001). At the C6 and C7 levels, the FDEA/VBA and FSEA/VBA ratios in the male group were higher than those in the female group (P = 0.009, P = 0.022, P = 0.019, P = 0.005, respectively).

CONCLUSION

Patients with acute cervical SCI exhibit significantly higher fatty infiltration and a greater degree of paravertebral extensor muscle degeneration compared to healthy controls. This finding underscores the importance of the paravertebral extensor muscles in the context of cervical SCI and may guide future therapeutic strategies.

Morphological Changes of Deep Extensor Neck Muscles in Relation to the Maximum Level of Cord Compression and Canal Compromise in Patients With Degenerative Cervical Myelopathy

STUDY DESIGN

Cross-sectional study.

OBJECTIVES

To examine the relationship between morphological changes of the deep extensor neck muscles in patients with degenerative cervical myelopathy (DCM) and the level of maximum spinal cord compression (MSCC) and canal compromise (MCC). A secondary objective was to examine the relationship between muscle morphological changes with neck pain and functional scores related to neck pain and interference.

METHODS

A total of 171 patients with DCM were included. Total cross-sectional area (CSA), functional CSA (fat free area, FCSA), ratio of FCSA/CSA (fatty infiltration) and asymmetry of the multifidus (MF) and semispinalis cervicis (SCer) together, and cervical muscle as a group (eg, MF, SCer, semispinalis capitis, splenius capitis) were obtained from T2-weighted axial MR images at mid-disc, at the level of maximum cord compression and the level below. The relationship between the muscle parameters of interest, MSCC, MCC and functional scores including the Neck Disability Index (NDI) was assessed using multivariate linear regression models, adjusting for age, body mass index and sex.

RESULTS

Greater MF + Scer fatty infiltration was associated with greater MCC (P = .032) and MSCC (P = .049) at the same level. Greater asymmetry in MF + SCer CSA was also associated with greater MCC (P = .006). Similarly, greater asymmetry in FCSA and FCSA/CSA of the entire extensor muscle was associated with greater MCC (P = .011, P = .013). There was a negative association between asymmetry in FCSA MF + SCer, FCSA/CSA MF + SCer and FCSA/CSA group muscles with NDI score at the level below.

CONCLUSION

Greater MCC is associated with increased fatty infiltration and greater asymmetry of the deep cervical muscles in patients with DCM. A negative association between muscle asymmetry and NDI scores was also observed which has implications for clinical prediction around axial neck pain.

The association between lateral hip muscle size/intramuscular fat infiltration and hip strength in active young adults with long standing hip/groin pain

OBJECTIVES

To investigate associations between lateral hip muscle size/intramuscular fat infiltration (MFI) and hip strength in active young adults with longstanding hip/groin pain.

DESIGN

Cross-sectional study.

SETTING

University/Clinical.

PARTICIPANTS

Sub-elite soccer and Australian Football players (n = 180; 37 female) with long standing hip/groin pain.

MAIN OUTCOME MEASURES

Muscle size (volume) and MFI of gluteus maximus, medius, and minimis, and tensor fascia latae (TFL) were assessed using magnetic resonance imaging. Isometric hip strength was measured with handheld dynamometry. Associations between muscle size/MFI were assessed using linear regression models, adjusted for body mass index and age, with sex-specific interactions.

RESULTS

Positive associations were identified between lateral hip muscle volume and hip muscle strength, particularly for gluteus maximus and gluteus minimus volume. For all muscles, hip abduction was associated with an increase in strength by up to 0.69 N (R2 ranging from 0.29 to 0.39). These relationships were consistent across sexes with no sex interactions observed. No associations were found between MFI and strength measures.

CONCLUSION

Greater lateral hip muscle volumes are associated with greater hip strength in active young adults with long standing hip/groin pain, irrespective of sex. Gluteus maximus and minimus volume showed the most consistent relationships with hip strength across multiple directions.

Advances in and applications of imaging and radiomics in head and neck cancer survivorship

PURPOSE OF REVIEW

Radiological imaging is an essential component of head/neck cancer (HNC) care. Advances in imaging modalities (including CT, PET, MRI and ultrasound) and analysis have enhanced our understanding of tumour characteristics and prognosis. However, the application of these methods to evaluate treatment-related toxicities and functional burden is still emerging. This review showcases recent literature applying advanced imaging and radiomics to the assessment and management of sequelae following chemoradiotherapy for HNC.

RECENT FINDINGS

Whilst primarily early-stage/exploratory studies, recent investigations have showcased the feasibility of using radiological imaging, particularly advanced/functional MRI (including diffusion-weighted and dynamic contrast-enhanced MRI), to quantify treatment-induced tissue change in the head/neck musculature, and the clinical manifestation of lymphoedema/fibrosis and dysphagia. Advanced feature analysis and radiomic studies have also begun to give specific focus to the prediction of functional endpoints, including dysphagia, trismus and fibrosis.

SUMMARY

There is demonstrated potential in the use of novel imaging techniques, to help better understand pathophysiology, and improve assessment and treatment of functional deficits following HNC treatment. As larger studies emerge, technologies continue to progress, and pathways to clinical translation are honed, the application of these methods offers an exciting opportunity to transform clinical practices and improve outcomes for HNC survivors.

The association between morphological characteristics of paraspinal muscle and spinal disorders

BACKGROUND

Spinal disorders affect millions of people worldwide, and can cause significant disability and pain. The paraspinal muscles, located on either side of the spinal column, play a crucial role in the movement, support, and stabilization of the spine. Many spinal disorders can affect paraspinal muscles, as evidenced by changes in their morphology, including hypertrophy, atrophy, and degeneration.

OBJECTIVES

The objectives of this review were to examine the current literature on the relationship between the paraspinal muscles and spinal disorders, summarize the methods used in previous studies, and identify areas for future research.

METHODS

We reviewed studies on the morphological characteristics of the paravertebral muscle and discussed their relationship with spinal disorders, as well as the current limitations and future research directions.

RESULTS

The paraspinal muscles play a critical role in spinal disorders and are important targets for the treatment and prevention of spinal disorders. Clinicians should consider the role of the paraspinal muscles in the development and progression of spinal disorders and incorporate assessments of the paraspinal muscle function in clinical practice.

CONCLUSION

The findings of this review highlight the need for further research to better understand the relationship between the paraspinal muscles and spinal disorders, and to develop effective interventions to improve spinal health and reduce the burden of spinal disorders.

IMFSegNet: Cost-effective and objective quantification of intramuscular fat in histological sections by deep learning

The assessment of muscle condition is of great importance in various research areas. In particular, evaluating the degree of intramuscular fat (IMF) in tissue sections is a challenging task, which today is still mostly performed qualitatively or quantitatively by a highly subjective and error-prone manual analysis. We here realize the mission to make automated IMF analysis possible that (i) minimizes subjectivity, (ii) provides accurate and quantitative results quickly, and (iii) is cost-effective using standard hematoxylin and eosin (H&E) stained tissue sections. To address all these needs in a deep learning approach, we utilized the convolutional encoder-decoder network SegNet to train the specialized network IMFSegNet allowing to accurately quantify the spatial distribution of IMF in histological sections. Our fully automated analysis was validated on 17 H&E-stained muscle sections from individual sheep and compared to various state-of-the-art approaches. Not only does IMFSegNet outperform all other approaches, but this neural network also provides fully automated and highly accurate results utilizing the most cost-effective procedures of sample preparation and imaging. Furthermore, we shed light on the opacity of black-box approaches such as neural networks by applying an explainable artificial intelligence technique to clarify that the success of IMFSegNet actually lies in identifying the hard-to-detect IMF structures. Embedded in our open-source visual programming language JIPipe that does not require programming skills, it can be expected that IMFSegNet advances muscle condition assessment in basic research across multiple areas as well as in research fields focusing on translational clinical applications.

Biopsychosocial sequelae and recovery trajectories from whiplash injury following a motor vehicle collision

BACKGROUND CONTEXT

Five out of 10 injured in a motor vehicle collision (MVC) will develop persistent pain and disability. It is unclear if prolonged symptoms are related to peritraumatic pain/disability, psychological distress, muscle fat, lower extremity weakness.

PURPOSE

To test if widespread muscle fat infiltration (MFI) was (1) unique to those with poor recovery, (2) present in the peritraumatic stage, (3) related to known risk factors.

STUDY DESIGN/SETTING

A cohort study, single-center academic hospital.

PATIENT SAMPLES

A total of 97 men and women (age 18-65) presenting to an urban academic emergency medicine department following MVC, but not requiring inpatient hospitalization.

PRIMARY OUTCOME MEASURE

Neck disability at 12-months.

METHODS

Participants underwent magnetic resonance imaging (MRI) to quantify neck and lower extremity MFI, completed questionnaires on pain/disability and psychological distress (< 1-week, 2-weeks, 3-, and 12-months) and underwent maximum volitional torque testing of their lower extremities (2-weeks, 3-, and 12-months). Percentage score on the Neck Disability Index at 12-months was used for a model of (1) Recovered (0%-8%), (2) Mild (10%-28%), and (3) Moderate/Severe (≥ 30%). This model was adjusted for BMI and age.

RESULTS

Significant differences for neck MFI were revealed, with the Recovered group having significantly lower neck MFI than the Mild and Moderate/Severe groups at all time points. The Mild group had significantly more leg MFI at 12-months (p=.02) than the Recovered group. There were no other significant differences at any other time point. Lower extremity torques revealed no group differences. The Traumatic Injury Distress Scale (TIDS) and MFI of the neck at 1-week postinjury significantly predicted NDI score at 12-months.

CONCLUSIONS

Higher neck MFI and distress may represent a risk factor though it is unclear whether this is a pre-existing phenotype or result of the trauma.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02157038.

Role of Artificial Intelligence and Machine Learning in the prediction of the pain: A scoping systematic review

Artificial Intelligence in healthcare is growing quickly in diagnostics and treatment management. Despite the quantity and variety of studies its role in clinical care is not clear. To identify the evidence gaps and characteristics of the Artificial Intelligence and Machine Learning techniques in predicting/diagnosing the pain? Pubmed/Embase were searched from the inception to October 2021 for articles without any language restrictions specifically addressing the following: the accuracy of AI in pain considering Brain Imaging, Patient-reported measures, and Electrophysiology, the ability of AI to differentiate stratify severity/types of pain, the ability of AI to predict pain and lastly the most accurate AI technique for given inputs. All the included studies were on humans. Eight hundred forty abstracts were reviewed, and 23 articles were finally included. Identified records were independently screened and relevant data was extracted. We performed conceptual synthesis by grouping the studies using available concepts of AL/ML techniques in diagnosing pain. Then we summarized the number of features/physiological measurements. Structured tabulation synthesis was used to show patterns predictions along with a narrative commentary. A total of 23 articles, published between 2015 and 2020 from 12 countries were included. Most studies were experimental in design. The most common design was cross sectional. Chronic or acute pains were predicted more often. Compared to control, the pain prediction was in the range of 57%–96% by AI techniques. Support Vector Machine and deep learning showed higher accuracy for classifying pain. From this study, it can be inferred that AI/ML can be used to differentiate healthy controls from patients. It can also facilitate categorizing them into new and different clinical subgroups. Lastly, it can predict future pain. The limitations are with respect to studies done after the search period. AL/ ML has a supportive role in pain diagnostics.

Convolutional neural networks for the automatic segmentation of lumbar paraspinal muscles in people with low back pain

The size, shape, and composition of paraspinal muscles have been widely reported in disorders of the cervical and lumbar spine. Measures of size, shape, and composition have required time-consuming and rater-dependent manual segmentation techniques. Convolutional neural networks (CNNs) provide alternate timesaving, state-of-the-art performance measures, which could realise clinical translation. Here we trained a CNN for the automatic segmentation of lumbar paraspinal muscles and determined the impact of CNN architecture and training choices on segmentation performance. T₂-weighted MRI axial images from 76 participants (46 female; age (SD): 45.6 (12.8) years) with low back pain were used to train CNN models to segment the multifidus, erector spinae, and psoas major muscles (left and right segmented separately). Using cross-validation, we compared 2D and 3D CNNs with and without data augmentation. Segmentation accuracy was compared between the models using the Sørensen-Dice index as the primary outcome measure. The effect of increasing network depth on segmentation accuracy was also investigated. Each model showed high segmentation accuracy (Sørensen-Dice index ≥ 0.885) and excellent reliability (ICC_2,1 ≥ 0.941). Overall, across all muscles, 2D models performed better than 3D models (p = 0.012), and training without data augmentation outperformed training with data augmentation (p < 0.001). The 2D model trained without data augmentation demonstrated the highest average segmentation accuracy. Increasing network depth did not improve accuracy (p = 0.771). All trained CNN models demonstrated high accuracy and excellent reliability for segmenting lumbar paraspinal muscles. CNNs can be used to efficiently and accurately extract measures of paraspinal muscle health from MRI.

Comparison between movement pattern training and strengthening on muscle volume, muscle fat, and strength in patients with hip-related groin pain: An exploratory analysis

The purpose of this exploratory analysis was to compare the impact of movement pattern training (MoveTrain) and standard strength and flexibility training (Standard) on muscle volume, strength and fatty infiltration in patients with hip-related groin pain (HRGP). We completed a secondary analysis of data collected during an assessor-blinded randomized control trial. Data were used from 27 patients with HRGP, 15-40 years, who were randomized into MoveTrain or Standard groups. Both groups participated in their training protocol (MoveTrain, n = 14 or Standard, n = 13) which included 10 supervised sessions over 12 weeks and a daily home exercise program. Outcome measures were collected at baseline and immediately after treatment. Magnetic resonance images data were used to determine muscle fat index (MFI) and muscle volume. A hand-held dynamometer was used to assess isometric hip abductor and extensor strength. The Standard group demonstrated a significant posttreatment increase in gluteus medius muscle volume compared to the MoveTrain group. Both groups demonstrated an increase in hip abductor strength and reduction in gluteus minimus and gluteus maximus MFI. The magnitude of change for all outcomes were modest. Statement of Clinical Significance: Movement pattern training or a program of strength/flexibility training may be effective at improving hipabductor strength and reducing fatty infiltration in the gluteal musculature among those with HRGP. Further research is needed to betterunderstand etiology of strength changes and impact of muscle volume and MFI in HRGP and the effect of exercise on muscle structure andfunction.

Reliability and agreement of lumbar multifidus volume and fat fraction quantification using magnetic resonance imaging

INTRODUCTION

Magnetic resonance imaging (MRI) is the standard to quantify size and structure of lumbar muscles. Three-dimensional volumetric measures are expected to be more closely related to muscle function than two-dimensional measures such as cross-sectional area. Reliability and agreement of a standardized method should be established to enable the use of MRI to assess lumbar muscle characteristics.

OBJECTIVES

This study investigates the intra- and inter-processor reliability for the quantification of (1) muscle volume and (2) fat fraction based on chemical shift MRI images using axial 3D-volume measurements of the lumbar multifidus in patients with low back pain.

METHODS

Two processors manually segmented the lumbar multifidus on the MRI scans of 18 patients with low back pain using Mevislab software following a well-defined method. Fat fraction of the segmented volume was calculated. Reliability and agreement were determined using intra-class correlation coefficients (ICC), Bland-Altman plots and calculation of the standard error of measurement (SEM) and minimal detectable change (MDC).

RESULTS

Excellent ICCs were found for both intra-processor and inter-processor analysis of lumbar multifidus volume measurement, with slightly better results for the intra-processor reliability. The SEMs for volume were lower than 4.1 cm³. Excellent reliability and agreement were also found for fat fraction measures, with ICCs of 0.985-0.998 and SEMs below 0.946%.

CONCLUSION

The proposed method to quantify muscle volume and fat fraction of the lumbar multifidus on MRI was highly reliable, and can be used in further research on lumbar multifidus structure.

Improving the measurement of intrinsic foot muscle morphology and composition from high-field (7T) magnetic resonance imaging

Magnetic resonance imaging (MRI) can be used to quantify intrinsic foot muscle morphology and composition. Due to the high spatial resolution required to adequately capture the architecturally complex anatomy, manual segmentation is time consuming and not clinically feasible. The aim of this study was to evaluate if a reduced number of MRI slices can be used to accurately estimate intrinsic foot muscle volume and composition. A three-dimensional 2-point Dixon sequence of the whole foot was acquired at 7-Tesla for thirteen asymptomatic individuals and twenty individuals with plantar heel pain. Slice intervals of 2, 3, 5, 10, 15 and 30 were used to calculate alternative muscle volume and composition, and were compared to reference values calculated from every available slice. Agreement between methods was assessed by calculating mean differences and 95% limits of agreement, and inspection of Bland -Altman plots. In both groups, slice intervals of 2, 3 and 5 provided excellent precision for all muscles (measurement error < 1%). Larger slice intervals of 10, 15 and 30 provided excellent precision for some muscles, but for other muscles (e.g. small forefoot muscles), error was up to 7.3%. Bland-Altman plots showed no systematic measurement bias. This study provides a quantitative basis for selecting a reduced number of slices to measure intrinsic foot muscle volume and composition from MRI. A slice interval of 10 may provide a balance between efficiency (36 mins vs. 6 h) and accuracy (error < 2.4%) across all intrinsic foot muscles in asymptomatic individuals and those with plantar heel pain.

Utilization of Mid-Thigh Magnetic Resonance Imaging to Predict Lean Body Mass and Knee Extensor Strength in Obese Adults

Purpose

To train and test a machine learning model to automatically measure mid-thigh muscle cross-sectional area (CSA) to provide rapid estimation of appendicular lean mass (ALM) and predict knee extensor torque of obese adults.

Methods

Obese adults [body mass index (BMI) = 30–40 kg/m², age = 30–50 years] were enrolled for this study. Participants received full-body dual-energy X-ray absorptiometry (DXA), mid-thigh MRI, and completed knee extensor and flexor torque assessments via isokinetic dynamometer. Manual segmentation of mid-thigh CSA was completed for all MRI scans. A convolutional neural network (CNN) was created based on the manual segmentation to develop automated quantification of mid-thigh CSA. Relationships were established between the automated CNN values to the manual CSA segmentation, ALM via DXA, knee extensor, and flexor torque.

Results

A total of 47 obese patients were enrolled in this study. Agreement between the CNN-automated measures and manual segmentation of mid-thigh CSA was high (>0.90). Automated measures of mid-thigh CSA were strongly related to the leg lean mass (r = 0.86, p < 0.001) and ALM (r = 0.87, p < 0.001). Additionally, mid-thigh CSA was strongly related to knee extensor strength (r = 0.76, p < 0.001) and moderately related to knee flexor strength (r = 0.48, p = 0.002).

Conclusion

CNN-measured mid-thigh CSA was accurate compared to the manual segmented values from the mid-thigh. These values were strongly predictive of clinical measures of ALM and knee extensor torque. Mid-thigh MRI may be utilized to accurately estimate clinical measures of lean mass and function in obese adults.

Classification of rotator cuff tears in ultrasound images using deep learning models

Rotator cuff tears (RCTs) are one of the most common shoulder injuries, which are typically diagnosed using relatively expensive and time-consuming diagnostic imaging tests such as magnetic resonance imaging or computed tomography. Deep learning algorithms are increasingly used to analyze medical images, but they have not been used to identify RCTs with ultrasound images. The aim of this study is to develop an approach to automatically classify RCTs and provide visualization of tear location using ultrasound images and convolutional neural networks (CNNs). The proposed method was developed using transfer learning and fine-tuning with five pre-trained deep models (VGG19, InceptionV3, Xception, ResNet50, and DenseNet121). The Bayesian optimization method was also used to optimize hyperparameters of the CNN models. A total of 194 ultrasound images from Kosin University Gospel Hospital were used to train and test the CNN models by five-fold cross-validation. Among the five models, DenseNet121 demonstrated the best classification performance with 88.2% accuracy, 93.8% sensitivity, 83.6% specificity, and AUC score of 0.832. A gradient-weighted class activation mapping (Grad-CAM) highlighted the sensitive features in the learning process on ultrasound images. The proposed approach demonstrates the feasibility of using deep learning and ultrasound images to assist RCTs' diagnosis.

Analysis of the paraspinal muscle morphology of the lumbar spine using a convolutional neural network (CNN)

PURPOSE

This single-center study aimed to develop a convolutional neural network to segment multiple consecutive axial magnetic resonance imaging (MRI) slices of the lumbar spinal muscles of patients with lower back pain and automatically classify fatty muscle degeneration.

METHODS

We developed a fully connected deep convolutional neural network (CNN) with a pre-trained U-Net model trained on a dataset of 3,650 axial T2-weighted MRI images from 100 patients with lower back pain. We included all qualities of MRI; the exclusion criteria were fractures, tumors, infection, or spine implants. The training was performed using k-fold cross-validation (k = 10), and performance was evaluated using the dice similarity coefficient (DSC) and cross-sectional area error (CSA error). For clinical correlation, we used a simplified Goutallier classification (SGC) system with three classes.

RESULTS

The mean DSC was high for overall muscle (0.91) and muscle tissue segmentation (0.83) but showed deficiencies in fatty tissue segmentation (0.51). The CSA error was small for the overall muscle area of 8.42%, and fatty tissue segmentation showed a high mean CSA error of 40.74%. The SGC classification was correctly predicted in 75% of the patients.

CONCLUSION

Our fully connected CNN segmented overall muscle and muscle tissue with high precision and recall, as well as good DSC values. The mean predicted SGC values of all available patient axial slices showed promising results. With an overall Error of 25%, further development is needed for clinical implementation. Larger datasets and training of other model architectures are required to segment fatty tissue more accurately.

Gluteal Muscle Atrophy and Increased Intramuscular Lipid Concentration Are Not Mitigated by Daily Artificial Gravity Following 60-Day Head-Down Tilt Bed Rest

Exposure to spaceflight and head-down tilt (HDT) bed rest leads to decreases in the mass of the gluteal muscle. Preliminary results have suggested that interventions, such as artificial gravity (AG), can partially mitigate some of the physiological adaptations induced by HDT bed rest. However, its effect on the gluteal muscles is currently unknown. This study investigated the effects of daily AG on the gluteal muscles during 60-day HDT bed rest. Twenty-four healthy individuals participated in the study: eight received 30 min of continuous AG; eight received 6 × 5 min of AG, interspersed with rest periods; eight belonged to a control group. T1-weighted Dixon magnetic resonance imaging of the hip region was conducted at baseline and day 59 of HDT bed rest to establish changes in volumes and intramuscular lipid concentration (ILC). Results showed that, across groups, muscle volumes decreased by 9.2% for gluteus maximus (GMAX), 8.0% for gluteus medius (GMED), and 10.5% for gluteus minimus after 59-day HDT bed rest (all p < 0.005). The ILC increased by 1.3% for GMAX and 0.5% for GMED (both p < 0.05). Neither of the AG protocols mitigated deconditioning of the gluteal muscles. Whereas all gluteal muscles atrophied, the ratio of lipids to intramuscular water increased only in GMAX and GMED muscles. These changes could impair the function of the hip joint and increased the risk of falls. The deconditioning of the gluteal muscles in space may negatively impact the hip joint stability of astronauts when reexpose to terrestrial gravity.

Multi-muscle deep learning segmentation to automate the quantification of muscle fat infiltration in cervical spine conditions

Muscle fat infiltration (MFI) has been widely reported across cervical spine disorders. The quantification of MFI requires time-consuming and rater-dependent manual segmentation techniques. A convolutional neural network (CNN) model was trained to segment seven cervical spine muscle groups (left and right muscles segmented separately, 14 muscles total) from Dixon MRI scans (n = 17, 17 scans < 2 weeks post motor vehicle collision (MVC), and 17 scans 12 months post MVC). The CNN MFI measures demonstrated high test reliability and accuracy in an independent testing dataset (n = 18, 9 scans < 2 weeks post MVC, and 9 scans 12 months post MVC). Using the CNN in 84 participants with scans < 2 weeks post MVC (61 females, 23 males, age = 34.2 ± 10.7 years) differences in MFI between the muscle groups and relationships between MFI and sex, age, and body mass index (BMI) were explored. Averaging across all muscles, females had significantly higher MFI than males (p = 0.026). The deep cervical muscles demonstrated significantly greater MFI than the more superficial muscles (p < 0.001), and only MFI within the deep cervical muscles was moderately correlated to age (r > 0.300, p ≤ 0.001). CNN's allow for the accurate and rapid, quantitative assessment of the composition of the architecturally complex muscles traversing the cervical spine. Acknowledging the wider reports of MFI in cervical spine disorders and the time required to manually segment the individual muscles, this CNN may have diagnostic, prognostic, and predictive value in disorders of the cervical spine.

Fatty infiltration in cervical flexors and extensors in patients with degenerative cervical myelopathy using a multi-muscle segmentation model

Background

In patients with degenerative cervical myelopathy (DCM) that have spinal cord compression and sensorimotor deficits, surgical decompression is often performed. However, there is heterogeneity in clinical presentation and post-surgical functional recovery.

Objectives

Primary: a) to assess differences in muscle fat infiltration (MFI) in patients with DCM versus controls, b) to assess association between MFI and clinical disability. Secondary: to assess association between MFI pre-surgery and post-surgical functional recovery.

Study design

Cross-sectional case control study.

Methods

Eighteen patients with DCM (58.6 ± 14.2 years, 10 M/8F) and 25 controls (52.6 ± 11.8 years, 13M/12 F) underwent 3D Dixon fat-water imaging. A convolutional neural network (CNN) was used to segment cervical muscles (MFSS- multifidus and semispinalis cervicis, LC- longus capitis/colli) and quantify MFI. Modified Japanese Orthopedic Association (mJOA) and Nurick were collected.

Results

Patients with DCM had significantly higher MFI in MFSS (20.63 ± 5.43 vs 17.04 ± 5.24, p = 0.043) and LC (18.74 ± 6.7 vs 13.66 ± 4.91, p = 0.021) than controls. Patients with increased MFI in LC and MFSS had higher disability (LC: Nurick (Spearman’s ρ = 0.436, p = 0.003) and mJOA (ρ = -0.399, p = 0.008)). Increased MFI in LC pre-surgery was associated with post-surgical improvement in Nurick (ρ = -0.664, p = 0.026) and mJOA (ρ = -0.603, p = 0.049).

Conclusion

In DCM, increased muscle adiposity is significantly associated with sensorimotor deficits, clinical disability, and functional recovery after surgery. Accurate and time efficient evaluation of fat infiltration in cervical muscles may be conducted through implementation of CNN models.

A Deep-Learning-Based, Fully Automated Program to Segment and Quantify Major Spinal Components on Axial Lumbar Spine Magnetic Resonance Images

OBJECTIVE

The paraspinal muscles have been extensively studied on axial lumbar magnetic resonance imaging (MRI) for better understanding of back pain; however, the acquisition of measurements mainly relies on manual segmentation, which is time consuming. The study objective was to develop and validate a deep-learning-based program for automated acquisition of quantitative measurements for major lumbar spine components on axial lumbar MRIs, the paraspinal muscles in particular.

METHODS

This study used a cross-sectional observational design. From the Hangzhou Lumbar Spine Study, T2-weighted axial MRIs at the L4-5 disk level of 120 participants (aged 54.8 years [SD = 15.0]) were selected to develop the deep-learning-based program Spine Explorer (Tulong). Another 30 axial lumbar MRIs were automatically measured by Spine Explorer and then manually measured using ImageJ to acquire quantitative size and compositional measurements for bilateral multifidus, erector spinae, and psoas muscles; the disk; and the spinal canal. Intersection-over-union and Dice score were used to evaluate the performance of automated segmentation. Intraclass coefficients and Bland-Altman plots were used to examine intersoftware agreements for various measurements.

RESULTS

After training, Spine Explorer (Tulong) measures an axial lumbar MRI in 1 second. The intersections-over-union were 83.3% to 88.4% for the paraspinal muscles and 92.2% and 82.1% for the disk and spinal canal, respectively. For various size and compositional measurements of paraspinal muscles, Spine Explorer (Tulong) was in good agreement with ImageJ (intraclass coefficient = 0.85 to approximately 0.99).

CONCLUSION

Spine Explorer (Tulong) is automated, efficient, and reliable in acquiring quantitative measurements for the paraspinal muscles, the disk, and the canal, and various size and compositional measurements were simultaneously obtained for the lumbar paraspinal muscles.

IMPACT

Such an automated program might encourage further epidemiological studies of the lumbar paraspinal muscle degeneration and enhance paraspinal muscle assessment in clinical practice.

A deep learning model for detection of cervical spinal cord compression in MRI scans

Magnetic Resonance Imaging (MRI) evidence of spinal cord compression plays a central role in the diagnosis of degenerative cervical myelopathy (DCM). There is growing recognition that deep learning models may assist in addressing the increasing volume of medical imaging data and provide initial interpretation of images gathered in a primary-care setting. We aimed to develop and validate a deep learning model for detection of cervical spinal cord compression in MRI scans. Patients undergoing surgery for DCM as a part of the AO Spine CSM-NA or CSM-I prospective cohort studies were included in our study. Patients were divided into a training/validation or holdout dataset. Images were labelled by two specialist physicians. We trained a deep convolutional neural network using images from the training/validation dataset and assessed model performance on the holdout dataset. The training/validation cohort included 201 patients with 6588 images and the holdout dataset included 88 patients with 2991 images. On the holdout dataset the deep learning model achieved an overall AUC of 0.94, sensitivity of 0.88, specificity of 0.89, and f1-score of 0.82. This model could improve the efficiency and objectivity of the interpretation of cervical spine MRI scans.

Changes in Muscle Morphology in Female Chronic Neck Pain Patients Using Magnetic Resonance Imaging

STUDY DESIGN

Population-based cross-sectional study.

OBJECTIVE

The aim of this study was to gain a better understanding of changes in muscle morphology in patients with chronic idiopathic neck pain (CINP) and chronic whiplash-associated disorder (CWAD).

SUMMARY OF BACKGROUND DATA

Worldwide, neck pain (NP) is a common health problem with high socioeconomic burden. A high percentage of these patients evolves toward chronic symptoms. Efficacy of treatments for these complaints remains variable. In current literature, changes in muscle morphology (muscle fat infiltration and cross-sectional area) have been reported in patients with NP, both CWAD and CINP. However, no strong conclusions could be made.

METHODS

In this study, magnetic resonance imaging was used to obtain data on muscle morphology from 14 cervical flexor and extensor muscles in 117 female subjects with NP (CWAD = 37; CINP = 45) and healthy controls (HC = 35).

RESULTS

The CWAD group had a significantly larger muscle fat infiltration in some extensor (semispinalis and splenius capitis, trapezius, obliquus capitis inferior) and flexor (sternocleidomastoid) muscles compared to the CINP and/or HC group. A significantly larger (muscle) cross-sectional area was found in some extensor (levator scapulae, semispinalis capitis, trapezius) and flexor (longus colli, longus capitis, sternocleidomastoid) muscles in the HC group compared to the CINP and/or CWAD group. No clear associations were found between group differences and factors as pain duration, kinesiophobia, and disability.

CONCLUSION

The results in this study suggest changes in muscle morphology in both NP cohorts. These results show some similarities with earlier findings in this research domain. Further studies based on controlled longitudinal designs are needed to facilitate data compilation, to draw stronger conclusions, and to integrate them into the treatment of patients with chronic NP.Level of Evidence: 4.

Applications of Machine Learning to Imaging of Spinal Disorders: Current Status and Future Directions

STUDY DESIGN

Narrative review.

OBJECTIVES

We aim to describe current progress in the application of artificial intelligence and machine learning technology to provide automated analysis of imaging in patients with spinal disorders.

METHODS

A literature search utilizing the PubMed database was performed. Relevant studies from all the evidence levels have been included.

RESULTS

Within spine surgery, artificial intelligence and machine learning technologies have achieved near-human performance in narrow image classification tasks on specific datasets in spinal degenerative disease, spinal deformity, spine trauma, and spine oncology.

CONCLUSION

Although substantial challenges remain to be overcome it is clear that artificial intelligence and machine learning technology will influence the practice of spine surgery in the future.

New insights into intrinsic foot muscle morphology and composition using ultra-high-field (7-Tesla) magnetic resonance imaging

Background

The intrinsic muscles of the foot are key contributors to foot function and are important to evaluate in lower limb disorders. Magnetic resonance imaging (MRI), provides a non-invasive option to measure muscle morphology and composition, which are primary determinants of muscle function. Ultra-high-field (7-T) magnetic resonance imaging provides sufficient signal to evaluate the morphology of the intrinsic foot muscles, and, when combined with chemical-shift sequences, measures of muscle composition can be obtained. Here we aim to provide a proof-of-concept method for measuring intrinsic foot muscle morphology and composition with high-field MRI.

Methods

One healthy female (age 39 years, mass 65 kg, height 1.73 m) underwent MRI. A T1-weighted VIBE – radio-frequency spoiled 3D steady state GRE – sequence of the whole foot was acquired on a Siemens 7T MAGNETOM scanner, as well as a 3T MAGNETOM Prisma scanner for comparison. A high-resolution fat/water separation image was also acquired using a 3D 2-point DIXON sequence at 7T. Coronal plane images from 3T and 7T scanners were compared. Using 3D Slicer software, regions of interest were manually contoured for each muscle on 7T images. Muscle volumes and percentage of muscle fat infiltration were calculated (muscle fat infiltration % = Fat/(Fat + Water) x100) for each muscle.

Results

Compared to the 3T images, the 7T images provided superior resolution, particularly at the forefoot, to facilitate segmentation of individual muscles. Muscle volumes ranged from 1.5 cm³ and 19.8 cm³, and percentage muscle fat infiltration ranged from 9.2–15.0%.

Conclusions

This proof-of-concept study demonstrates a feasible method of quantifying muscle morphology and composition for individual intrinsic foot muscles using advanced high-field MRI techniques. This method can be used in future studies to better understand intrinsic foot muscle morphology and composition in healthy individuals, as well as those with lower disorders.

Texture analysis and machine learning to predict water T2 and fat fraction from non-quantitative MRI of thigh muscles in Facioscapulohumeral muscular dystrophy

PURPOSE

Quantitative MRI (qMRI) plays a crucial role for assessing disease progression and treatment response in neuromuscular disorders, but the required MRI sequences are not routinely available in every center. The aim of this study was to predict qMRI values of water T2 (wT2) and fat fraction (FF) from conventional MRI, using texture analysis and machine learning.

METHOD

Fourteen patients affected by Facioscapulohumeral muscular dystrophy were imaged at both thighs using conventional and quantitative MR sequences. Muscle FF and wT2 were calculated for each muscle of the thighs. Forty-seven texture features were extracted for each muscle on the images obtained with conventional MRI. Multiple machine learning regressors were trained to predict qMRI values from the texture analysis dataset.

RESULTS

Eight machine learning methods (linear, ridge and lasso regression, tree, random forest (RF), generalized additive model (GAM), k-nearest-neighbor (kNN) and support vector machine (SVM) provided mean absolute errors ranging from 0.110 to 0.133 for FF and 0.068 to 0.115 for wT2. The most accurate methods were RF, SVM and kNN to predict FF, and tree, RF and kNN to predict wT2.

CONCLUSION

This study demonstrates that it is possible to estimate with good accuracy qMRI parameters starting from texture analysis of conventional MRI.

Large-scale analysis of iliopsoas muscle volumes in the UK Biobank

Psoas muscle measurements are frequently used as markers of sarcopenia and predictors of health. Manually measured cross-sectional areas are most commonly used, but there is a lack of consistency regarding the position of the measurement and manual annotations are not practical for large population studies. We have developed a fully automated method to measure iliopsoas muscle volume (comprised of the psoas and iliacus muscles) using a convolutional neural network. Magnetic resonance images were obtained from the UK Biobank for 5000 participants, balanced for age, gender and BMI. Ninety manual annotations were available for model training and validation. The model showed excellent performance against out-of-sample data (average dice score coefficient of 0.9046 ± 0.0058 for six-fold cross-validation). Iliopsoas muscle volumes were successfully measured in all 5000 participants. Iliopsoas volume was greater in male compared with female subjects. There was a small but significant asymmetry between left and right iliopsoas muscle volumes. We also found that iliopsoas volume was significantly related to height, BMI and age, and that there was an acceleration in muscle volume decrease in men with age. Our method provides a robust technique for measuring iliopsoas muscle volume that can be applied to large cohorts.

LUMINOUS database: lumbar multifidus muscle segmentation from ultrasound images

BACKGROUND

Among the paraspinal muscles, the structure and function of the lumbar multifidus (LM) has become of great interest to researchers and clinicians involved in lower back pain and muscle rehabilitation. Ultrasound (US) imaging of the LM muscle is a useful clinical tool which can be used in the assessment of muscle morphology and function. US is widely used due to its portability, cost-effectiveness, and ease-of-use. In order to assess muscle function, quantitative information of the LM must be extracted from the US image by means of manual segmentation. However, manual segmentation requires a higher level of training and experience and is characterized by a level of difficulty and subjectivity associated with image interpretation. Thus, the development of automated segmentation methods is warranted and would strongly benefit clinicians and researchers. The aim of this study is to provide a database which will contribute to the development of automated segmentation algorithms of the LM.

CONSTRUCTION AND CONTENT

This database provides the US ground truth of the left and right LM muscles at the L5 level (in prone and standing positions) of 109 young athletic adults involved in Concordia University's varsity teams. The LUMINOUS database contains the US images with their corresponding manually segmented binary masks, serving as the ground truth. The purpose of the database is to enable development and validation of deep learning algorithms used for automatic segmentation tasks related to the assessment of the LM cross-sectional area (CSA) and echo intensity (EI). The LUMINOUS database is publicly available at http://data.sonography.ai .

CONCLUSION

The development of automated segmentation algorithms based on this database will promote the standardization of LM measurements and facilitate comparison among studies. Moreover, it can accelerate the clinical implementation of quantitative muscle assessment in clinical and research settings.

Negative emotion diffusion and intervention countermeasures of social networks based on deep learning

Aiming at the limitation of using only word features in traditional deep learning sentiment classification, this paper combines topic features with deep learning models to build a topic-fused deep learning sentiment classification model. The model can fuse topic features to obtain high-quality high-level text features. Experiments show that in binary sentiment classification, the highest classification accuracy of the model can reach more than 90%, which is higher than that of commonly used deep learning models. This paper focuses on the combination of deep neural networks and emerging text processing technologies, and improves and perfects them from two aspects of model architecture and training methods, and designs an efficient deep network sentiment analysis model. A CNN (Convolutional Neural Network) model based on polymorphism is proposed. The model constructs the CNN input matrix by combining the word vector information of the text, the emotion information of the words, and the position information of the words, and adjusts the importance of different feature information in the training process by means of weight control. The multi-objective sample data set is used to verify the effectiveness of the proposed model in the sentiment analysis task of related objects from the classification effect and training performance.

Deep learning for the rapid automatic quantification and characterization of rotator cuff muscle degeneration from shoulder CT datasets

Objectives

This study aimed at developing a convolutional neural network (CNN) able to automatically quantify and characterize the level of degeneration of rotator cuff (RC) muscles from shoulder CT images including muscle atrophy and fatty infiltration.

Methods

One hundred three shoulder CT scans from 95 patients with primary glenohumeral osteoarthritis undergoing anatomical total shoulder arthroplasty were retrospectively retrieved. Three independent radiologists manually segmented the premorbid boundaries of all four RC muscles on standardized sagittal-oblique CT sections. This premorbid muscle segmentation was further automatically predicted using a CNN. Automatically predicted premorbid segmentations were then used to quantify the ratio of muscle atrophy, fatty infiltration, secondary bone formation, and overall muscle degeneration. These muscle parameters were compared with measures obtained manually by human raters.

Results

Average Dice similarity coefficients for muscle segmentations obtained automatically with the CNN (88% ± 9%) and manually by human raters (89% ± 6%) were comparable. No significant differences were observed for the subscapularis, supraspinatus, and teres minor muscles (p > 0.120), whereas Dice coefficients of the automatic segmentation were significantly higher for the infraspinatus (p < 0.012). The automatic approach was able to provide good–very good estimates of muscle atrophy (R² = 0.87), fatty infiltration (R² = 0.91), and overall muscle degeneration (R² = 0.91). However, CNN-derived segmentations showed a higher variability in quantifying secondary bone formation (R² = 0.61) than human raters (R² = 0.87).

Conclusions

Deep learning provides a rapid and reliable automatic quantification of RC muscle atrophy, fatty infiltration, and overall muscle degeneration directly from preoperative shoulder CT scans of osteoarthritic patients, with an accuracy comparable with that of human raters.

Key Points

• Deep learning can not only segment RC muscles currently available in CT images but also learn their pre-existing locations and shapes from invariant anatomical structures visible on CT sections.

• Our automatic method is able to provide a rapid and reliable quantification of RC muscle atrophy and fatty infiltration from conventional shoulder CT scans.

• The accuracy of our automatic quantitative technique is comparable with that of human raters.

A Review on the Use of Artificial Intelligence in Spinal Diseases

Artificial neural networks (ANNs) have been used in a wide variety of real-world applications and it emerges as a promising field across various branches of medicine. This review aims to identify the role of ANNs in spinal diseases. Literature were searched from electronic databases of Scopus and Medline from 1993 to 2020 with English publications reported on the application of ANNs in spinal diseases. The search strategy was set as the combinations of the following keywords: “artificial neural networks,” “spine,” “back pain,” “prognosis,” “grading,” “classification,” “prediction,” “segmentation,” “biomechanics,” “deep learning,” and “imaging.” The main findings of the included studies were summarized, with an emphasis on the recent advances in spinal diseases and its application in the diagnostic and prognostic procedures. According to the search strategy, a set of 3,653 articles were retrieved from Medline and Scopus databases. After careful evaluation of the abstracts, the full texts of 89 eligible papers were further examined, of which 79 articles satisfied the inclusion criteria of this review. Our review indicates several applications of ANNs in the management of spinal diseases including (1) diagnosis and assessment of spinal disease progression in the patients with low back pain, perioperative complications, and readmission rate following spine surgery; (2) enhancement of the clinically relevant information extracted from radiographic images to predict Pfirrmann grades, Modic changes, and spinal stenosis grades on magnetic resonance images automatically; (3) prediction of outcomes in lumbar spinal stenosis, lumbar disc herniation and patient-reported outcomes in lumbar fusion surgery, and preoperative planning and intraoperative assistance; and (4) its application in the biomechanical assessment of spinal diseases. The evidence suggests that ANNs can be successfully used for optimizing the diagnosis, prognosis and outcome prediction in spinal diseases. Therefore, incorporation of ANNs into spine clinical practice may improve clinical decision making.

Age- and gender-related variations of cervical muscle composition using chemical shift encoding-based water-fat MRI

PURPOSE

To understand fat distribution patterns and ectopic fat deposition in healthy adults and to provide normative data, encompassing the borders of physiological regional muscle composition. For this purpose chemical shift encoding-based water-fat Magnetic Resonance Imaging (MRI) was used for proton density fat fraction (PDFF) calculations.

MATERIAL AND METHODS

91 volunteers were enrolled (male: n = 28, age = 36.6 ± 11.4 years; female: n = 63, age = 38.5 ± 15.1 years). PDFF values combined for the multifidus, semispinalis and spinalis cervicis muscles at the level of the 3rd cervical vertebral body (C3), the 5th cervical vertebral body (C5) and the first thoracic vertebral body (Th1) were extracted.

RESULTS

The paraspinal musculature at C3 (14.8 ± 10.1 % vs. 19.2 ± 11.0 %; p = 0.029) and Th1 (13.8 ± 7.0 % vs 17.7 ± 7.4 %; p = 0.011) showed significantly lower PDFF values in men compared to women. Partial correlation testing with BMI as control variable revealed highly significant correlations between the paraspinal musculature PDFF at C3 (men: r = 0.504, p = 0.007; women: r = 0.279, p = 0.028), C5 (men: r = 0.450, p = 0.019; women: r = 0.347, p = 0.006) and Th1 (men: r = 0.652, p < 0.0001; women: r = 0.443, p < 0.0001) with age in both genders.

CONCLUSION

The present data suggest gender and age-specific fat deposition patterns of the cervical and the upper cervicothoracic paraspinal muscles and may provide reference values for pathology detection.

Design of a Human Evaluator Model for the Ride Comfort of Vehicle on a Speed Bump Using a Neural Artistic Style Extraction

The subjective evaluation of vehicle ride comfort is costly and time-consuming but is crucial for vehicle development. To reduce the cost and time, the objectification of subjective evaluation has been widely studied, and most of the approaches use a regression model between objective metrics and subjective ratings. However, the accuracy of these approaches is highly dependent on the selection of the objective metrics. In most of the methods, it is not clear that the selected metrics are sufficiently significant or whether all significant metrics are included in the selection. This paper presents a method to build a correlation model between measurements and subjective evaluations without using predefined features or objective metrics. A numerical representation of ride comfort was extracted from raw signals based on the idea of the artistic style transfer method. The correlation model was designed based on the extracted numerical representation and subjective ratings. The model has a much better accuracy than any other correlation models in the literature. This better accuracy is contributed to not only by using a neural network, but also by the extraction of the numerical representation of ride comfort using a pre-trained neural network.

The relation between local and distal muscle fat infiltration in chronic whiplash using magnetic resonance imaging

The objective of this study was to investigate the relationship between fat infiltration in the cervical multifidi and fat infiltration measured in the lower extremities to move further into understanding the complex signs and symptoms arising from a whiplash trauma. Thirty-one individuals with chronic whiplash associated disorders, stratified into a mild/moderate group and a severe group, together with 31 age- and gender matched controls were enrolled in this study. Magnetic resonance imaging was used to acquire a 3D volume of the neck and of the whole-body. Cervical multifidi was used to represent muscles local to the whiplash trauma and all muscles below the hip joint, the lower extremities, were representing widespread muscles distal to the site of the trauma. The fat infiltration was determined by fat fraction in the segmented images. There was a linear correlation between local and distal muscle fat infiltration (p<0.001, r² = 0.28). The correlation remained significant when adjusting for age and WAD group (p = 0.009) as well as when correcting for age, WAD group and BMI (p = 0.002). There was a correlation between local and distal muscle fat infiltration within the severe WAD group (p = 0.0016, r² = 0.69) and in the healthy group (p = 0.022, r² = 0.17) but not in the mild/moderate group (p = 0.29, r² = 0.06). No significant differences (p = 0.11) in the lower extremities’ MFI between the different groups were found. The absence of differences between the groups in terms of lower extremities’ muscle fat infiltration indicates that, in this particular population, the whiplash trauma has a local effect on muscle fat infiltration rather than a generalized.