Fully Automated and Standardized Segmentation of Adipose Tissue Compartments via Deep Learning in 3D Whole-Body MRI of Epidemiologic Cohort Studies

Reproducibility of automatic adipose tissue segmentation using proton density fat fraction images between 1.5 and 3.0 T magnetic resonance

Background

Deep learning (DL)-based adipose tissue segmentation methods have shown great performance and efficacy for adipose tissue distribution analysis using magnetic resonance (MR) images, an important indicator of metabolic health and disease. The aim of this study was to evaluate the reproducibility of whole-body adipose tissue distribution analysis using proton density fat fraction (PDFF) images at different MR strengths.

Methods

A total of 24 volunteers were imaged using both 1.5 and 3.0 T clinical MR imaging (MRI) scanners at two sites. Whole-body PDFF images were acquired covering from neck to knee, and grouped into three subparts: thorax, abdomen, and thigh. The PDFF images were then segmented automatically into subcutaneous adipose tissue (SAT) and internal adipose tissue (IAT) using a U-Net DL model. The volumes of whole body (WH), total adipose tissue (TAT), SAT, and IAT for total body and each subpart were measured, and the volume ratio of TAT/WH, SAT/WH, IAT/WH, SAT/TAT, and IAT/SAT were also calculated. Additionally, the reproducibility of PDFF values of SAT and IAT for total body and subparts were evaluated.

Results

The intraclass correlation coefficient (ICC) and Pearson correlation coefficient of these volumes and volume ratios in whole-body between the two scanners were very close to one. The paired t-test and Bland-Altman plots for all comparisons showed no significant differences (P>0.05) when comparing the results from the 1.5 T scanner minus those from the 3.0 T scanner. The mean bias for WH, TAT, SAT, and IAT was -6.89 cm3 (P=0.95), -67.21 cm3 (P=0.40), 19.31 cm3 (P=0.74), and -18.84 cm3 (P=0.69), respectively. Good reproducibility performances were also found in each subpart, except for the indices of IAT volume, TAT/WH ratio, and SAT/TAT ratio in the thorax due to different susceptibility effects across MR strengths. The results also demonstrated good reproducibility between PDFF values of the two scanners with the mean bias for WH, thorax, abdomen, and thigh being -0.19% (P=0.219), -0.30% (P=0.118), 0.086% (P=0.494), and 0.24% (P=0.186) for SAT, respectively, as well as 0.35% (P=0.136), 0.46% (P=0.150), 0.58% (P=0.255), and 0.40% (P=0.169) for IAT, respectively.

Conclusions

Good reproducibility of whole-body adipose tissue distribution analysis using the DL method between 1.5 and 3.0 T MR images was demonstrated, which may facilitate the whole-body adipose tissue distribution analysis using the quantitative MR-PDFF images.

Using UK Biobank data to establish population-specific atlases from whole body MRI

BACKGROUND

Reliable reference data in medical imaging is largely unavailable. Developing tools that allow for the comparison of individual patient data to reference data has a high potential to improve diagnostic imaging. Population atlases are a commonly used tool in medical imaging to facilitate this. Constructing such atlases becomes particularly challenging when working with highly heterogeneous datasets, such as whole-body images, which contain significant anatomical variations.

METHOD

In this work, we propose a pipeline for generating a standardised whole-body atlas for a highly heterogeneous population by partitioning the population into anatomically meaningful subgroups. Using magnetic resonance images from the UK Biobank dataset, we create six whole-body atlases representing a healthy population average. We furthermore unbias them, and this way obtain a realistic representation of the population. In addition to the anatomical atlases, we generate probabilistic atlases that capture the distributions of abdominal fat (visceral and subcutaneous) and five abdominal organs across the population (liver, spleen, pancreas, left and right kidneys).

RESULTS

Our pipeline effectively generates high-quality, realistic whole-body atlases with clinical applicability. The probabilistic atlases show differences in fat distribution between subjects with medical conditions such as diabetes and cardiovascular diseases and healthy subjects in the atlas space.

CONCLUSIONS

With this work, we make the constructed anatomical and label atlases publically available, with the expectation that they will support medical research involving whole-body MR images.

Evaluating Sex-specific Differences in Abdominal Fat Volume and Proton Density Fat Fraction at MRI Using Automated nnU-Net-based Segmentation

Sex-specific abdominal organ volume and proton density fat fraction (PDFF) in people with obesity during a weight loss intervention was assessed with automated multiorgan segmentation of quantitative water-fat MRI. An nnU-Net architecture was employed for automatic segmentation of abdominal organs, including visceral and subcutaneous adipose tissue, liver, and psoas and erector spinae muscle, based on quantitative chemical shift-encoded MRI and using ground truth labels generated from participants of the Lifestyle Intervention (LION) study. Each organ's volume and fat content were examined in 127 participants (73 female and 54 male participants; body mass index, 30-39.9 kg/m2) and in 81 (54 female and 32 male participants) of these participants after an 8-week formula-based low-calorie diet. Dice scores ranging from 0.91 to 0.97 were achieved for the automatic segmentation. PDFF was found to be lower in visceral adipose tissue compared with subcutaneous adipose tissue in both male and female participants. Before intervention, female participants exhibited higher PDFF in subcutaneous adipose tissue (90.6% vs 89.7%; P < .001) and lower PDFF in liver (8.6% vs 13.3%; P < .001) and visceral adipose tissue (76.4% vs 81.3%; P < .001) compared with male participants. This relation persisted after intervention. As a response to caloric restriction, male participants lost significantly more visceral adipose tissue volume (1.76 L vs 0.91 L; P < .001) and showed a higher decrease in subcutaneous adipose tissue PDFF (2.7% vs 1.5%; P < .001) than female participants. Automated body composition analysis on quantitative water-fat MRI data provides new insights for understanding sex-specific metabolic response to caloric restriction and weight loss in people with obesity. Keywords: Obesity, Chemical Shift-encoded MRI, Abdominal Fat Volume, Proton Density Fat Fraction, nnU-Net ClinicalTrials.gov registration no. NCT04023942 Supplemental material is available for this article. Published under a CC BY 4.0 license.

Automated abdominal adipose tissue segmentation and volume quantification on longitudinal MRI using 3D convolutional neural networks with multi-contrast inputs

OBJECTIVE

Increased subcutaneous and visceral adipose tissue (SAT/VAT) volume is associated with risk for cardiometabolic diseases. This work aimed to develop and evaluate automated abdominal SAT/VAT segmentation on longitudinal MRI in adults with overweight/obesity using attention-based competitive dense (ACD) 3D U-Net and 3D nnU-Net with full field-of-view volumetric multi-contrast inputs.

MATERIALS AND METHODS

920 adults with overweight/obesity were scanned twice at multiple 3 T MRI scanners and institutions. The first scan was divided into training/validation/testing sets (n = 646/92/182). The second scan from the subjects in the testing set was used to evaluate the generalizability for longitudinal analysis. Segmentation performance was assessed by measuring Dice scores (DICE-SAT, DICE-VAT), false negatives (FN), and false positives (FP). Volume agreement was assessed using the intraclass correlation coefficient (ICC).

RESULTS

ACD 3D U-Net achieved rapid (< 4.8 s/subject) segmentation with high DICE-SAT (median ≥ 0.994) and DICE-VAT (median ≥ 0.976), small FN (median ≤ 0.7%), and FP (median ≤ 1.1%). 3D nnU-Net yielded rapid (< 2.5 s/subject) segmentation with similar DICE-SAT (median ≥ 0.992), DICE-VAT (median ≥ 0.979), FN (median ≤ 1.1%) and FP (median ≤ 1.2%). Both models yielded excellent agreement in SAT/VAT volume versus reference measurements (ICC > 0.997) in longitudinal analysis.

DISCUSSION

ACD 3D U-Net and 3D nnU-Net can be automated tools to quantify abdominal SAT/VAT volume rapidly, accurately, and longitudinally in adults with overweight/obesity.

Quantification of Visceral Fat at the L5 Vertebral Body Level in Patients with Crohn's Disease Using T2-Weighted MRI

The umbilical or L3 vertebral body level is often used for body fat quantification using computed tomography. To explore the feasibility of using clinically acquired pelvic magnetic resonance imaging (MRI) for visceral fat measurement, we examined the correlation of visceral fat parameters at the umbilical and L5 vertebral body levels. We retrospectively analyzed T2-weighted half-Fourier acquisition single-shot turbo spin echo (HASTE) MR axial images from Crohn's disease patients who underwent MRI enterography of the abdomen and pelvis over a three-year period. We determined the area/volume of subcutaneous and visceral fat from the umbilical and L5 levels and calculated the visceral fat ratio (VFR = visceral fat/subcutaneous fat) and visceral fat index (VFI = visceral fat/total fat). Statistical analyses involved correlation analysis between both levels, inter-rater analysis between two investigators, and inter-platform analysis between two image-analysis platforms. Correlational analysis of 32 patients yielded significant associations for VFI (r = 0.85; p < 0.0001) and VFR (r = 0.74; p < 0.0001). Intraclass coefficients for VFI and VFR were 0.846 and 0.875 (good agreement) between investigators and 0.831 and 0.728 (good and moderate agreement) between platforms. Our study suggests that the L5 level on clinically acquired pelvic MRIs may serve as a reference point for visceral fat quantification.

An Empirical Approach to Derive Water T1 from Multiparametric MR Images Using an Automated Pipeline and Comparison With Liver Stiffness

BACKGROUND

Water T1 of the liver has been shown to be promising in discriminating the progressive forms of fatty liver diseases, inflammation, and fibrosis, yet proper correction for iron and lipid is required.

PURPOSE

To examine the feasibility of an empirical approach for iron and lipid correction when measuring imaging-based T1 and to validate this approach by spectroscopy on in vivo data.

STUDY TYPE

Retrospective.

POPULATION

Next to mixed lipid-iron phantoms, individuals with different hepatic lipid content were investigated, including people with type 1 diabetes (N = 15, %female = 15.6, age = 43.5 ± 14.0), or type 2 diabetes mellitus (N = 21, %female = 28.9, age = 59.8 ± 9.7) and healthy volunteers (N = 9, %female = 11.1, age = 58.0 ± 8.1).

FIELD STRENGTH/SEQUENCES

3 T, balanced steady-state free precession MOdified Look-Locker Inversion recovery (MOLLI), multi- and dual-echo gradient echo Dixon, gradient echo magnetic resonance elastography (MRE).

ASSESSMENT

T1 values were measured in phantoms to determine the respective correction factors. The correction was tested in vivo and validated by proton magnetic resonance spectroscopy (1 H-MRS). The quantification of liver T1 based on automatic segmentation was compared to the T1 values based on manual segmentation. The association of T1 with MRE-derived liver stiffness was evaluated.

STATISTICAL TESTS

Bland-Altman plots and intraclass correlation coefficients (ICCs) were used for MOLLI vs. 1 H-MRS agreement and to compare liver T1 values from automatic vs. manual segmentation. Pearson's r correlation coefficients for T1 with hepatic lipids and liver stiffness were determined. A P-value of 0.05 was considered statistically significant.

RESULTS

MOLLI T1 values after correction were found in better agreement with the 1 H-MRS-derived water T1 (ICC = 0.60 [0.37; 0.76]) in comparison with the uncorrected T1 values (ICC = 0.18 [-0.09; 0.44]). Automatic quantification yielded similar liver T1 values (ICC = 0.9995 [0.9991; 0.9997]) as with manual segmentation. A significant correlation of T1 with liver stiffness (r = 0.43 [0.11; 0.67]) was found. A marked and significant reduction in the correlation strength of T1 with liver stiffness (r = 0.05 [-0.28; 0.38], P = 0.77) was found after correction for hepatic lipid content.

DATA CONCLUSION

Imaging-based correction factors enable accurate estimation of water T1 in vivo.

LEVEL OF EVIDENCE

1 TECHNICAL EFFICACY: Stage 1.

Automated Deep Learning-Based Segmentation of Abdominal Adipose Tissue on Dixon MRI in Adolescents: A Prospective Population-Based Study

BACKGROUND. The prevalence of childhood obesity has increased significantly worldwide, highlighting a need for accurate noninvasive quantification of body fat distribution in children. OBJECTIVE. The purpose of this study was to develop and test an automated deep learning method for subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) segmentation using Dixon MRI acquisitions in adolescents. METHODS. This study was embedded within the Generation R Study, a prospective population-based cohort study in Rotterdam, The Netherlands. The current study included 2989 children (1432 boys, 1557 girls; mean age, 13.5 years) who underwent investigational whole-body Dixon MRI after reaching the age of 13 years during the follow-up phase of the Generation R Study. A 2D competitive dense fully convolutional neural network model (2D-CDFNet) was trained from scratch to segment abdominal SAT and VAT using Dixon MRI-based images. The model underwent training, validation, and testing in 62, eight, and 15 children, respectively, who were selected by stratified random sampling, with manual segmentations used as reference. Segmentation performance was assessed using the Dice similarity coefficient and volumetric similarity. Two observers independently performed subjective visual assessments of automated segmentations in 504 children, selected by stratified random sampling, with undersegmentation and oversegmentation scored on a scale of 0-3 (with a score of 3 denoting nearly perfect segmentation). For 2820 children for whom complete data were available, Spearman correlation coefficients were computed among MRI measurements and BMI and dual-energy x-ray absorptiometry (DEXA)-based measurements. The model used (gitlab.com/radiology/msk/genr/abdomen/cdfnet) is publicly available. RESULTS. In the test dataset, the mean Dice similarity coefficient and mean volu-metric similarity, respectively, were 0.94 ± 0.03 [SD] and 0.98 ± 0.01 [SD] for SAT and 0.85 ± 0.05 and 0.92 ± 0.04 for VAT. The two observers assigned a score of 3 for SAT in 94% and 93% for the undersegmentation proportion and in 99% and 99% for the oversegmentation proportion, and they assigned a score of 3 for VAT in 99% and 99% for the undersegmentation proportion and in 95% and 97% for the oversegmentation proportion. Correlations with SAT and VAT were 0.808 and 0.698 for BMI and 0.941 and 0.801 for DEXA-derived fat mass. CONCLUSION. We trained and evaluated the 2D-CDFNet model on Dixon MRI in adolescents. Quantitative and qualitative measures of automated SAT and VAT segmentations indicated strong model performance. CLINICAL IMPACT. The automated model may facilitate large-scale studies investigating abdominal fat distribution on MRI among adolescents as well as associations of fat distribution with clinical outcomes.

Abdominal fat quantification using convolutional networks

OBJECTIVES

To present software for automated adipose tissue quantification of abdominal magnetic resonance imaging (MRI) data using fully convolutional networks (FCN) and to evaluate its overall performance-accuracy, reliability, processing effort, and time-in comparison with an interactive reference method.

MATERIALS AND METHODS

Single-center data of patients with obesity were analyzed retrospectively with institutional review board approval. Ground truth for subcutaneous (SAT) and visceral adipose tissue (VAT) segmentation was provided by semiautomated region-of-interest (ROI) histogram thresholding of 331 full abdominal image series. Automated analyses were implemented using UNet-based FCN architectures and data augmentation techniques. Cross-validation was performed on hold-out data using standard similarity and error measures.

RESULTS

The FCN models reached Dice coefficients of up to 0.954 for SAT and 0.889 for VAT segmentation during cross-validation. Volumetric SAT (VAT) assessment resulted in a Pearson correlation coefficient of 0.999 (0.997), relative bias of 0.7% (0.8%), and standard deviation of 1.2% (3.1%). Intraclass correlation (coefficient of variation) within the same cohort was 0.999 (1.4%) for SAT and 0.996 (3.1%) for VAT.

CONCLUSION

The presented methods for automated adipose-tissue quantification showed substantial improvements over common semiautomated approaches (no reader dependence, less effort) and thus provide a promising option for adipose tissue quantification.

CLINICAL RELEVANCE STATEMENT

Deep learning techniques will likely enable image-based body composition analyses on a routine basis. The presented fully convolutional network models are well suited for full abdominopelvic adipose tissue quantification in patients with obesity.

KEY POINTS

• This work compared the performance of different deep-learning approaches for adipose tissue quantification in patients with obesity. • Supervised deep learning-based methods using fully convolutional networks  were suited best. • Measures of accuracy were equal to or better than the operator-driven approach.

A Deep Learning Algorithm to Identify Anatomical Landmarks on Computed Tomography of the Temporal Bone

BACKGROUND

Petrous temporal bone cone-beam computed tomography scans help aid diagnosis and accurate identification of key operative landmarks in temporal bone and mastoid surgery. Our primary objective was to determine the accuracy of using a deep learning convolutional neural network algorithm to augment identification of structures on petrous temporal bone cone-beam computed tomography. Our secondary objective was to compare the accuracy of convolutional neural network structure identification when trained by a senior versus junior clinician.

METHODS

A total of 129 petrous temporal bone cone-beam computed tomography scans were obtained from an Australian public tertiary hospital. Key intraoperative landmarks were labeled in 68 scans using bounding boxes on axial and coronal slices at the level of the malleoincudal joint by an otolaryngology registrar and board-certified otolaryngologist. Automated structure identification was performed on axial and coronal slices of the remaining 61 scans using a convolutional neural network (Microsoft Custom Vision) trained using the labeled dataset. Convolutional neural network structure identification accuracy was manually verified by an otolaryngologist, and accuracy when trained by the registrar and otolaryngologist labeled datasets respectively was compared.

RESULTS

The convolutional neural network was able to perform automated structure identification in petrous temporal bone cone-beam computed tomography scans with a high degree of accuracy in both axial (0.958) and coronal (0.924) slices (P < .001). Convolutional neural network accuracy was proportionate to the seniority of the training clinician in structures with features more difficult to distinguish on single slices such as the cochlea, vestibule, and carotid canal.

CONCLUSION

Convolutional neural networks can perform automated structure identification in petrous temporal bone cone-beam computed tomography scans with a high degree of accuracy, with the performance being proportionate to the seniority of the training clinician. Training of the convolutional neural network by the most senior clinician is desirable to maximize the accuracy of the results.

A Review of the Clinical Applications of Artificial Intelligence in Abdominal Imaging

Artificial intelligence (AI) has been a topic of substantial interest for radiologists in recent years. Although many of the first clinical applications were in the neuro, cardiothoracic, and breast imaging subspecialties, the number of investigated and real-world applications of body imaging has been increasing, with more than 30 FDA-approved algorithms now available for applications in the abdomen and pelvis. In this manuscript, we explore some of the fundamentals of artificial intelligence and machine learning, review major functions that AI algorithms may perform, introduce current and potential future applications of AI in abdominal imaging, provide a basic understanding of the pathways by which AI algorithms can receive FDA approval, and explore some of the challenges with the implementation of AI in clinical practice.

Deep learning of image-derived measures of body composition in pediatric, adolescent, and young adult lymphoma: association with late treatment effects

OBJECTIVES

The objective of this study was to translate a deep learning (DL) approach for semiautomated analysis of body composition (BC) measures from standard of care CT images to investigate the prognostic value of BC in pediatric, adolescent, and young adult (AYA) patients with lymphoma.

METHODS

This 10-year retrospective, single-site study of 110 pediatric and AYA patients with lymphoma involved manual segmentation of fat and muscle tissue from 260 CT imaging datasets obtained as part of routine imaging at initial staging and first therapeutic follow-up. A DL model was trained to perform semiautomated image segmentation of adipose and muscle tissue. The association between BC measures and the occurrence of 3-year late effects was evaluated using Cox proportional hazards regression analyses.

RESULTS

DL-guided measures of BC were in close agreement with those obtained by a human rater, as demonstrated by high Dice scores (≥ 0.95) and correlations (r > 0.99) for each tissue of interest. Cox proportional hazards regression analyses revealed that patients with elevated subcutaneous adipose tissue at baseline and first follow-up, along with patients who possessed lower volumes of skeletal muscle at first follow-up, have increased risk of late effects compared to their peers.

CONCLUSIONS

DL provides rapid and accurate quantification of image-derived measures of BC that are associated with risk for treatment-related late effects in pediatric and AYA patients with lymphoma. Image-based monitoring of BC measures may enhance future opportunities for personalized medicine for children with lymphoma by identifying patients at the highest risk for late effects of treatment.

KEY POINTS

• Deep learning-guided CT image analysis of body composition measures achieved high agreement level with manual image analysis. • Pediatric patients with more fat and less muscle during the course of cancer treatment were more likely to experience a serious adverse event compared to their clinical counterparts. • Deep learning of body composition may add value to routine CT imaging by offering real-time monitoring of pediatric, adolescent, and young adults at high risk for late effects of cancer treatment.

Whole-Body Magnetic Resonance Imaging Assessment of the Contributions of Adipose and Nonadipose Tissues to Cardiovascular Remodeling in Adolescents

Background Greater body mass index is associated with cardiovascular remodeling in adolescents. However, body mass index cannot differentiate between adipose and nonadipose tissues. We examined how visceral and subcutaneous adipose tissue are linked with markers of early cardiovascular remodeling, independently from nonadipose tissue. Methods and Results Whole-body magnetic resonance imaging was done in 82 adolescents (39 overweight/obese; 36 female; median age, 16.3 [interquartile range, 14.4-18.1] years) to measure body composition and cardiovascular remodeling markers. Left ventricular diastolic function was assessed by echocardiography. Waist, waist:height ratio, and body mass index z scores were calculated. Residualized nonadipose tissue, subcutaneous adipose tissue, and visceral adipose tissue variables, uncorrelated with each other, were constructed using partial regression modeling to allow comparison of their individual contributions in a 3-compartment body composition model. Cardiovascular variables mostly related to nonadipose rather than adipose tissue. Nonadipose tissue was correlated positively with left ventricular mass (r=0.81), end-diastolic volume (r=0.70), stroke volume (r=0.64), left ventricular mass:end-diastolic volume (r=0.37), and systolic blood pressure (r=0.35), and negatively with heart rate (r=-0.33) (all P<0.01). Subcutaneous adipose tissue was associated with worse left ventricular diastolic function (r=-0.42 to -0.48, P=0.0007-0.02) and higher heart rates (r=0.34, P=0.007) but linked with better systemic vascular resistance (r=-0.35, P=0.006). There were no significant relationships with visceral adipose tissue and no associations of any compartment with pulse wave velocity. Conclusions Simple anthropometry does not reflect independent effects of nonadipose tissue and subcutaneous adipose tissue on the adolescent cardiovascular system. This could result in normal cardiovascular adaptations to growth being misinterpreted as pathological sequelae of excess adiposity in studies reliant on such measures.

More than skin-deep: visceral fat is strongly associated with disease activity, function and metabolic indices in psoriatic disease

OBJECTIVE

To compare body composition between patients with psoriatic disease (PsD), including cutaneous psoriasis (PsO) and psoriatic arthritis (PsA), and controls, and to explore associations between disease activity and measures of function and metabolic derangement.

METHODS

Body composition was assessed by air displacement plethysmography (ADP) and MRI-derived fat segmentation using an automated pipeline (FatSegNet). Function was assessed by Health Assessment Questionnaire (HAQ) and metabolic status by fasting lipid profile, insulin and adiponectin. Active and inactive PsO and PsA were defined by body surface area (BSA) and Psoriasis Area Severity Index (PASI) and minimal disease activity (MDA), respectively.

RESULTS

Thirty patients (median disease duration 15 years; median age 52 years) and 30 BMI-matched controls were enrolled. Compared with controls, all MRI-derived body composition parameters-whole-body volume, subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), abdominal adipose tissue (AAT), VAT/AAT and VAT/SAT-were higher in the PsD group, specifically, those with active disease. Body mass, body fat, whole-body volume and whole-body VAT were correlated with higher triglycerides, cholesterol:HDL (high-density lipoprotein), insulin resistance and lower adiponectin as well as higher HAQ and lower MDA.

CONCLUSIONS

In this pilot study, patients with PsD revealed excessive total adipose tissue and a greater volume of metabolically unfavourable ectopic fat, including VAT, compared with BMI-matched controls, which also correlated with HAQ, disease activity and overall dysmetabolism. We also provide the first evidence in patients with PsD for the clinical application of FatSegNet: a novel, automated and rapid deep learning pipeline for providing accurate MRI-based measurement of fat segmentation. Our findings suggest the need for a more integrated approach to the management of PsD, which considers both the metabolic and inflammatory burden of disease. More specifically, visceral fat is a surrogate marker of uncontrolled PsD and may be an important future target for both pharmacological and lifestyle interventions.

Automated whole-volume measurement of CT fat fraction of the pancreas: correlation with Dixon MR imaging

OBJECTIVES

This study aimed to assess the feasibility of pancreatic steatosis quantification by automated whole-volume measurement of the fat fraction of the pancreas on CT in comparison to MRI using proton-density fat fraction (PDFF) techniques.

METHODS

Fifty-nine patients who underwent both CT and MRI were analyzed. Automated whole-volume measurement of pancreatic fat on unenhanced CT was performed by a histogram analysis with local thresholding. Three sets of CT fat volume fraction (FVF) (%) values with thresholds of -30 Hounsfield unit (HU), -20 HU and -10 HU were compared to MR-FVF (%) values measured on a PDFF map.

RESULTS

The median -30 HU CT-FVF, -20 HU CT-FVF, -10 HU CT-FVF and MR-FVF values of the pancreas were 8.6% (interquartile range (IQR), 11.3), 10.5% (IQR, 13.2), 13.4% (IQR, 16.1) and 10.9% (IQR, 9.7), respectively. The -30 HU CT-FVF (%), -20 HU CT-FVF (%) and -10 HU CT-FVF (%) of the pancreas showed a significant positive correlation with the MR-FVF (%) of the pancreas (ρ = 0.898, p < 0.001, ρ = 0.905, p < 0.001, ρ = 0.909, p < 0.001, respectively). The -20 HU CT-FVF (%) displayed reasonable agreement with the MR-FVF (%) with a low absolute fixed bias (mean difference, 0.32%; limit of agreement from -10.1 to 10.7%).

CONCLUSION

The automated whole-volume measurement of the CT fat fraction of the pancreas using the threshold CT attenuation value of -20 HU may be a feasible, non-invasive, and convenient technique for quantifying pancreatic steatosis.

ADVANCES IN KNOWLEDGE

CT-FVF value of the pancreas had a positive correlation with the MR-FVF value. The -20 HU CT-FVF may be a convenient technique for quantifying pancreatic steatosis.

Deep learning-based automatic pipeline for quantitative assessment of thigh muscle morphology and fatty infiltration

PURPOSE

Fast and accurate thigh muscle segmentation from MRI is important for quantitative assessment of thigh muscle morphology and composition. A novel deep learning (DL) based thigh muscle and surrounding tissues segmentation model was developed for fully automatic and reproducible cross-sectional area (CSA) and fat fraction (FF) quantification and tested in patients at 10 years after anterior cruciate ligament reconstructions.

METHODS

A DL model combining UNet and DenseNet was trained and tested using manually segmented thighs from 16 patients (32 legs). Segmentation accuracy was evaluated using Dice similarity coefficients (DSC) and average symmetric surface distance (ASSD). A UNet model was trained for comparison. These segmentations were used to obtain CSA and FF quantification. Reproducibility of CSA and FF quantification was tested with scan and rescan of six healthy subjects.

RESULTS

The proposed UNet and DenseNet had high agreement with manual segmentation (DSC >0.97, ASSD < 0.24) and improved performance compared with UNet. For hamstrings of the operated knee, the automated pipeline had largest absolute difference of 6.01% for CSA and 0.47% for FF as compared to manual segmentation. In reproducibility analysis, the average difference (absolute) in CSA quantification between scan and rescan was better for the automatic method as compared with manual segmentation (2.27% vs. 3.34%), whereas the average difference (absolute) in FF quantification were similar.

CONCLUSIONS

The proposed method exhibits excellent accuracy and reproducibility in CSA and FF quantification compared with manual segmentation and can be used in large-scale patient studies.

Analysis of volume and topography of adipose tissue in the trunk: Results of MRI of 11,141 participants in the German National Cohort

This research addresses the assessment of adipose tissue (AT) and spatial distribution of visceral (VAT) and subcutaneous fat (SAT) in the trunk from standardized magnetic resonance imaging at 3 T, thereby demonstrating the feasibility of deep learning (DL)-based image segmentation in a large population-based cohort in Germany (five sites). Volume and distribution of AT play an essential role in the pathogenesis of insulin resistance, a risk factor of developing metabolic/cardiovascular diseases. Cross-validated training of the DL-segmentation model led to a mean Dice similarity coefficient of >0.94, corresponding to a mean absolute volume deviation of about 22 ml. SAT is significantly increased in women compared to men, whereas VAT is increased in males. Spatial distribution shows age- and body mass index-related displacements. DL-based image segmentation provides robust and fast quantification of AT (≈15 s per dataset versus 3 to 4 hours for manual processing) and assessment of its spatial distribution from magnetic resonance images in large cohort studies.

CNN-based fully automatic wrist cartilage volume quantification in MR images: A comparative analysis between different CNN architectures

PURPOSE

Automatic measurement of wrist cartilage volume in MR images.

METHODS

We assessed the performance of four manually optimized variants of the U-Net architecture, nnU-Net and Mask R-CNN frameworks for the segmentation of wrist cartilage. The results were compared to those from a patch-based convolutional neural network (CNN) we previously designed. The segmentation quality was assessed on the basis of a comparative analysis with manual segmentation. The best networks were compared using a cross-validation approach on a dataset of 33 3D VIBE images of mostly healthy volunteers. Influence of some image parameters on the segmentation reproducibility was assessed.

RESULTS

The U-Net-based networks outperformed the patch-based CNN in terms of segmentation homogeneity and quality, while Mask R-CNN did not show an acceptable performance. The median 3D DSC value computed with the U-Net_AL (0.817) was significantly larger than DSC values computed with the other networks. In addition, the U-Net_AL provided the lowest mean volume error (17%) and the highest Pearson correlation coefficient (0.765) with respect to the ground truth values. Of interest, the reproducibility computed using U-Net_AL was larger than the reproducibility of the manual segmentation. Moreover, the results indicate that the MRI-based wrist cartilage volume is strongly affected by the image resolution.

CONCLUSIONS

U-Net CNN with attention layers provided the best wrist cartilage segmentation performance. In order to be used in clinical conditions, the trained network can be fine-tuned on a dataset representing a group of specific patients. The error of cartilage volume measurement should be assessed independently using a non-MRI method.

Automated volume measurement of abdominal adipose tissue from entire abdominal cavity in Dixon MR images using deep learning

The purpose of this study was to realize an automated volume measurement of abdominal adipose tissue from the entire abdominal cavity in Dixon magnetic resonance (MR) images using deep learning. Our algorithm involves a combination of extraction of the abdominal cavity and body trunk regions using deep learning and extraction of a fat region based on automatic thresholding. To evaluate the proposed method, we calculated the Dice coefficient (DC) between the extracted regions using deep learning and labeled images. We also compared the visceral adipose tissue (VAT) and subcutaneous adipose tissue volumes calculated by employing the proposed method with those calculated from computed tomography (CT) images scanned on the same day using the automatic calculation method previously developed by our group. We implemented our method as a plug-in in a web-based medical image processing platform. The DCs of the abdominal cavity and body trunk regions were 0.952 ± 0.014 and 0.995 ± 0.002, respectively. The VAT volume measured from MR images using the proposed method was almost equivalent to that measured from CT images. The time required for our plug-in to process the test set was 118.9 ± 28.0 s. Using our proposed method, the VAT volume measured from MR images can be an alternative to that measured from CT images.

Automated imaging-based abdominal organ segmentation and quality control in 20,000 participants of the UK Biobank and German National Cohort Studies

Large epidemiological studies such as the UK Biobank (UKBB) or German National Cohort (NAKO) provide unprecedented health-related data of the general population aiming to better understand determinants of health and disease. As part of these studies, Magnetic Resonance Imaging (MRI) is performed in a subset of participants allowing for phenotypical and functional characterization of different organ systems. Due to the large amount of imaging data, automated image analysis is required, which can be performed using deep learning methods, e. g. for automated organ segmentation. In this paper we describe a computational pipeline for automated segmentation of abdominal organs on MRI data from 20,000 participants of UKBB and NAKO and provide results of the quality control process. We found that approx. 90% of data sets showed no relevant segmentation errors while relevant errors occurred in a varying proportion of data sets depending on the organ of interest. Image-derived features based on automated organ segmentations showed relevant deviations of varying degree in the presence of segmentation errors. These results show that large-scale, deep learning-based abdominal organ segmentation on MRI data is feasible with overall high accuracy, but visual quality control remains an important step ensuring the validity of down-stream analyses in large epidemiological imaging studies.

Imaging of Colorectal Liver Metastasis

Colorectal cancer (CRC) is one of the most common cancers in the world. The most important determinant of survival and prognosis is the stage and presence of metastasis. The liver is the most common location for CRC metastasis. The only curative treatment for CRC liver metastasis (CRLM) is resection; however, many patients are ineligible for surgical resection of CRLM. Locoregional treatments such as ablation and intra-arterial therapy are also available for patients with CRLM. Assessment of response after chemotherapy is challenging due to anatomical and functional changes. Antiangiogenic agents such as bevacizumab that are used in the treatment of CRLM may show atypical patterns of response on imaging. It is vital to distinguish patterns of response in addition to toxicities to various treatments. Imaging plays a critical role in evaluating the characteristics of CRLM and the approach to treatment. CT is the modality of choice in the diagnosis and management of CRLM. MRI is best used for indeterminate lesions and to assess response to intra-arterial therapy. PET-CT is often utilized to detect extrahepatic metastasis. State-of-the-art imaging is critical to characterize patterns of response to various treatments. We herein review the imaging characteristics of CRLM with an emphasis on imaging changes following the most common CRLM treatments.

3D Neural Networks for Visceral and Subcutaneous Adipose Tissue Segmentation using Volumetric Multi-Contrast MRI

Individuals with obesity have larger amounts of visceral (VAT) and subcutaneous adipose tissue (SAT) in their body, increasing the risk for cardiometabolic diseases. The reference standard to quantify SAT and VAT uses manual annotations of magnetic resonance images (MRI), which requires expert knowledge and is time-consuming. Although there have been studies investigating deep learning-based methods for automated SAT and VAT segmentation, the performance for VAT remains suboptimal (Dice scores of 0.43 to 0.89). Previous work had key limitations of not fully considering the multi-contrast information from MRI and the 3D anatomical context, which are critical for addressing the complex spatially varying structure of VAT. An additional challenge is the imbalance between the number and distribution of pixels representing SAT/VAT. This work proposes a network based on 3D U-Net that utilizes the full field-of-view volumetric T₁-weighted, water, and fat images from dual-echo Dixon MRI as the multi-channel input to automatically segment SAT and VAT in adults with overweight/obesity. In addition, this work extends the 3D U-Net to a new Attention-based Competitive Dense 3D U-Net (ACD 3D U-Net) trained with a class frequency-balancing Dice loss (FBDL). In an initial testing dataset, the proposed 3D U-Net and ACD 3D U-Net with FBDL achieved 3D Dice scores of (mean ± standard deviation) 0.99 ±0.01 and 0.99±0.01 for SAT, and 0.95±0.04 and 0.96 ±0.04 for VAT, respectively, compared to manual annotations. The proposed 3D networks had rapid inference time (<60 ms/slice) and can enable automated segmentation of SAT and VAT.Clinical relevance- This work developed 3D neural networks to automatically, accurately, and rapidly segment visceral and subcutaneous adipose tissue on MRI, which can help to characterize the risk for cardiometabolic diseases such as diabetes, elevated glucose levels, and hypertension.

Clinical Artificial Intelligence Applications in Radiology: Chest and Abdomen

Organ segmentation, chest radiograph classification, and lung and liver nodule detections are some of the popular artificial intelligence (AI) tasks in chest and abdominal radiology due to the wide availability of public datasets. AI algorithms have achieved performance comparable to humans in less time for several organ segmentation tasks, and some lesion detection and classification tasks. This article introduces the current published articles of AI applied to chest and abdominal radiology, including organ segmentation, lesion detection, classification, and predicting prognosis.

Automated Segmentation of Visceral, Deep Subcutaneous, and Superficial Subcutaneous Adipose Tissue Volumes in MRI of Neonates and Young Children

Purpose

To develop and evaluate an automated segmentation method for accurate quantification of abdominal adipose tissue (AAT) depots (superficial subcutaneous adipose tissue [SSAT], deep subcutaneous adipose tissue [DSAT], and visceral adipose tissue [VAT]) in neonates and young children.

Materials and Methods

This was a secondary analysis of prospectively collected data, which used abdominal MRI data from Growing Up in Singapore Towards healthy Outcomes, or GUSTO, a longitudinal mother-offspring cohort, to train and evaluate a convolutional neural network for volumetric AAT segmentation. The data comprised imaging volumes of 333 neonates obtained at early infancy (age ≤2 weeks, 180 male neonates) and 755 children aged either 4.5 years (n = 316, 150 male children) or 6 years (n = 439, 219 male children). The network was trained on images of 761 randomly selected volumes (neonates and children combined) and evaluated on 100 neonatal volumes and 227 child volumes by using 10-fold validation. Automated segmentations were compared with expert-generated manual segmentation. Segmentation performance was assessed using Dice scores.

Results

When the model was tested on the test datasets across the 10 folds, the model had strong agreement with the ground truth for all testing sets, with mean Dice similarity scores for SSAT, DSAT, and VAT, respectively, of 0.960, 0.909, and 0.872 in neonates and 0.944, 0.851, and 0.960 in children. The model generalized well to different body sizes and ages and to all abdominal levels.

Conclusion

The proposed segmentation approach provided accurate automated volumetric assessment of AAT compartments on MR images of neonates and children.Keywords Pediatrics, Deep Learning, Convolutional Neural Networks, Water-Fat MRI, Image Segmentation, Deep and Superficial Subcutaneous Adipose Tissue, Visceral Adipose TissueClinical trial registration no. NCT01174875 Supplemental material is available for this article. © RSNA, 2021.

Artificial intelligence and abdominal adipose tissue analysis: a literature review

Body composition imaging relies on assessment of tissues composition and distribution. Quantitative data provided by body composition imaging analysis have been linked to pathogenesis, risk, and clinical outcomes of a wide spectrum of diseases, including cardiovascular and oncologic. Manual segmentation of imaging data allows to obtain information on abdominal adipose tissue; however, this procedure can be cumbersome and time-consuming. On the other hand, quantitative imaging analysis based on artificial intelligence (AI) has been proposed as a fast and reliable automatic technique for segmentation of abdominal adipose tissue compartments, possibly improving the current standard of care. AI holds the potential to extract quantitative data from computed tomography (CT) and magnetic resonance (MR) images, which in most of the cases are acquired for other purposes. This information is of great importance for physicians dealing with a wide spectrum of diseases, including cardiovascular and oncologic, for the assessment of risk, pathogenesis, clinical outcomes, response to treatments, and complications. In this review we summarize the available evidence on AI algorithms aimed to the segmentation of visceral and subcutaneous adipose tissue compartments on CT and MR images.

Automatic segmentation of whole-body adipose tissue from magnetic resonance fat fraction images based on machine learning

OBJECTIVE

To propose a fully automated algorithm, which is implemented to segment subcutaneous adipose tissue (SAT) and internal adipose tissue (IAT) from the total adipose tissue for whole-body fat distribution analysis using proton density fat fraction (PDFF) magnetic resonance images.

MATERIALS AND METHODS

Adipose tissue segmentation was implemented using the U-Net deep neural network model. All datasets were collected using a 3.0 T magnetic resonance imaging (MRI) scanner for whole-body scan of 20 volunteers covering from neck to knee with about 160 images for each volunteer. PDFF images were reconstructed based on chemical-shift-encoded fat-water imaging. After selecting the representative PDFF images (total 906 images), the manual labeling of the SAT area was used for model training (504 images), validation (168 images), and testing (234 images).

RESULTS

The automatic segmentation model was validated through three indices using the validation and test sets. The dice similarity coefficient, precision rate, and recall rate were 0.976 ± 0.048, 0.978 ± 0.048, and 0.978 ± 0.050, respectively, in both validation and test sets.

CONCLUSION

The proposed algorithm can reliably and automatically segment SAT and IAT from whole-body MRI PDFF images. The proposed method provides a simple and automatic tool for whole-body fat distribution analysis to explore the relationship between fat deposition and metabolic-related chronic diseases.

Role of 3D Printing and Modeling to Aid in Neuroradiology Education for Medical Trainees

BACKGROUND

Applications of 3-dimensional (3D) printing in medical imaging and health care are expanding. Currently, primary uses involve presurgical planning and patient and medical trainee education. Neuroradiology is a complex subdiscipline of radiology that requires further training beyond radiology residency. This review seeks to explore the clinical value of 3D printing and modeling specifically in enhancing neuroradiology education for radiology physician residents and medical trainees.

METHODS

A brief review summarizing the key steps from radiologic image to 3D printed model is provided, including storage of computed tomography and magnetic resonance imaging data as digital imaging and communications in medicine files; conversion to standard tessellation language (STL) format; manipulation of STL files in interactive medical image control system software (Materialise) to create 3D models; and 3D printing using various resins via a Formlabs 2 printer.

RESULTS

For the purposes of demonstration and proof of concept, neuroanatomy models deemed crucial in early radiology education were created via open-source hardware designs under free or open licenses. 3D-printed objects included a sphenoid bone, cerebellum, skull base, middle ear labyrinth and ossicles, mandible, circle of Willis, carotid aneurysm, and lumbar spine using a combination of clear, white, and elastic resins.

CONCLUSIONS

Based on this single-institution experience, 3D-printed complex neuroanatomical structures seem feasible and may enhance resident education and patient safety. These same steps and principles may be applied to other subspecialties of radiology. Artificial intelligence also has the potential to advance the 3D process.