Development and Validation of a Deep Learning System for Segmentation of Abdominal Muscle and Fat on Computed Tomography

Using a new artificial intelligence-aided method to assess body composition CT segmentation in colorectal cancer patients

INTRODUCTION

This study aimed to evaluate the accuracy of our own artificial intelligence (AI)-generated model to assess automated segmentation and quantification of body composition-derived computed tomography (CT) slices from the lumber (L3) region in colorectal cancer (CRC) patients.

METHODS

A total of 541 axial CT slices at the L3 vertebra were retrospectively collected from 319 patients with CRC diagnosed during 2012-2019 at a single Australian tertiary institution, Western Health in Melbourne. A two-dimensional U-Net convolutional network was trained on 338 slices to segment muscle, visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT). Manual reading of these same slices of muscle, VAT and SAT was created to serve as ground truth data. The Dice similarity coefficient was used to assess the U-Net-based segmentation performance on both a validation dataset (68 slices) and a test dataset (203 slices). The measurement of cross-sectional area and Hounsfield unit (HU) density of muscle, VAT and SAT were compared between two methods.

RESULTS

The segmentation for muscle, VAT and SAT demonstrated excellent performance for both the validation (Dice similarity coefficients >0.98, respectively) and test (Dice similarity coefficients >0.97, respectively) datasets. There was a strong positive correlation between manual and AI segmentation measurements of body composition for both datasets (Spearman's correlation coefficients: 0.944-0.999, P < 0.001).

CONCLUSIONS

Compared to the gold standard, this fully automated segmentation system exhibited a high accuracy for assessing segmentation and quantification of abdominal muscle and adipose tissues of CT slices at the L3 in CRC patients.

Multi-organ segmentation: a progressive exploration of learning paradigms under scarce annotation

Precise delineation of multiple organs or abnormal regions in the human body from medical images plays an essential role in computer-aided diagnosis, surgical simulation, image-guided interventions, and especially in radiotherapy treatment planning. Thus, it is of great significance to explore automatic segmentation approaches, among which deep learning-based approaches have evolved rapidly and witnessed remarkable progress in multi-organ segmentation. However, obtaining an appropriately sized and fine-grained annotated dataset of multiple organs is extremely hard and expensive. Such scarce annotation limits the development of high-performance multi-organ segmentation models but promotes many annotation-efficient learning paradigms. Among these, studies on transfer learning leveraging external datasets, semi-supervised learning including unannotated datasets and partially-supervised learning integrating partially-labeled datasets have led the dominant way to break such dilemmas in multi-organ segmentation. We first review the fully supervised method, then present a comprehensive and systematic elaboration of the 3 abovementioned learning paradigms in the context of multi-organ segmentation from both technical and methodological perspectives, and finally summarize their challenges and future trends.

Visceral obesity and sarcopenia as predictors of efficacy and hematological toxicity in patients with metastatic breast cancer treated with CDK 4/6 inhibitors

PURPOSE

We aimed to investigate whether visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), and skeletal muscle area (SMA) index are predictive for efficacy and hematological toxicity in ER + HER2-metastatic breast cancer (BC) patients who received CDK 4/6 inhibitors.

METHODS

This retrospective cohort study analyzed 52 patients who were treated with CDK 4/6 inhibitors between January 2018 and February 2021. The values of VAT, SAT, SMA indices and hematological parameters were noted before the start, at the third and sixth months of this treatment. The skeletal muscle area (SMA) and adipose tissue measurements were calculated at the level of the third lumbar vertebra. A SMA-index value of <40 cm2/m2 was accepted as the threshold value for sarcopenia.

RESULTS

Patients with sarcopenia had a worse progression-free survival (PFS) compared to patients without sarcopenia (19.6 vs. 9.0 months, p = 0.005). Patients with a high-VAT-index had a better PFS (20.4 vs. 9.3 months, p = 0.033). Only the baseline low-SMA- index (HR: 3.89; 95% CI: 1.35-11.25, p = 0.012) and baseline low-VAT-index (HR: 2.15; 95% CI: 1.02-4.53, p = 0.042) had significantly related to poor PFS in univariate analyses. The low-SMA-index was the only independent factor associated with poor PFS (HR: 3.99; 95% CI: 1.38-11.54, p = 0.011). No relationship was observed between body composition parameters and grade 3-4 hematological toxicity.

CONCLUSION

The present study supported the significance of sarcopenia and low visceral adipose tissue as potential early indicators of poor PFS in patients treated with CDK 4/6 inhibitors.

Abdominal adipose tissue and type 2 diabetic kidney disease: adipose radiology assessment, impact, and mechanisms

Diabetic kidney disease (DKD) is a significant healthcare burden worldwide that substantially increases the risk of kidney failure and cardiovascular events. To reduce the prevalence of DKD, extensive research is being conducted to determine the risk factors and consequently implement early interventions. Patients with type 2 diabetes mellitus (T2DM) are more likely to be obese. Abdominal adiposity is associated with a greater risk of kidney damage than general obesity. Abdominal adipose tissue can be divided into different fat depots according to the location and function, including visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), perirenal adipose tissue (PAT), and renal sinus adipose tissue (RSAT), which can be accurately measured by radiology techniques, such as computed tomography (CT) and magnetic resonance imaging (MRI). Abdominal fat depots may affect the development of DKD through different mechanisms, and radiologic abdominal adipose characteristics may serve as imaging indicators of DKD risk. This review will first describe the CT/MRI-based assessment of abdominal adipose depots and subsequently describe the current studies on abdominal adipose tissue and DKD development, as well as the underlying mechanisms in patients of T2DM with DKD.

Association between visceral obesity and tumor recurrence in hepatocellular carcinoma recipients undergoing liver transplantation

BACKGROUND

Excessive visceral obesity in recipients of living donor liver transplantation (LDLT) is associated with mortality, and a recent study reported the correlation between visceral adiposity of male LDLT recipients and hepatocellular carcinoma (HCC) recurrence. However, there is no study on the relationship between the donor's visceral adiposity and surgical outcomes in LDLT recipients. We investigated the association of the visceral-to-subcutaneous fat area ratio (VSR) in donors and recipients with HCC recurrence and mortality in LDLT.

METHODS

We analyzed 1386 sets of donors and recipients who underwent LDLT between January 2008 and January 2018. The maximal chi-square method was used to determine the optimal cutoff values for VSR for predicting overall HCC recurrence and mortality. Cox regression analyses were performed to evaluate the association of donor VSR and recipient VSR with overall HCC recurrence and mortality in recipients.

RESULTS

The cutoff values of VSR was determined as 0.73 in males and 0.31 in females. High donor VSR was significantly associated with overall HCC recurrence (adjusted hazard ratio [HR]: 1.43, 95% confidence interval [CI]: 1.06-1.93, p = 0.019) and mortality (HR: 1.35, 95% CI: 1.03-1.76, p = 0.030). High recipient VSR was significantly associated with overall HCC recurrence (HR: 1.40, 95% CI: 1.04-1.88, p = 0.027) and mortality (HR: 1.50, 95% CI: 1.14-1.96, p = 0.003).

CONCLUSIONS

Both recipient VSR and donor VSR were significant risk factors for HCC recurrence and mortality in LDLT recipients. Preoperative donor VSR and recipient VSR may be strong predictors of the surgical outcomes of LDLT recipients with HCC.

Obesity paradox in patients with non-small cell lung cancer undergoing immune checkpoint inhibitor therapy

BACKGROUND

The obesity paradox in patients with advanced non-small cell lung cancer receiving immune checkpoint inhibitor therapy has been observed, but its underlying mechanism is not fully understood. We aimed to investigate whether body composition affects the prognostic impact of obesity, as determined by body mass index (BMI), on survival.

METHODS

This retrospective study evaluated the data collected from Asian patients who were treated with immune checkpoint inhibitors for advanced non-small cell lung cancer between October 2015 and October 2021. We used abdominal cross-sectional imaging to calculate the skeletal muscle and visceral fat indices (cm2 /m2 ) by dividing the cross-sectional areas of the skeletal muscle and visceral fat by the height squared. Cox proportional-hazards regression was performed to determine the correlation between BMI according to the Asia-Pacific classification, body composition metrics and overall survival.

RESULTS

We analysed the data of 820 patients (630 men and 190 women, with a mean age of 64.3 years [standard deviation: 10.4 years]) and observed 572 (69.8%) deaths with the 1-year mortality rate of 0.58 (95% confidence interval, 0.55-0.62). Obese BMI was associated with longer overall survival, independent of clinical covariates (hazard ratio, 0.64; 95% confidence interval: 0.52-0.80). The prognostic value of obese BMI remained after additional adjustments for skeletal muscle index (hazard ratio, 0.68; 95% confidence interval, 0.53-0.87) or visceral fat index (hazard ratio, 0.54; 95% confidence interval: 0.41-0.70). No association was observed between sex and the impact of BMI on overall survival (P-value for interaction >0.05).

CONCLUSIONS

In Asian patients with advanced non-small cell lung cancer who received immune checkpoint inhibitors, obese BMI was associated with favourable overall survival independent of skeletal muscle or visceral fat mass.

Reporting Quality of Research Studies on AI Applications in Medical Images According to the CLAIM Guidelines in a Radiology Journal With a Strong Prominence in Asia

OBJECTIVE

We aimed to evaluate the reporting quality of research articles that applied deep learning to medical imaging. Using the Checklist for Artificial Intelligence in Medical Imaging (CLAIM) guidelines and a journal with prominence in Asia as a sample, we intended to provide an insight into reporting quality in the Asian region and establish a journal-specific audit.

MATERIALS AND METHODS

A total of 38 articles published in the Korean Journal of Radiology between June 2018 and January 2023 were analyzed. The analysis included calculating the percentage of studies that adhered to each CLAIM item and identifying items that were met by ≤ 50% of the studies. The article review was initially conducted independently by two reviewers, and the consensus results were used for the final analysis. We also compared adherence rates to CLAIM before and after December 2020.

RESULTS

Of the 42 items in the CLAIM guidelines, 12 items (29%) were satisfied by ≤ 50% of the included articles. None of the studies reported handling missing data (item #13). Only one study respectively presented the use of de-identification methods (#12), intended sample size (#19), robustness or sensitivity analysis (#30), and full study protocol (#41). Of the studies, 35% reported the selection of data subsets (#10), 40% reported registration information (#40), and 50% measured inter and intrarater variability (#18). No significant changes were observed in the rates of adherence to these 12 items before and after December 2020.

CONCLUSION

The reporting quality of artificial intelligence studies according to CLAIM guidelines, in our study sample, showed room for improvement. We recommend that the authors and reviewers have a solid understanding of the relevant reporting guidelines and ensure that the essential elements are adequately reported when writing and reviewing the manuscripts for publication.

Assessment of skeletal muscle using deep learning on low-dose CT images

The visceral fat area obtained by computed tomography (CT) at the navel level is clinically used as an indicator of visceral fat obesity in Japan. Analysis of skeletal muscle mass using CT images at the navel level may potentially support concurrent assessment of sarcopenia and sarcopenic obesity. The purpose of this study was to assess the performance of deep learning models (DLMs) for skeletal muscle mass measurement using low-dose abdominal CT. The primary dataset used in this study included 11,494 low-dose abdominal CT images at navel level acquired in 7,370 subjects for metabolic syndrome screening. The publicly available Cancer Imaging Archive (TCIA) dataset, including 5,801 abdominal CT images, was used as a complementary dataset. For abdominal CT image segmentation, we used the SegU-net DLM with different filter size and hierarchical depth. The segmentation accuracy was assessed by measuring the dice similarity coefficient (DSC), cross-sectional area (CSA) error, and Bland-Altman plots. The proposed DLM achieved a DSC of 0.992 ± 0.012, a CSA error of 0.41 ± 1.89%, and a Bland-Altman percent difference of -0.1 ± 3.8%. The proposed DLM was able to automatically segment skeletal muscle mass measurements from low-dose abdominal CT with high accuracy.

Identification of Differences in Body Composition Measures Using 3D-Derived Artificial Intelligence from Multiple CT Scans across the L3 Vertebra Compared to a Single Mid-Point L3 CT Scan

Purpose

Body composition analysis in colorectal cancer (CRC) typically utilises a single 2D-abdominal axial CT slice taken at the mid-L3 level. The use of artificial intelligence (AI) allows for analysis of the entire L3 vertebra (non-mid-L3 and mid-L3). The goal of this study was to determine if the use of an AI approach offered any additional information on capturing body composition measures.

Methods

A total of 2203 axial CT slices of the entire L3 level (4-46 slices were available per patient) were retrospectively collected from 203 CRC patients treated at Western Health, Melbourne (97 males; 47.8%). A pretrained artificial intelligence (AI) model was used to segment muscle, visceral adipose tissue (VAT), and subcutaneous adipose tissue (SAT) on these slices. The difference in body composition measures between mid-L3 and non-mid-L3 scans was compared for each patient, and for males and females separately.

Results

Body composition measures derived from non-mid-L3 scans exhibited a median range of 0.85% to 6.28% (average percent difference) when compared to the use of a single mid-L3 scan. Significant variation in the VAT surface area (p = 0.02) was observed in females compared to males, whereas male patients exhibited a greater variation in SAT surface area (p < 0.001) and radiodensity (p = 0.007).

Conclusion

Significant differences in various body composition measures were observed when comparing non-mid-L3 slices to only the mid-L3 slice. Researchers should be aware that considering only the use of a single midpoint L3 CT scan slice will impact the estimate of body composition measurements.

Automatic segmentation of large-scale CT image datasets for detailed body composition analysis

BACKGROUND

Body composition (BC) is an important factor in determining the risk of type 2-diabetes and cardiovascular disease. Computed tomography (CT) is a useful imaging technique for studying BC, however manual segmentation of CT images is time-consuming and subjective. The purpose of this study is to develop and evaluate fully automated segmentation techniques applicable to a 3-slice CT imaging protocol, consisting of single slices at the level of the liver, abdomen, and thigh, allowing detailed analysis of numerous tissues and organs.

METHODS

The study used more than 4000 CT subjects acquired from the large-scale SCAPIS and IGT cohort to train and evaluate four convolutional neural network based architectures: ResUNET, UNET++, Ghost-UNET, and the proposed Ghost-UNET++. The segmentation techniques were developed and evaluated for automated segmentation of the liver, spleen, skeletal muscle, bone marrow, cortical bone, and various adipose tissue depots, including visceral (VAT), intraperitoneal (IPAT), retroperitoneal (RPAT), subcutaneous (SAT), deep (DSAT), and superficial SAT (SSAT), as well as intermuscular adipose tissue (IMAT). The models were trained and validated for each target using tenfold cross-validation and test sets.

RESULTS

The Dice scores on cross validation in SCAPIS were: ResUNET 0.964 (0.909-0.996), UNET++ 0.981 (0.927-0.996), Ghost-UNET 0.961 (0.904-0.991), and Ghost-UNET++ 0.968 (0.910-0.994). All four models showed relatively strong results, however UNET++ had the best performance overall. Ghost-UNET++ performed competitively compared to UNET++ and showed a more computationally efficient approach.

CONCLUSION

Fully automated segmentation techniques can be successfully applied to a 3-slice CT imaging protocol to analyze multiple tissues and organs related to BC. The overall best performance was achieved by UNET++, against which Ghost-UNET++ showed competitive results based on a more computationally efficient approach. The use of fully automated segmentation methods can reduce analysis time and provide objective results in large-scale studies of BC.

Development and validation of a new equation based on plasma creatinine and muscle mass assessed by CT scan to estimate glomerular filtration rate: a cross-sectional study

Background

Inter-individual variations of non-glomerular filtration rate (GFR) determinants of serum creatinine, such as muscle mass, account for the imperfect performance of estimated GFR (eGFR) equations. We aimed to develop an equation based on creatinine and total lumbar muscle cross-sectional area measured by unenhanced computed tomography scan at the third lumbar vertebra.

Methods

The muscle mass-based eGFR (MMB-eGFR) equation was developed in 118 kidney donor candidates (iohexol clearance) using linear regression. Validation cohorts included 114 healthy subjects from another center (⁵¹Cr-EDTA clearance, validation population 1), 55 patients with chronic diseases (iohexol, validation population 2), and 60 patients with highly discordant creatinine and cystatin C-based eGFR, thus presumed to have atypical non-GFR determinants of creatinine (⁵¹Cr-EDTA, validation population 3). Mean bias was the mean difference between eGFR and measured GFR, precision the standard deviation (SD) of the bias, and accuracy the percentage of eGFR values falling within 20% and 30% of measured GFR.

Results

In validation population 1, performance of MMB-eGFR was not different from those of CKD-EPI_Cr2009 and CKD-EPI_Cr2021. In validation population 2, MMB-eGFR was unbiased and displayed better precision than CKD-EPI_Cr2009, CKD-EPI_Cr2021 and EKFC (SD of the biases: 13.1 vs 16.5, 16.8 and 15.9 mL/min/1.73 m²). In validation population 3, MMB-eGFR had better precision and accuracy {accuracy within 30%: 75.0% [95% confidence interval (CI) 64.0-86.0] vs 51.5% (95% CI 39.0-64.3) for CKD-EPI_Cr2009, 43.3% (95% CI 31.0-55.9) for CKD-EPICr2021, and 53.3% (95% CI 40.7-66.0) for EKFC}. Difference in bias between Black and white subjects was -2.1 mL/min/1.73 m² (95% CI -7.2 to 3.0), vs -8.4 mL/min/1.73 m² (95% CI -13.2 to -3.6) for CKD-EPI_Cr2021.

Conclusion

MMB-eGFR displayed better performances than equations based on demographics, and could be applied to subjects of various ethnic backgrounds.

Abdominal CT Segmentation for Body Composition Assessment Using Network Consistency Learning

Estimating skeletal muscle (SM) and adipose tissues is an invaluable prognostic indicator in cancer treatment, major surgeries, and general health screening. Body composition is usually measured with abdominal computed tomography (CT) scans acquired in clinical settings. The whole-body SM volume is correlated with the estimated SM based on the measurement of a single two-dimensional vertebral slice. It is necessary to label a CT image at the pixel level to estimate SM, known as semantic segmentation. In this work, we trained a segmentation model using the labeled abdominal CT slices and the additional unlabeled slices. In particular, we trained two identical segmentation networks with differently initialized weights. Network Consistency Learning (NCL) allowed learning from unlabeled images by forcing the predictions from both networks to be the same. We segmented abdominal CT images from a newly created in-house dataset. The proposed approach gained 10% better performance in terms of Dice similarity score (DSC) than that obtained by a standard supervised network demonstrating the effectiveness of NCL in exploiting unlabeled images.Clinical relevance- An efficient and cost-effective method is proposed for assessing body composition from limited labeled and abundant unlabeled CT images to facilitate fast diagnosis, prognosis, and interventions.

Artificial Intelligence in Gastric Cancer Imaging With Emphasis on Diagnostic Imaging and Body Morphometry

Gastric cancer remains a significant global health concern, coercing the need for advancements in imaging techniques for ensuring accurate diagnosis and effective treatment planning. Artificial intelligence (AI) has emerged as a potent tool for gastric-cancer imaging, particularly for diagnostic imaging and body morphometry. This review article offers a comprehensive overview of the recent developments and applications of AI in gastric cancer imaging. We investigated the role of AI imaging in gastric cancer diagnosis and staging, showcasing its potential to enhance the accuracy and efficiency of these crucial aspects of patient management. Additionally, we explored the application of AI body morphometry specifically for assessing the clinical impact of gastrectomy. This aspect of AI utilization holds significant promise for understanding postoperative changes and optimizing patient outcomes. Furthermore, we examine the current state of AI techniques for the prognosis of patients with gastric cancer. These prognostic models leverage AI algorithms to predict long-term survival outcomes and assist clinicians in making informed treatment decisions. However, the implementation of AI techniques for gastric cancer imaging has several limitations. As AI continues to evolve, we hope to witness the translation of cutting-edge technologies into routine clinical practice, ultimately improving patient care and outcomes in the fight against gastric cancer.

Segmentation of multi-regional skeletal muscle in abdominal CT image for cirrhotic sarcopenia diagnosis

Background

Sarcopenia is generally diagnosed by the total area of skeletal muscle in the CT axial slice located in the third lumbar (L3) vertebra. However, patients with severe liver cirrhosis cannot accurately obtain the corresponding total skeletal muscle because their abdominal muscles are squeezed, which affects the diagnosis of sarcopenia.

Purpose

This study proposes a novel lumbar skeletal muscle network to automatically segment multi-regional skeletal muscle from CT images, and explores the relationship between cirrhotic sarcopenia and each skeletal muscle region.

Methods

This study utilizes the skeletal muscle characteristics of different spatial regions to improve the 2.5D U-Net enhanced by residual structure. Specifically, a 3D texture attention enhancement block is proposed to tackle the issue of blurred edges with similar intensities and poor segmentation between different skeletal muscle regions, which contains skeletal muscle shape and muscle fibre texture to spatially constrain the integrity of skeletal muscle region and alleviate the difficulty of identifying muscle boundaries in axial slices. Subsequentially, a 3D encoding branch is constructed in conjunction with a 2.5D U-Net, which segments the lumbar skeletal muscle in multiple L3-related axial CT slices into four regions. Furthermore, the diagnostic cut-off values of the L3 skeletal muscle index (L3SMI) are investigated for identifying cirrhotic sarcopenia in four muscle regions segmented from CT images of 98 patients with liver cirrhosis.

Results

Our method is evaluated on 317 CT images using the five-fold cross-validation method. For the four skeletal muscle regions segmented in the images from the independent test set, the avg. DSC is 0.937 and the avg. surface distance is 0.558 mm. For sarcopenia diagnosis in 98 patients with liver cirrhosis, the cut-off values of Rectus Abdominis, Right Psoas, Left Psoas, and Paravertebral are 16.67, 4.14, 3.76, and 13.20 cm²/m² in females, and 22.51, 5.84, 6.10, and 17.28 cm²/m² in males, respectively.

Conclusion

The proposed method can segment four skeletal muscle regions related to the L3 vertebra with high accuracy. Furthermore, the analysis shows that the Rectus Abdominis region can be used to assist in the diagnosis of sarcopenia when the total muscle is not available.

Relationship between body composition and PBRM1 mutations in clear cell renal cell carcinoma: a propensity score matching analysis

OBJECTIVE

This study aimed to examine the relationship between body muscle and adipose tissue composition in clear cell renal cell carcinoma patients with PBRM1 gene mutation.

METHODS

Cancer Genome Atlas Kidney clear cell renal cell carcinoma and Clinical Proteomic Tumor Analysis Consortium clear cell renal cell carcinoma collections were retrieved from the Cancer Imaging Archive. A total of 291 clear cell renal cell carcinoma patients were included in the study retrospectively. Patients' characteristics were obtained from Cancer Imaging Archive. Body composition was assessed with abdominal computed tomography using the automated artificial intelligence software (AID-U™, iAID Inc., Seoul, Korea). Body composition parameters of the patients were calculated. To investigate the net effect of body composition, the propensity score matching procedure was applied over age, gender, and T-stage parameters.

RESULTS

 Of the patients, 184 were males and 107 were females. Mutations in the PBRM1 gene were detected in 77 of the patients. While there was no difference in adipose tissue areas between the PBRM1 mutation group and those without PBRM1 mutation, statistically significant differences were found in normal attenuated muscle area parameters.

CONCLUSION

 This study shows that there was no difference between adipose tissue areas in patients with PBMR1 mutation, but normal attenuated muscle area was found to be higher in PBRM1 patients.

The effects of muscle mass and quality on mortality of patients with acute kidney injury requiring continuous renal replacement therapy

This study examined the effects of muscle mass on mortality in patients with acute kidney injury requiring continuous renal replacement therapy. It was conducted in eight medical centers between 2006 and 2021. The data of 2200 patients over the age of 18 years with acute kidney injury who required continuous renal replacement therapy were retrospectively collected. Skeletal muscle areas, categorized into normal and low attenuation muscle areas, were obtained from computed tomography images at the level of the third lumbar vertebra. Cox proportional hazards models were used to investigate the association between mortality within 1, 3, and 30 days and skeletal muscle index. Sixty percent of patients were male, and the 30-day mortality rate was 52%. Increased skeletal muscle areas/body mass index was associated with decreased mortality risk. We also identified a 26% decreased risk of low attenuation muscle area/body mass index on mortality. We established that muscle mass had protective effects on the mortality of patients with acute kidney injury requiring continuous renal replacement therapy. This study showed that muscle mass is a significant determinant of mortality, even if the density is low.

Serum triglyceride glucose index is a valuable predictor for visceral obesity in patients with type 2 diabetes: a cross-sectional study

BACKGROUND

Since the triglyceride glucose (TyG) index can reflect insulin resistance, it has been proven to be an efficient predictor of glycolipid-metabolism-related diseases. Therefore, this study aimed to investigate the predictive value of the TyG index for visceral obesity (VO) and body fat distribution in patients with type 2 diabetes mellitus (T2DM).

METHODS

Abdominal adipose tissue characteristics in patients with T2DM, including visceral adipose area (VAA), subcutaneous adipose area (SAA), VAA-to-SAA ratio (VSR), visceral adipose density (VAD), and subcutaneous adipose density (SAD), were obtained through analyses of computed tomography images at the lumbar 2/3 level. VO was diagnosed according to the VAA (> 142 cm² for males and > 115 cm² for females). Logistic regression was performed to identify independent factors of VO, and receiver operating characteristic (ROC) curves were used to compare the diagnostic performance according to the area under the ROC curve (AUC).

RESULTS

A total of 976 patients were included in this study. VO patients showed significantly higher TyG values than non-VO patients in males (9.74 vs. 8.88) and females (9.59 vs. 9.01). The TyG index showed significant positive correlations with VAA, SAA, and VSR and negative correlations with VAD and SAD. The TyG index was an independent factor for VO in both males (odds ratio [OR] = 2.997) and females (OR = 2.233). The TyG index ranked second to body mass index (BMI) for predicting VO in male (AUC = 0.770) and female patients (AUC = 0.720). Patients with higher BMI and TyG index values showed a significantly higher risk of VO than the other patients. TyG-BMI, the combination index of TyG and BMI, showed significantly higher predictive power than BMI for VO in male patients (AUC = 0.879 and 0.835, respectively) but showed no significance when compared with BMI in female patients (AUC = 0.865 and 0.835, respectively).

CONCLUSIONS

. TyG is a comprehensive indicator of adipose volume, density, and distribution in patients with T2DM and is a valuable predictor for VO in combination with anthropometric indices, such as BMI.

Can automated CT body composition analysis predict high-grade Clavien-Dindo complications in patients with RCC undergoing partial and radical nephrectomy?

INTRODUCTION

This study investigated the relationship between body tissue composition analysis and complications according to the Clavien-Dindo classification in patients with renal cell carcinoma (RCC) who underwent partial (PN) or radical nephrectomies (RN).

METHODS

We obtained all data of 210 patients with RCC from the 2019 Kidney and Kidney Tumor Segmentation Challenge (C4KC-KiTS) dataset and obtained radiological images from the cancer image archive. Body composition was assessed with automated artificial intelligence software using the convolutional network segmentation technique from abdominal computed tomography images. We included 125 PN and 63 RN in the study. The relationship between body fat and muscle tissue distribution and complications according to the Clavien-Dindo classification was evaluated between these two groups.

RESULTS

Clavien-Dindo 3A and higher (high grade) complications were developed in 9 of 125 patients who underwent PN and 7 of 63 patients who underwent RN. There was no significant difference between all body composition values between patients with and without high-grade complications.

CONCLUSION

This study showed that body muscle-fat tissue distribution did not affect patients with 3A and above complications according to the Clavien-Dindo classification in patients who underwent nephrectomy due to RCC.

A deep learning model based on the attention mechanism for automatic segmentation of abdominal muscle and fat for body composition assessment

Background

Quantitative muscle and fat data obtained through body composition analysis are expected to be a new stable biomarker for the early and accurate prediction of treatment-related toxicity, treatment response, and prognosis in patients with lung cancer. The use of these biomarkers can enable the adjustment of individualized treatment regimens in a timely manner, which is critical to further improving patient prognosis and quality of life. We aimed to develop a deep learning model based on attention for fully automated segmentation of the abdomen from computed tomography (CT) to quantify body composition.

Methods

A fully automatic segmentation deep learning model was designed based on the attention mechanism and using U-Net as the framework. Subcutaneous fat, skeletal muscle, and visceral fat were manually segmented by two experts to serve as ground truth labels. The performance of the model was evaluated using Dice similarity coefficients (DSCs) and Hausdorff distance at 95th percentile (HD95).

Results

The mean DSC for subcutaneous fat and skeletal muscle were high for both the enhanced CT test set (0.93±0.06 and 0.96±0.02, respectively) and the plain CT test set (0.90±0.09 and 0.95±0.01, respectively). Nevertheless, the model did not perform well in the segmentation performance of visceral fat, especially for the enhanced CT test set. The mean DSC for the enhanced CT test set was 0.87±0.11, while the mean DSC for the plain CT test set was 0.92±0.03. We discuss the reasons for this result.

Conclusions

This work demonstrates a method for the automatic outlining of subcutaneous fat, skeletal muscle, and visceral fat areas at L3.

Association of sex-specific donor skeletal muscle index with surgical outcomes in living donor liver transplantation recipients

BACKGROUND

A recent study reported a correlation between the muscle mass of male donors and graft failure in living donor liver transplantation (LDLT) recipients. We investigated the association of sex-specific donor skeletal muscle index (SMI) with mortality and graft failure in LDLT recipients.

METHODS

We retrospectively analysed 2750 sets of donors and recipients between January 2008 and January 2018. The recipient outcomes were analysed by dividing the data according to donor sex. Cox regression analyses were performed to evaluate the association between donor SMI by sex and 1-year mortality and graft failure in recipients.

RESULTS

In the male donor group, robust donor (increased SMI) was significantly associated with higher risks for mortality (hazard ratio [HR]: 1.03, 95% confidence interval [CI]: 1.00-1.06, p = .023) and graft failure (HR: 1.04, 95% CI: 1.01-1.06, p = .007) at 1 year. In the female donor group, the robust donor was significantly associated with lower risks for mortality (HR: 0.92, 95% CI: 0.87-0.97, p = .003) and graft failure (HR: 0.95, 95% CI: 0.90-1.00, p = .032) at 1 year.

CONCLUSIONS

Donor SMI was associated with surgical outcomes in recipients. Robust male and female donors were a significant negative and protective factor for grafts respectively.

Impact of Sarcopenia on Clinical Course of Inflammatory Bowel Disease in Korea

BACKGROUND AND AIMS

Reduced body muscle mass is a poor prognostic factor for inflammatory bowel disease (IBD). In this study, we investigated the prevalence of sarcopenia at diagnosis and its clinical significance in Korean patients with IBD.

METHODS

The prevalence of sarcopenia in IBD patients between June 1989 and December 2016 was investigated using a well-characterized referral center-based cohort. Abdominopelvic computed tomography within six months from IBD diagnosis was used for the evaluation. Sarcopenia was defined as an L3 skeletal muscle index of < 49 cm²/m² for male and < 31 cm²/m² for female. The clinical characteristics and outcomes were evaluated with respect to sarcopenia.

RESULTS

A total of 1,027 patients (854 Crohn's disease [CD]; 173 ulcerative colitis [UC]) were evaluated. Sarcopenia was found in 56.8% of the population (CD, 57.5%; UC, 53.2%), and male were more likely to be sarcopenic (CD, 94.3%; UC, 91.6%). There were no significant differences in the cumulative risk of using steroids, immunomodulators, biologics, and bowel resections (or colectomy) with or without sarcopenia during follow-up (median: CD, 5.8 years; UC, 3.7 years). In sarcopenic patients with CD, there was a significantly higher cumulative risk of perianal surgeries than in non-sarcopenic patients with CD (Log-rank test; P = 0.001). However, the risk of perianal surgeries was not significant in multivariate analysis (Odds ratio 1.368; 95% confidence interval 0.782-2.391; P = 0.272).

CONCLUSION

Sarcopenia at diagnosis may have no significant prognostic value for medical treatment and bowel resection, but it may be associated with perianal CD.

Deep Learning-Based Medical Images Segmentation of Musculoskeletal Anatomical Structures: A Survey of Bottlenecks and Strategies

By leveraging the recent development of artificial intelligence algorithms, several medical sectors have benefited from using automatic segmentation tools from bioimaging to segment anatomical structures. Segmentation of the musculoskeletal system is key for studying alterations in anatomical tissue and supporting medical interventions. The clinical use of such tools requires an understanding of the proper method for interpreting data and evaluating their performance. The current systematic review aims to present the common bottlenecks for musculoskeletal structures analysis (e.g., small sample size, data inhomogeneity) and the related strategies utilized by different authors. A search was performed using the PUBMED database with the following keywords: deep learning, musculoskeletal system, segmentation. A total of 140 articles published up until February 2022 were obtained and analyzed according to the PRISMA framework in terms of anatomical structures, bioimaging techniques, pre/post-processing operations, training/validation/testing subset creation, network architecture, loss functions, performance indicators and so on. Several common trends emerged from this survey; however, the different methods need to be compared and discussed based on each specific case study (anatomical region, medical imaging acquisition setting, study population, etc.). These findings can be used to guide clinicians (as end users) to better understand the potential benefits and limitations of these tools.

Association between hypertension and myosteatosis evaluated by abdominal computed tomography

Few studies have examined the relationship between myosteatosis and hypertension, and no studies have enrolled an Asian population. Existing studies also found discordant results, possibly due to the use of conventional myosteatosis indices that are not sufficiently reliable and representative. Therefore, we investigated the association between myosteatosis and hypertension in Asian individuals using novel, objective computed tomography (CT) markers. The total abdominal muscle area (TAMA) was determined from abdominal CT scans taken at the L3 level. Based on the mean CT attenuation, the TAMA was divided into intramuscular adipose tissue and skeletal muscle area (SMA), which was further segmented into normal attenuation muscle area (NAMA) and low attenuation muscle area (LAMA). Among SMA/body mass index (BMI), NAMA/BMI, LAMA/BMI, and the NAMA/TAMA index, NAMA/BMI was chosen through receiver operating characteristic curves as the best predictive marker for hypertension. The hypertension risk for each quartile of NAMA/BMI was calculated by logistic regression analysis. Among the 19,766 participants, 40.3% of men and 23.8% of women had hypertension. People with hypertension showed unhealthier myosteatosis profiles than normotensive controls. Similarly, a lower NAMA/BMI was significantly associated with a greater hypertension risk. The lowest quartile group of NAMA/BMI exhibited 2.3- and 2.6-fold higher risks of hypertension than the highest quartile in men and women, respectively. In conclusion, advanced myosteatosis assessed by abdominal CT was significantly correlated with a higher risk of hypertension. Improving myosteatosis may be a new approach for preventing cardiovascular diseases, including hypertension. Advanced myosteatosis measured by abdominal CT taken at the L3 level was significantly correlated with a higher risk of hypertension even after adjusting for health behaviors, intake of lipid-lowering drugs, plasma lipid levels, and other ectopic fat distribution.

Impact of the preoperative skeletal muscle index on early remnant liver regeneration in living donors after liver transplantation

Background

We investigated the correlation between the preoperative skeletal muscle index (SMI) and remnant liver regeneration after right hemihepatectomy for living-donor liver transplantation and aimed to identify preoperative predictors of greater early remnant liver regeneration in living donors.

Methods

This retrospective study included 525 right hemiliver donors (mean age, 28.9±8.3 years; 345 male patients) between 2017 and 2018, who underwent computed tomography before surgery and on postoperative day (POD) 7. Preoperative anthropometry, laboratory parameters, skeletal muscle area at the third lumbar vertebral level, and liver volume before and after surgery were evaluated. Correlations were analyzed using Pearson correlation coefficients, and stepwise multiple regression analysis was performed to identify independent predictors of greater remnant liver regeneration.

Results

Remnant liver regeneration volume on POD 7 was positively correlated with body mass index (BMI; r=0.280, P<0.001) and SMI (r=0.322, P<0.001), and negatively correlated with age (r=-0.154, P<0.001) and the ratio of future remnant liver volume (FRLV) to total liver volume (TLV; r=-0.261, P<0.001). Stepwise multiple regression analysis showed that high BMI (β=0.146; P=0.001) and SMI (β=0.228, P<0.001), young age (β=-0.091, P=0.025), and a low FRLV/TLV ratio (β=-0.225, P<0.001) were predictors of greater remnant liver regeneration.

Conclusions

High SMI and BMI, young age, and a low FRLV/TLV ratio may predict greater early remnant liver regeneration in living donors after LDLT.

Testosterone is associated with abdominal body composition derived from computed tomography: a large cross sectional study

The aim of this study was to evaluate the association between serum testosterone and abdominal body composition based on abdominopelvic computed tomography (APCT) measurements after adjusting for individual metabolic syndrome components. We performed a cross-sectional study using male subjects (age range: 22-84 years) who underwent a general health examination with abdominopelvic computed tomography and testosterone measurements. Body composition was evaluated with APCT. To confirm an association between testosterone and abdominal body composition, we conducted linear regression analysis. The effect of abdominal body composition was adjusted for important clinical factors such as age, albumin, and metabolic components in the multivariable regression analysis. Overall, 1453 subjects were included in the primary analysis. After adjustment for age, individual metabolic components, albumin, hemoglobin A1c, and C-reactive protein, we found that subcutaneous fat area index (β = - 0.042, p < 0.001), total abdominal muscle area index (β = 0.115, p < 0.001), normal attenuation muscle area index (β = 0.070, p < 0.001), and log_e-transformed lower attenuation muscle area index (β = 0.140, p = 0.002) had an association with log_e-transformed testosterone level. After adjusting for individual metabolic syndrome components, testosterone was associated negatively with subcutaneous fat, but not visceral fat. In addition, testosterone was positively correlated with abdominal muscle regardless of qualitative features such as fat-rich and fat-free.

Serial Changes in Body Composition and the Association with Disease Activity during Treatment in Patients with Crohn's Disease

Objectives: To analyze serial changes in body composition and investigate the association between body composition changes and disease activity changes in patients with Crohn’s disease (CD). Methods: Seventy-one patients with CD who had been treated and followed-up at our institution were included. Two to four computed tomography images were acquired at baseline, and the 2−5-year, 5−8-year, and last follow-ups were selected per patient for body composition and disease activity analyses. Visceral fat area (VFA), skeletal muscle index (SMI; skeletal muscle area/height2), and subcutaneous fat area (SFA) were assessed using an artificial-intelligence-driven fully automated method. Disease activity was assessed using a modified computed tomography scoring system and the Simple Endoscopic Score for Crohn’s Disease. The associations between body composition, disease activity, and remission were investigated. Results: The mean age was 29.83 ± 11.27 years; most patients were men (48/71, 67.6%); and the median follow-up was 144 (12−264) months. Overall, VFA and SFA gradually increased, while SMI decreased during the follow-up. Sarcopenia was associated with the female sex, higher disease activities at baseline (p = 0.01) and the last follow-up (p = 0.001). SMI and SFA inversely correlated with the disease activity, i.e., the more severe the disease activity, the lower the SMI and SFA (p < 0.05). SMI at the last follow-up was the only significant predictor of remission (OR = 1.21, 95% confidence interval: 1.03−1.42, p = 0.021). Conclusion: SMI decreased while VFA and SFA increased during the treatment follow-up in patients with CD. Sarcopenia was associated with higher disease activity, and SMI and SFA inversely correlated with disease activity. SMI at the last follow-up was the significant factor for remission.

Prognostic Value of Sarcopenia and Myosteatosis in Patients with Resectable Pancreatic Ductal Adenocarcinoma

OBJECTIVE

The clinical relevance of myosteatosis has not been well evaluated in patients with pancreatic ductal adenocarcinoma (PDAC), although sarcopenia has been extensively researched. Therefore, we evaluated the prognostic value of muscle quality, including myosteatosis, in patients with resectable PDAC treated surgically.

MATERIALS AND METHODS

We retrospectively evaluated 347 patients with resectable PDAC who underwent curative surgery (mean age ± standard deviation, 63.6 ± 9.6 years; 202 male). Automatic muscle segmentation was performed on preoperative computed tomography (CT) images using an artificial intelligence program. A single axial image of the portal phase at the inferior endplate level of the L3 vertebra was used for analysis in each patient. Sarcopenia was evaluated using the skeletal muscle index, calculated as the skeletal muscle area (SMA) divided by the height squared. The mean SMA attenuation was used to evaluate myosteatosis. Diagnostic cutoff values for sarcopenia and myosteatosis were devised using the Contal and O'Quigley methods, and patients were classified according to normal (nMT), sarcopenic (sMT), myosteatotic (mMT), or combined (cMT) muscle quality types. Multivariable Cox regression analyses were conducted to assess the effects of muscle type on the overall survival (OS) and recurrence-free survival (RFS) after surgery.

RESULTS

Eighty-four (24.2%), 73 (21.0%), 75 (21.6%), and 115 (33.1%) patients were classified as having nMT, sMT, mMT, and cMT, respectively. Compared to nMT, mMT and cMT were significantly associated with poorer OS, with hazard ratios (HRs) of 1.49 (95% confidence interval, 1.00-2.22) and 1.68 (1.16-2.43), respectively, while sMT was not (HR of 1.40 [0.94-2.10]). Only mMT was significantly associated with poorer RFS, with an HR of 1.59 (1.07-2.35), while sMT and cMT were not.

CONCLUSION

Myosteatosis was associated with poor OS and RFS in patients with resectable PDAC who underwent curative surgery.

A new finding for the obesity paradox? Evaluation of the relationship between muscle and adipose tissue in nuclear grade prediction in patients with clear cell renal cell carcinoma

Background

Obesity is associated with an increased risk of developing clear cell renal cell carcinoma (ccRCC), but paradoxically there is a positive association between obesity and surveillance.

Purpose

To investigate the relationship between nucleus grade classification and body composition in patients with matched co-morbid conditions with non-metastatic ccRCC.

Materials and Methods

A total of 253 patients with non-metastatic ccRCC were included in the study. Body composition was assessed with abdominal computed tomography (CT) using an automated artificial intelligence software. Both adipose and muscle tissue parameters of the patients were calculated. In order to investigate the net effect of body composition, propensity score matching (PSM) procedure was applied over age, sex, and T stage parameters. In this way, selection bias and imbalance between groups were minimized. Univariate and multivariate logistic regression analyses were performed to identify the association between body composition and WHO/ISUP grade (I–IV).

Result

When the body composition of the patients was examined without matching the conditions, it was found that the subcutaneous adipose tissue (SAT) values were higher in patients with low grades ( P = 0.001). Normal attenuation muscle area (NAMA) was higher in high-grade patients than low-grade patients ( P < 0.05). In the post-matching evaluation, only SAT/NAMA was found to be associated with high-grade ccRCC (univariate analysis: odds ratio [OR]=0.899, 95% confidence interval [CI]=0.817−0.988, P = 0.028; multivariate analysis: OR=0.922, 95% CI=0.901−0.974, P = 0.042).

Conclusion

CT-based body composition parameters can be used as a prognostic marker in predicting nuclear grade when age, sex, and T stage match conditions. This finding offers a new perspective on the obesity paradox.

Sex-specific Cutoff Values of Visceral Fat Area for Lean vs. Overweight/Obese Nonalcoholic Fatty Liver Disease in Asians

BACKGROUND AND AIMS

Visceral obesity is a risk factor for nonalcoholic fatty liver disease (NAFLD). We investigated sex-specific optimal cutoff values for visceral fat area (VFA) associated with lean and overweight/obese NAFLD in an Asian population.

METHODS

This retrospective study included 678 potential living liver donors (mean age, 30.8±9.4 years; 434 men and 244 women) who had undergone abdominal computed tomography (CT) imaging and liver biopsy between November 2016 and October 2017. VFA was measured using single-slice abdominal CT. NAFLD was evaluated by liver biopsy (≥5% hepatic steatosis). Receiver operating characteristic curve analysis was used to determine cutoff values for VFA associated with lean (body mass index [BMI] <23 kg/m²) and overweight/obese (BMI ≥23 kg/m²) NAFLD.

RESULTS

Area under the curve (AUC) values with 95% confidence intervals (CI) for VFA were 0.82 (95% CI, 0.75-0.88) for lean and 0.74 (95% CI, 0.69-0.79) for overweight/obese men with NAFLD. The AUC values were 0.67 (95% CI, 0.58-0.75) for lean and 0.71 (95% CI, 0.62-0.80) for overweight/obese women with NAFLD. The cutoff values for VFA associated with lean NAFLD were 50.2 cm² in men and 40.5 cm² in women. The optimal cutoff values for VFA associated with overweight/obese NAFLD were 100.6 cm² in men and 68.0 cm² in women.

CONCLUSIONS

Sex-specific cutoff values for VFA may be useful for identifying subjects at risk of lean and overweight/obese NAFLD.

Associations of sarcopenia with graft failure and mortality in patients undergoing living donor liver transplantation

Recent studies have reported that sarcopenia influences morbidity and mortality in surgical patients. However, few studies have investigated the associations of sarcopenia with short-term and long-term graft failure in recipients after living donor liver transplantation (LDLT). In this study, we investigated the associations between sarcopenia and graft failure/mortality in patients undergoing LDLT. We retrospectively examined 2816 recipients who underwent LDLT between January 2008 and January 2018. Cox regression analysis was performed to evaluate the associations between sarcopenia and graft failure/mortality in recipients at 60 days, 180 days, and 1 year and overall. Sarcopenia in the recipient was significantly associated with 60-day graft failure (adjusted hazard ratio [HR], 1.98; 95% confidence interval [CI], 1.09-3.61; p = 0.03), 180-day graft failure (HR, 1.85; 95% CI, 1.19-2.88; p = 0.01), 1-year graft failure (HR, 1.45; 95% CI, 1.01-2.17; p = 0.05), and overall graft failure (HR, 1.42; 95% CI, 1.08-1.87; p = 0.01). In addition, recipient sarcopenia was associated with 180-day mortality (HR, 1.88; 95% CI, 1.17-3.01; p = 0.01), 1-year mortality (HR, 1.53; 95% CI, 1.01-2.29; p = 0.04), and overall mortality (HR, 1.43; 95% CI, 1.08-1.90; p = 0.01). Preoperative sarcopenia was associated with high rates of graft failure and mortality in LDLT recipients. Therefore, preoperative sarcopenia may be a strong predictor of the surgical prognosis in LDLT recipients.

Development and validation of an ensemble artificial intelligence model for comprehensive imaging quality check to classify body parts and contrast enhancement

Background

Despite the dramatic increase in the use of medical imaging in various therapeutic fields of clinical trials, the first step of image quality check (image QC), which aims to check whether images are uploaded appropriately according to the predefined rules, is still performed manually by image analysts, which requires a lot of manpower and time.

Methods

In this retrospective study, 1669 computed tomography (CT) images with five specific anatomical locations were collected from Asan Medical Center and Kangdong Sacred Heart Hospital. To generate the ground truth, two radiologists reviewed the anatomical locations and presence of contrast enhancement using the collected data. The individual deep learning model is developed through InceptionResNetv2 and transfer learning, and we propose Image Quality Check-Net (Image QC-Net), an ensemble AI model that utilizes it. To evaluate their clinical effectiveness, the overall accuracy and time spent on image quality check of a conventional model and ImageQC-net were compared.

Results

ImageQC-net body part classification showed excellent performance in both internal (precision, 100%; recall, 100% accuracy, 100%) and external verification sets (precision, 99.8%; recovery rate, 99.8%, accuracy, 99.8%). In addition, contrast enhancement classification performance achieved 100% precision, recall, and accuracy in the internal verification set and achieved (precision, 100%; recall, 100%; accuracy 100%) in the external dataset. In the case of clinical effects, the reduction of time by checking the quality of artificial intelligence (AI) support by analysts 1 and 2 (49.7% and 48.3%, respectively) was statistically significant (p < 0.001).

Conclusions

Comprehensive AI techniques to identify body parts and contrast enhancement on CT images are highly accurate and can significantly reduce the time spent on image quality checks.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12880-022-00815-4.

Sex Differences of Visceral Fat Area and Visceral-to-Subcutaneous Fat Ratio for the Risk of Incident Type 2 Diabetes Mellitus

BACKGROUND

This study aimed to determine the optimal cut-off values of visceral fat area (VFA) and visceral-to-subcutaneous fat ratio (VSR) for predicting incident type 2 diabetes mellitus (T2DM).

METHODS

A total of 10,882 individuals (6,835 men; 4,047 women) free of T2DM at baseline aged between 30 and 79 years who underwent abdominal computed tomography scan between 2012 and 2013 as a part of routine health check-ups were included and followed. VFA, subcutaneous fat area, and VSR on L3 vertebral level were measured at baseline.

RESULTS

During a median follow-up of 4.8 years, 730 (8.1% for men; 4.3% for women) incident cases of T2DM were identified. Receiver operating characteristic curve analysis showed that the optimal cut-off values of VFA and VSR for predicting incident T2DM were 130.03 cm2 and 1.08 in men, respectively, and 85.7 cm2 and 0.48 in women, respectively. Regardless of sex, higher VFA and VSR were significantly associated with a higher risk of incident T2DM. Compared with the lowest quartiles of VFA and VSR, the highest quartiles had adjusted odds ratios of 2.62 (95% confidence interval [CI], 1.73 to 3.97) and 1.55 (95% CI, 1.14 to 2.11) in men, respectively, and 32.49 (95% CI, 7.42 to 142.02) and 11.07 (95% CI, 3.89 to 31.50) in women, respectively.

CONCLUSION

Higher VFA and VSR at baseline were independent risk factors for the development of T2DM. Sex-specific reference values for visceral fat obesity (VFA ≥130 cm2 or VSR ≥1.0 in men; VFA ≥85 cm2 or VSR ≥0.5 in women) are proposed for the prediction of incident T2DM.

Association between type 2 diabetes and skeletal muscle quality assessed by abdominal computed tomography scan

AIM

To examine the association between type 2 diabetes and the amount and quality of trunk muscle as assessed by computed tomography (CT) scan.

MATERIALS AND METHODS

A total of 20,986 subjects (13,007 men and 7979 women) who underwent abdominal CT scan as part of a routine health check-up were included. The total abdominal muscle area (TAMA) measured at the third lumbar vertebrae was classified into skeletal muscle area (SMA), and intermuscular adipose tissue area. SMA was divided into good quality muscles (normal attenuation muscle area [NAMA]) and poor quality muscles (low attenuation muscle area). NAMA/TAMA index was calculated.

RESULTS

Subjects with type 2 diabetes had higher values of TAMA and SMA but significantly lower values of NAMA and NAMA/TAMA index. Compared with those in the lowest quartile of NAMA/TAMA index, subjects in the highest quartile had metabolically favourable laboratory findings, a lower prevalence of type 2 diabetes (Q1 vs. Q4: 19.3% vs. 9.5% in men, 12.3% vs. 3.0% in women) and inverse association with type 2 diabetes (odds ratio for Q2, Q3, and Q4: 0.87, 0.78, and 0.75 in men; 0.82, 0.70, and 0.68 in women) after multivariable adjustment.

CONCLUSIONS

The amount of good quality muscle on CT scan was associated with a lower prevalence of type 2 diabetes.

Automated segmentation of the fractured vertebrae on CT and its applicability in a radiomics model to predict fracture malignancy

Although CT radiomics has shown promising results in the evaluation of vertebral fractures, the need for manual segmentation of fractured vertebrae limited the routine clinical implementation of radiomics. Therefore, automated segmentation of fractured vertebrae is needed for successful clinical use of radiomics. In this study, we aimed to develop and validate an automated algorithm for segmentation of fractured vertebral bodies on CT, and to evaluate the applicability of the algorithm in a radiomics prediction model to differentiate benign and malignant fractures. A convolutional neural network was trained to perform automated segmentation of fractured vertebral bodies using 341 vertebrae with benign or malignant fractures from 158 patients, and was validated on independent test sets (internal test, 86 vertebrae [59 patients]; external test, 102 vertebrae [59 patients]). Then, a radiomics model predicting fracture malignancy on CT was constructed, and the prediction performance was compared between automated and human expert segmentations. The algorithm achieved good agreement with human expert segmentation at testing (Dice similarity coefficient, 0.93-0.94; cross-sectional area error, 2.66-2.97%; average surface distance, 0.40-0.54 mm). The radiomics model demonstrated good performance in the training set (AUC, 0.93). In the test sets, automated and human expert segmentations showed comparable prediction performances (AUC, internal test, 0.80 vs 0.87, p = 0.044; external test, 0.83 vs 0.80, p = 0.37). In summary, we developed and validated an automated segmentation algorithm that showed comparable performance to human expert segmentation in a CT radiomics model to predict fracture malignancy, which may enable more practical clinical utilization of radiomics.

Change of Computed Tomography-Based Body Composition after Adrenalectomy in Patients with Pheochromocytoma

Despite the potential biological importance of the sympathetic nervous system on fat and skeletal muscle metabolism in animal and in vitro studies, its relevance in humans remains undetermined. To clarify the influence of catecholamine excess on human body composition, we performed a retrospective longitudinal cohort study including 313 consecutive patients with histologically confirmed pheochromocytoma who underwent repeat abdominal computed tomography (CT) scans before and after adrenalectomy. Changes in CT-determined visceral fat area (VFA), subcutaneous fat area (SFA), skeletal muscle area (SMA), and skeletal muscle index (SMI) were measured at the level of the third lumbar vertebra. The mean age of all patients was 50.6 ± 13.6 years, and 171/313 (54.6%) were women. The median follow-up duration for repeat CTs was 25.0 months. VFA and SFA were 14.5% and 15.8% higher, respectively (both p < 0.001), after adrenalectomy, whereas SMA and SMI remained unchanged. Similarly, patients with visceral obesity significantly increased from 103 (32.9%) at baseline to 138 (44.1%) following surgery (p < 0.001); however, the prevalence of sarcopenia was unchanged. This study provides important clinical evidence that sympathetic hyperactivity can contribute to lipolysis in visceral and subcutaneous adipose tissues, but its impact on human skeletal muscle is unclear.

Artificial Intelligence Mortality Prediction Model for Gastric Cancer Surgery Based on Body Morphometry, Nutritional, and Surgical Information: Feasibility Study

The objective of this study is to develop a mortality prediction model for patients undergoing gastric cancer surgery based on body morphometry, nutritional, and surgical information. Using a prospectively built gastric surgery registry from the Asan Medical Center (AMC), 621 gastric cancer patients, who were treated with surgery with no recurrence of cancer, were selected for the development of the prediction model. Input features (i.e., body morphometry, nutritional, surgical, and clinicopathologic information) were selected in the collected data based on the XGBoost analysis results and experts’ opinions. A convolutional neural network (CNN) framework was developed to predict the mortality of patients undergoing gastric cancer surgery. Internal validation was performed in split datasets of the AMC, whereas external validation was performed in patients in the Ajou University Hospital. Fifteen features were selected for the prediction of survival probability based on the XGBoost analysis results and experts’ suggestions. Accuracy, F1 score, and area under the curve of our CNN model were 0.900, 0.909, and 0.900 in the internal validation set and 0.879, 0.882, and 0.881 in the external validation set, respectively. Our developed CNN model was published on a website where anyone could predict mortality using individual patients’ data. Our CNN model provides substantially good performance in predicting mortality in patients undergoing surgery for gastric cancer, mainly based on body morphometry, nutritional, and surgical information. Using the web application, clinicians and gastric cancer patients will be able to efficiently manage mortality risk factors.

Relationship of body mass index and abdominal fat with radiation dose received during preoperative liver CT in potential living liver donors: a cross-sectional study

BACKGROUND

Although contrast-enhanced computed tomography (CT) is currently the most widely-used imaging modality for the preoperative evaluation of potential living liver donors, radiation exposure remains a major concern. The present study aimed to determine the relationship of body mass index (BMI) and abdominal fat with the effective radiation dose received during liver CT scans as part of a pre-donation work-up in potential living donors.

METHODS

This retrospective cross-sectional study included 695 potential living donors (mean age, 30.5±9.7 years; 445 men and 250 women) who had undergone preoperative liver CT scans between 2017 and 2018. The following measures were evaluated: BMI, abdominal fat as measured at the level of the third lumbar vertebra, and effective dose based on the dose length product (DLP). Correlations between the effective dose and other variables were evaluated using Pearson's correlation coefficient.

RESULTS

The mean BMI, total fat area (TFA), and effective dose were 23.6±3.3 kg/m2, 218.7±110.0 cm2, and 9.4±3.3 mSv, respectively. The effective dose during liver CT scans had a strong positive correlation with both BMI (r=0.715; P<0.001) and TFA (r=0.792; P<0.001). As BMI and TFA increased, so did the effective dose.

CONCLUSIONS

Higher BMI and TFA significantly increased the radiation dose received during liver CT scans in potential living donors.

The Value of Artificial Intelligence-Assisted Imaging in Identifying Diagnostic Markers of Sarcopenia in Patients with Cancer

Sarcopenia is defined as the loss of skeletal muscle mass and muscle function. It is common in patients with malignancies and often associated with adverse clinical outcomes. The presence of sarcopenia in patients with cancer is determined by body composition, and recently, radiologic technology for the accurate estimation of body composition is under development. Artificial intelligence- (AI-) assisted image measurement facilitates the detection of sarcopenia in clinical practice. Sarcopenia is a prognostic factor for patients with cancer, and confirming its presence helps to recognize those patients at the greatest risk, which provides a guide for designing individualized cancer treatments. In this review, we examine the recent literature (2017-2021) on AI-assisted image assessment of body composition and sarcopenia, seeking to synthesize current information on the mechanism and the importance of sarcopenia, its diagnostic image markers, and the interventions for sarcopenia in the medical care of patients with cancer. We concluded that AI-assisted image analysis is a reliable automatic technique for segmentation of abdominal adipose tissue. It has the potential to improve diagnosis of sarcopenia and facilitates identification of oncology patients at the greatest risk, supporting individualized prevention planning and treatment evaluation. The capability of AI approaches in analyzing series of big data and extracting features beyond manual skills would no doubt progressively provide impactful information and greatly refine the standard for assessing sarcopenia risk in patients with cancer.

Validation of a deep learning segmentation algorithm to quantify the skeletal muscle index and sarcopenia in metastatic renal carcinoma

OBJECTIVES

To validate a deep learning (DL) algorithm for measurement of skeletal muscular index (SMI) and prediction of overall survival in oncology populations.

METHODS

A retrospective single-center observational study included patients with metastatic renal cell carcinoma between 2007 and 2019. A set of 37 patients was used for technical validation of the algorithm, comparing manual vs DL-based evaluations. Segmentations were compared using mean Dice similarity coefficient (DSC), SMI using concordance correlation coefficient (CCC) and Bland-Altman plots. Overall survivals (OS) were compared using log-rank (Kaplan-Meier) and Mann-Whitney tests. Generalizability of the prognostic value was tested in an independent validation population (N = 87).

RESULTS

Differences between two manual segmentations (DSC = 0.91, CCC = 0.98 for areas) or manual vs. automated segmentation (DSC = 0.90, CCC = 0.98 for areas, CCC = 0.97 for SMI) had the same order of magnitude. Bland-Altman plots showed a mean difference of -3.33 cm² [95%CI: -15.98, 9.1] between two manual segmentations, and -3.28 cm² [95% CI: -14.77, 8.21] for manual vs. automated segmentations. With each method, 20/37 (56%) patients were classified as sarcopenic. Sarcopenic vs. non-sarcopenic groups had statistically different survival curves with median OS of 6.0 vs. 12.5 (p = 0.008) and 6.0 vs. 13.9 (p = 0.014) months respectively for manual and DL methods. In the independent validation population, sarcopenic patients according to DL had a lower OS (10.7 vs. 17.3 months, p = 0.033).

CONCLUSION

A DL algorithm allowed accurate estimation of SMI compared to manual reference standard. The DL-calculated SMI demonstrated a prognostic value in terms of OS.

KEY POINTS

• A deep learning algorithm allows accurate estimation of skeletal muscle index compared to a manual reference standard with a concordance correlation coefficient of 0.97. • Sarcopenic patients according to SMI thresholds after segmentation by the deep learning algorithm had statistically significantly lower overall survival compared to non-sarcopenic patients.

Visceral Fat Area Is an Independent Risk Factor for Overweight or Obese Nonalcoholic Fatty Liver Disease in Potential Living Liver Donors

BACKGROUND

The present study aimed to evaluate the correlation between hepatic steatosis (HS), determined by biopsy, and visceral adiposity, measured by computed tomography (CT), in overweight or obese potential living liver donors, and to investigate the risk factors for overweight or obese nonalcoholic fatty liver disease (NAFLD).

METHODS

This retrospective study included 375 overweight or obese (body mass index ≥23 kg/m²) potential living liver donors (mean age, 30.4 ± 9.5 years; 273 men) who underwent liver biopsies and abdominal CT examinations in 2017 and 2018. Anthropometry, laboratory parameters, body composition, and HS were assessed. Correlations were analyzed using Pearson's correlation coefficient, and logistic regression was used to identify independent predictors of overweight or obese NAFLD.

RESULTS

Visceral fat area (VFA) was positively correlated with the degree of HS in men (r = 0.307; P < .001) and women (r = 0.387; P < .001). Multivariable logistic regression analysis showed that alanine aminotransferase (odds ratio [OR], 1.017; 95% confidence interval [CI], 1.001-1.033; P = .033) and VFA (OR, 1.015; 95% CI, 1.008-1.022; P < .001) were independent risk factors for overweight or obese NAFLD in men, and VFA (OR, 1.029; 95% CI, 1.011-1.047; P = .002) was an independent risk factor for overweight or obese NAFLD in women.

CONCLUSION

Visceral adiposity was positively correlated with the degree of HS in overweight or obese potential living liver donors. Additionally, visceral adiposity may be an independent risk factor for overweight or obese NAFLD in potential living liver donors.

Transfer learning for data-efficient abdominal muscle segmentation with convolutional neural networks

Background

Skeletal muscle segmentation is an important procedure for assessing sarcopenia, an emerging imaging biomarker of patient frailty. Data annotation remains the bottleneck for training deep learning auto‐segmentation models.

Purpose

There is a need to define methodologies for applying models to different domains (e.g., anatomical regions or imaging modalities) without dramatically increasing data annotation.

Methods

To address this problem, we empirically evaluate the generalizability of various source tasks for transfer learning: natural image classification, natural image segmentation, unsupervised image reconstruction, and self‐supervised jigsaw solving. Axial CT slices at L3 were extracted from PET‐CT scans for 204 oesophago‐gastric cancer patients and the skeletal muscle manually delineated by an expert. Features were transferred and segmentation models trained on subsets (n=5,10,25,50,75,100,125) of the manually annotated training set. Four‐fold cross‐validation was performed to evaluate model generalizability. Human‐level performance was established by performing an inter‐observer study consisting of ten trained radiographers.

Results

We find that accurate segmentation models can be trained on a fraction of the data required by current approaches. The Dice similarity coefficient and root mean square distance‐to‐agreement were calculated for each prediction and used to assess model performance. Models pre‐trained on a segmentation task and fine‐tuned on 10 images produce delineations that are comparable to those from trained observers and extract reliable measures of muscle health.

Conclusions

Appropriate transfer learning can generate convolutional neural networks for abdominal muscle segmentation that achieve human‐level performance while decreasing the required data by an order of magnitude, compared to previous methods (n=160→10). This work enables the development of future models for assessing skeletal muscle at other anatomical sites where large annotated data sets are scarce and clinical needs are yet to be addressed.

DeepImageTranslator: A free, user-friendly graphical interface for image translation using deep-learning and its applications in 3D CT image analysis

The advent of deep-learning has set new standards in an array of image translation applications. At present, the use of these methods often requires computer programming experience. Non-commercial programs with graphical interface usually do not allow users to fully customize their deep-learning pipeline. Therefore, our primary objective is to provide a simple graphical interface that allows researchers with no programming experience to easily create, train, and evaluate custom deep-learning models for image translation. We also aimed to test the applicability of our tool in CT image semantic segmentation and noise reduction. DeepImageTranslator was implemented using the Tkinter library, the standard Python interface to the Tk graphical user interface toolkit; backend computations were implemented using data augmentation packages such as Pillow, Numpy, OpenCV, Augmentor, Tensorflow, and Keras libraries. Convolutional neural networks (CNNs) were trained using DeepImageTranslator. The effects of data augmentation, deep-supervision, and sample size on model accuracy were also systematically assessed. The DeepImageTranslator a simple tool that allows users to customize all aspects of their deep-learning pipeline, including the CNN, training optimizer, loss function, and the types of training image augmentation scheme. We showed that DeepImageTranslator can be used to achieve state-of-the-art accuracy and generalizability in semantic segmentation and noise reduction. Highly accurate 3D segmentation models for body composition can be obtained using training sample sizes as small as 17 images. In conclusion, an open-source deep-learning tool for accurate image translation with a user-friendly graphical interface was presented and evaluated. This standalone software can be downloaded at: https://sourceforge.net/projects/deepimagetranslator/.

Deep Learning-Based Abdominal Muscle Segmentation on CT Images of Surgical Patient Populations

Computed tomography (CT) is commonly used for the characterization and tracking of abdominal muscle mass in surgical patients for both pre-surgical outcome predictions and post-surgical monitoring of response to therapy. In order to accurately track changes of abdominal muscle mass, radiologists must manually segment CT slices of patients, a time-consuming task with potential for variability. In this work, we combined a fully convolutional neural network (CNN) with high levels of preprocessing to improve segmentation quality. We utilized a CNN based approach to remove patients' arms and fat from each slice and then applied a series of registrations with a diverse set of abdominal muscle segmentations to identify a best fit mask. Using this best fit mask, we were able to remove many parts of the abdominal cavity, such as the liver, kidneys, and intestines. This preprocessing was able to achieve a mean Dice similarity coefficient (DSC) of 0.53 on our validation set and 0.50 on our test set by only using traditional computer vision techniques and no artificial intelligence. The preprocessed images were then fed into a similar CNN previously presented in a hybrid computer vision-artificial intelligence approach and was able to achieve a mean DSC of 0.94 on testing data. The preprocessing and deep learning-based method is able to accurately segment and quantify abdominal muscle mass on CT images.

Association of skeletal muscle index with postoperative acute kidney injury in living donor hepatectomy: A retrospective single-centre cohort study

BACKGROUND

Although living donor liver transplantation (LDLT) is the standard treatment option for patients with end-stage liver disease, it always entails ethical concerns about the risk of living donors. Recent studies have reported a correlation between sarcopenia and surgical prognosis in recipients. However, there are few studies of donor sarcopenia and the surgical prognosis of donors. This study investigated the association between sarcopenia and postoperative acute kidney injury in liver donors.

METHODS

This retrospective study analysed 2892 donors who underwent donor hepatectomy for LDLT between January 2008 and January 2018. Sarcopenia was classified into pre-sarcopenia and severe sarcopenia, which were determined to be -1 standard deviation (SD), and -2 SD from the mean baseline of the skeletal muscle index, respectively. Multivariate regression analysis was performed to evaluate the association between donor sarcopenia and postoperative AKI. Additionally, we assessed the association between donor sarcopenia and delayed recovery of liver function (DRHF).

RESULTS

In the multivariate analysis, donor sarcopenia was significantly associated a higher incidence of postoperative AKI (adjusted odds ratio [OR]: 2.65, 95% confidence interval [CI]: 1.15-6.11, P = .022 in pre-sarcopenia, OR: 5.59, 95% CI: 1.11-28.15, P = .037 in severe sarcopenia, respectively). Additionally, hypertension and synthetic colloid use were significantly associated with postoperative AKI. In the multivariate analysis, risk factors of DRHF were male gender, indocyanine green retention rate at 15 minutes, and graft type, however, donor sarcopenia was not a risk factor.

CONCLUSIONS

Donor sarcopenia is associated with postoperative AKI following donor hepatectomy.

A reliable and robust method for the upper thigh muscle quantification on computed tomography: toward a quantitative biomarker for sarcopenia

Background

We aimed to evaluate the feasibility of the upper thigh level as a landmark to measure muscle area for sarcopenia assessment on computed tomography (CT).

Methods

In the 116 healthy subjects who performed CT scans covering from mid-abdomen to feet, the skeletal muscle area in the upper thigh level at the inferior tip of ischial tuberosity (SMA_UT), the mid-thigh level (SMA_MT), and L3 inferior endplate level (SMA_L3) were measured by two independent readers. Pearson correlation coefficients between SMA_UT, SMA_MT, and SMA_L3 were calculated. Inter-reader agreement between the two readers were evaluated using intraclass correlation coefficient (ICC) and Bland-Altman plots with 95% limit of agreement (LOA).

Results

In readers 1 and 2, very high positive correlations were observed between SMA_UT and SMA_MT (r = 0.91 and 0.92, respectively) and between SMA_UT and SMA_L3 (r = 0.90 and 0.91, respectively), while high positive correlation were observed between SMA_MT and SMA_L3 (r = 0.87 and 0.87, respectively). Based on ICC values, the inter-reader agreement was the best in the SMA_UT (0.999), followed by the SMA_L3 (0.990) and SMA_MT (0.956). The 95% LOAs in the Bland-Altman plots indicated that the inter-reader agreement of the SMA_UT (− 0.462 to 1.513) was the best, followed by the SMA_L3 (− 9.949 to 7.636) and SMA_MT (− 12.105 to 14.605).

Conclusion

Muscle area measurement at the upper thigh level correlates well with those with the mid-thigh and L3 inferior endpoint level and shows the highest inter-reader agreement. Thus, the upper thigh level might be an excellent landmark enabling SMA_UT as a reliable and robust biomarker for muscle area measurement for sarcopenia assessment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-022-05032-2.

Validation for measurements of skeletal muscle areas using low-dose chest computed tomography

Various methods were suggested to measure skeletal muscle areas (SMAs) using chest low-dose computed tomography (chest LDCT) as a substitute for SMA at 3rd lumbar vertebra level (L3-SMA). In this study, four SMAs (L1-SMA, T12-erector spinae muscle areas, chest wall muscle area at carina level, pectoralis muscle area at aortic arch level) were segmented semi-automatically in 780 individuals taking concurrent chest and abdomen LDCT for healthcare screening. Four SMAs were compared to L3-SMA and annual changes were calculated from individuals with multiple examinations (n = 101). Skeletal muscle index (SMI; SMA/height²) cut-off for sarcopenia was determined by lower 5th percentile of young individuals (age ≤ 40 years). L1-SMA showed the greatest correlation to L3-SMA (men, R² = 0.7920; women, R² = 0.7396), and the smallest annual changes (0.3300 ± 4.7365%) among four SMAs. L1-SMI cut-offs for determining sarcopenia were 39.2cm²/m² in men, and 27.5cm²/m² in women. Forty-six men (9.5%) and ten women (3.4%) were found to have sarcopenia using L1-SMI cut-offs. In conclusion, L1-SMA could be a reasonable substitute for L3-SMA in chest LDCT. Suggested L1-SMI cut-offs for sarcopenia were 39.2cm²/m² for men and 27.5cm²/m² for women in Asian.

Larger Remaining Stomach Volume Is Associated With Better Nutrition and Muscle Preservation in Patients With Gastric Cancer Receiving Distal Gastrectomy With Gastroduodenostomy

Purpose

Weight loss and deterioration in body composition are observed in patients with gastric cancer (GC) following gastrectomy. This study aimed to investigate the impact of residual stomach volume (RSV) on the nutritional status and body composition of patients with GC treated with distal gastrectomy.

Materials and Methods

In total, 227 patients who underwent minimally invasive distal gastrectomy with Billroth 1 anastomosis for stage I GC between February 2015 and May 2018 were enrolled. Clinicodemographic and laboratory data were collected from the GC registry. The RSV, abdominal muscle area, and subcutaneous/visceral fat areas were measured using computed tomography data.

Results

A larger RSV was associated with a lower decrease in the nutritional risk index (P=0.004) and hemoglobin level (P=0.003) during the first 3 months after surgery, and better recovery at 12 months. A larger RSV demonstrated an advantage in the preservation of abdominal muscle area (P=0.02) and visceral fat (P=0.04) after surgery, as well as less reduction in weight (P=0.02) and body mass index (P=0.03).

Conclusions

Larger RSV was associated with improved nutritional status and better preservation of muscle and fat after distal gastrectomy.