Abdominal CT Images Research Articles

Domain generalization on medical imaging classification using episodic training with task augmentation

Medical imaging datasets usually exhibit domain shift due to the variations of scanner vendors, imaging protocols, etc. This raises the concern about the generalization capacity of machine learning models. Domain generalization (DG), which aims to learn a model from multiple source domains such that it can be directly generalized to unseen test domains, seems particularly promising to medical imaging community. To address DG, recent model-agnostic meta-learning (MAML) has been introduced, which transfers the knowledge from previous training tasks to facilitate the learning of novel testing tasks. However, in clinical practice, there are usually only a few annotated source domains available, which decreases the capacity of training task generation and thus increases the risk of overfitting to training tasks in the paradigm. In this paper, we propose a novel DG scheme of episodic training with task augmentation on medical imaging classification. Based on meta-learning, we develop the paradigm of episodic training to construct the knowledge transfer from episodic training-task simulation to the real testing task of DG. Motivated by the limited number of source domains in real-world medical deployment, we consider the unique task-level overfitting and we propose task augmentation to enhance the variety during training task generation to alleviate it. With the established learning framework, we further exploit a novel meta-objective to regularize the deep embedding of training domains. To validate the effectiveness of the proposed method, we perform experiments on histopathological images and abdominal CT images.

Elimination of biological and physical artifacts in abdomen and brain computed tomography procedures using filtering techniques

IntroductionMedical images are usually affected by biological and physical artifacts or noise, which reduces image quality and hence poses difficulties in visual analysis, interpretation and thus requires higher doses and increased radiographs repetition rate. ObjectivesThis study aims at assessing image quality during CT abdomen and brain examinations using filtering techniques as well as estimating the radiogenic risk associated with CT abdomen and brain examinations. Materials and MethodsThe data were collected from the Radiology Department at Royal Care International (RCI) Hospital, Khartoum, Sudan. The study included 100 abdominal CT images and 100 brain CT images selected from adult patients. Filters applied are namely: Mean filter, Gaussian filter, Median filter and Minimum filter. In this study, image quality after denoising is measured based on the Mean Squared Error (MSE), Peak Signal-to-Noise Ratio (PSNR), and the Structural Similarity Index Metric (SSIM). ResultsThe results show that the images quality parameters become higher after applications of filters. Median filter showed improved image quality as interpreted by the measured parameters: PSNR and SSIM, and it is thus considered as a better filter for removing the noise from all other applied filters. DiscussionThe noise removed by the different filters applied to the CT images resulted in enhancing high quality images thereby effectively revealing the important details of the images without increasing the patients’ risks from higher doses. ConclusionsFiltering and image reconstruction techniques not only reduce the dose and thus the radiation risks, but also enhances high quality imaging which allows better diagnosis.

Comparative Image Quality Analysis of Spatial Filters for Pre-processing of CT Abdominal Images

Aim: To determine the efficient noise reduction filter for abdominal CT images. Background: Image enrichment is the first and foremost step that has to be done in all image processing applications. It is used to enhance the quality of digital images. Digital images are liable to addition of noise from various sources such as error in instrument calibration, excess staining of images, etc., Image de-noising is an enhancement technique used to remove / reduce noise present in an image. Reducing the noise of images and preserving its edges are always critical and challenging in image processing. Materials and Method: In this paper, four different spatial filters namely Mean, Median, Gaussian and Wiener were used on 100 CT abdominal images and their performances were compared against the following four parameters: Peak signal to noise ratio (PSNR), Mean Square Error (MSE), Normalised correlation coefficient (NCC) and Normalised Absolute Error (NAE) to determine the best denoising filter for the abdominal CT images. Result: Based on the experimental parameters, the median filter had the maximum efficiency in managing salt and pepper noise than the other three filters. Both Median and Wiener filters showed efficiency in removing the Gaussian noise. Whereas, the Wiener filter demonstrated higher efficiency in reducing both Poisson and Speckle noise. Conclusion: Based on the results of this study, we can conclude that the median filter can be used to reduce Salt and Pepper noises. Median and Wiener filters are significantly better for Gaussian Noise and the Wiener filter can be used to reduce both Poisson & Speckle noise in abdominal CT images.

Mesenteric ischemia in patients with COVID-19: an updated systematic review of abdominal CT findings in 75 patients.

BackgroundAcute mesenteric ischemia (AMI) is a less common but devastating complication of COVID-19 disease. The aim of this systematic review was to assess the most common CT imaging features of AMI in COVID-19 and also provide an updated review of the literature on symptoms, treatment, histopathological and operative findings, and follow-up of these patients.MethodsA systematic literature search of four databases: Pubmed, EMBASE, WHO database, and Google Scholar, was performed to identify all the articles which described abdominal CT imaging findings of AMI in COVID-19.ResultsA total of 47 studies comprising 75 patients were included in the final review. Small bowel ischemia (46.67%) was the most prevalent abdominal CT finding, followed by ischemic colitis (37.3%). Non-occlusive mesenteric ischemia (NOMI; 67.9%) indicating microvascular involvement was the most common pattern of bowel involvement. Bowel wall thickening/edema (50.9%) was more common than bowel hypoperfusion (20.7%). While ileum and colon both were equally involved bowel segments (32.07% each), SMA (24.9%), SMV (14.3%), and the spleen (12.5%) were the most commonly involved artery, vein, and solid organ, respectively. 50% of the patients receiving conservative/medical management died, highlighting high mortality without surgery. Findings on laparotomy and histopathology corroborated strikingly with CT imaging findings.ConclusionIn COVID-19 patients with AMI, small bowel ischemia is the most prevalent imaging diagnosis and NOMI is the most common pattern of bowel involvement. Contrast-enhanced CT is a powerful decision-making tool for prompt diagnosis of AMI in COVID-19, thereby potentially improving time to treat as well as clinical outcomes. Graphical abstract Supplementary InformationThe online version contains supplementary material available at 10.1007/s00261-021-03337-9.

An Automatic Detection of Liver Tumor from CT Abdominal Images—A Comparative Approach

The liver is a vital organ in human body. Liver performs an important function including metabolism, digestion, and detoxification. Liver is a significant organ in an abdomen, and is connected to the nearby organ such as spleen, pancreas, gallbladder, abdomen, and gut through blood vessels. Specific approaches such as image gradient and region growing are not quite reliable for the segmentation of the liver tumor. A level-set approach is evaluated in this paper compared with the active contour approach of segmentation of the liver imaging from the image of the CT abdomen and Unified level set method, spatial Fuzzy C-means method for segmenting tumor from segmented liver images is appraised. The proposed approach is implemented by using the 3DIRCADB dataset available to the public as well as non-public datasets taken from Arthi Hospital, Chennai and Tirunelveli scanning centre. For validating the system based on the diverse quantitative measures, including space overlap, coefficient of similarity, Jaccard indices, using ground truth images, which are available in the public data set 3DIRCADB and the expert segmentation results which are manually identified by the clinical partner for nonpublic datasets. The analysis of the algorithm shows the better results for segmenting liver using level set system and spatial segmentation of Fuzzy C means of the tumor segmentation.

Abnormal body composition in patients with adrenal adenomas.

Increased visceral fat and sarcopenia are cardiovascular risk factors that may explain increased cardiovascular morbidity and frailty in patients with adrenal adenomas. Our objective was to compare body composition measurement of patients with adrenal adenomas to referent subjects without adrenal disease. Cross-sectional study, 2014-2018. Participants were adults with nonfunctioning adrenal tumor (NFAT), mild autonomous cortisol secretion (MACS), and Cushing syndrome (CS) and age, sex, and BMI 1:1 matched referent subjects without adrenal disorders. Main outcome measures were body composition measurements calculated from abdominal CT imaging. Intra-abdominal adipose tissue and muscle mass measurements were performed at the third lumbar spine level. Of 227 patients with adrenal adenomas, 20 were diagnosed with CS, 76 with MACS, and 131 with NFAT. Median age was 56 years (range: 18-89), and 67% were women. When compared to referent subjects, patients with CS, MACS, and NFAT demonstrated a higher visceral fat (odds ratio (OR): 2.2 (95% CI: 0.9-6.5), 2.0 (1.3-3.2), and 1.8 (1.2-2.7) and a lower skeletal muscle area (OR: 0.01 (95% CI: 0-0.09), 0.31 (0.18-0.49), and 0.3 (1.2-2.7)) respectively. For every 1 µg/dL cortisol increase after overnight dexamethasone, visceral fat/muscle area ratio increased by 2.3 (P = 0.02) and mean total skeletal muscle area decreased by 2.2 cm2 (P = 0.03). Patients with adrenal adenomas demonstrate a lower muscle mass and a higher proportion of visceral fat when compared to referent subjects, including patients with NFAT. Even a subtle abnormality in cortisol secretion may impact health of patients with adenomas.

Early Prediction of Acute Biliary Pancreatitis Using Clinical and Abdominal CT Features.

Background Assessment of the biliary origin of acute pancreatitis (AP) is crucial because it affects patient treatment to avoid recurrence. Although CT is systematically performed to determine severity in AP, its usefulness in assessing AP biliary origin has not been evaluated. Purpose To assess abdominal CT features associated with acute biliary pancreatitis (ABP) and to evaluate the predictive value of a combination of CT and clinical data for determining a biliary origin in a first episode of AP. Materials and Methods From December 2014 to May 2019, all consecutive patients who presented with a first episode of AP and with at least 6 months of follow-up were retrospectively reviewed. Evidence of gallstones was mandatory for a clinical diagnosis of ABP. Abdominal CT images were reviewed by two abdominal radiologists. Univariable and multivariable statistical analyses were performed, and a nomogram was constructed on the basis of the combination of clinical and CT features. This nomogram was validated in a further independent internal cohort of patients. Results A total of 271 patients (mean age ± standard deviation, 56 years ± 20; 160 men) were evaluated. Of these, 170 (63%) had ABP. At multivariable analysis, age (odds ratio [OR], 1.06; 95% CI: 1.03, 1.09; P < .001), alanine aminotransferase level (OR, 1.00; 95% CI: 1.00, 1.01; P = .009), gallbladder gallstone (OR, 15.59; 95% CI: 4.61, 68.62; P < .001), choledochal ring sign (OR, 5.73; 95% CI: 2.11, 17.05; P < .001), liver spontaneous attenuation (OR, 1.07; 95% CI: 1.04, 1.11; P < .001), and duodenal thickening (OR, 0.17; 95% CI: 0.03, 0.61; P = .01) were independently associated with ABP. The matching nomogram combining both clinical and CT features displayed an area under the curve of 0.94 (95% CI: 0.91, 0.97) in the study sample (n = 271) and 0.91 (95% CI: 0.84, 0.99) in the validation cohort (n = 51). Conclusion Abdominal CT provided useful features for diagnosis of acute biliary pancreatitis (ABP). Combining CT and clinical features in a nomogram showed good diagnostic performance for early diagnosis of ABP. © RSNA, 2021 Online supplemental material is available for this article. See also the editorial by Chang in this issue.

Low skeletal muscle mass by computerized tomography is associated with increased mortality risk in end-stage kidney disease patients on hemodialysis.

Skeletal muscle (SM) area, as measured by abdominal CT at the level of the third lumbar vertebra (L3), has been proposed as a proxy of whole body muscle mass. However, population-specific reference values are lacking. In the present study we aimed at: (1) detecting low SM area on abdominal CT images in patients on hemodialysis by applying cut-offs derived from a group of healthy subjects, and (2) estimating the independent risk of all-cause mortality associated with low SM area. We retrospectively enrolled 212 adult patients on hemodialysis, undergoing abdominal CT scan (study group), and 87 healthy kidney donors (reference group). We obtained the gender-specific 5th percentile values of the abdominal SM area distribution from both the whole control group and the subgroup of younger (29-60years) subjects, which we used as reference cut-offs. Then we applied those cut-offs in the study group to identify patients with low SM area. We used survival and Cox regression analysis to evaluate the risk of all-cause mortality associated with low abdominal SM area. In the fully adjusted Cox regression analysis, the patients with low abdominal SM area had a higher risk of death than the patients with values above the reference cut-off derived in the subgroup of younger controls (adjHR = 1.79 (1.21; 2.67), P = 0.004). Abdominal CT imaging can be used to detect low abdominal SM area in patients on hemodialysis by applying cut-offs derived from healthy subjects sharing a similar ethnic background. Low SM area as assessed by CT is independently associated with all-cause mortality in ESKD patients on hemodialysis.

The Effect of Deep Learning-Based QSM Magnetic Resonance Imaging on the Subthalamic Nucleus.

In order to study the influence of quantitative magnetic susceptibility mapping (QSM) on them. A 2.5D Attention U-Net Network based on multiple input and multiple output, a method for segmenting RN, SN, and STN regions in high-resolution QSM images is proposed, and deep learning realizes accurate segmentation of deep nuclei in brain QSM images. Experimental results show data first cuts each layer of 0 100 case data, based on the image center, from 384 × 288 to the size of 128 × 128. Image combination: each layer of the image in the layer direction combines with two adjacent images into a 2.5D image, i.e., (It − m It; It + i), where It represents the layer i image. At this time, the size of the image changes from 128 × 128 to 128 × 128 × 3, in which 3 represents three consecutive layers of images. The SNR of SWP I to STN is twice that of SWI. The small deep gray matter nuclei (RN, SN, and STN) in QSM images of the brain and the pancreas with irregular shape and large individual differences in abdominal CT images can be automatically segmented.

The frequency of the bony parameters of femoroacetabular impingement syndrome in young asymptomatic individuals: a computed tomography study

Objectives: Femoroacetabular impingement syndrome is a painful disorder of the hip formed by movements in the hip joint between the femoral head-neck junction and the acetabular edge. The aim of this study was to determine the prevalence of radiological femoroacetabular impingement syndrome findings in a young, asymptomatic Turkish population. Methods: A total of 1000 abdominal pelvic CT images were collected from patients aged between 18–40 years. Measurements were taken on the CT images of the alpha angle and femoral head offset as signs of cam-type deformity, and of the acetabular version angle and center-edge angle as signs of pincer-type deformity. Results: Femoroacetabular impingement syndrome was determined in 2.3% of the individuals. Pincer type deformity was determined in 56.5%, being 83.3% in females and 47.1% in males. The cam-type deformity was present in 43.5% of all cases, being in 16.7% of the females and in 52.9% of the males. No difference was determined between the genders in respect of alpha angle which is &gt;55°. A head-neck offset which is &lt;8 mm was at a higher rate in females (13.3%) than males. An acetabular anteversion angle which is &lt;15° and centre-edge angle which is &gt;40° was more in males than females at a rate of 11% and 25.2% respectively. Conclusion: The results of this study showed that the incidence of cam-type deformity in the adult males and the incidence of pincer-type deformity in the adult females were lower in the asymptomatic Turkish population than previously reported in literature.

Classification of focal liver lesions in CT images using convolutional neural networks with lesion information augmented patches and synthetic data augmentation.

We propose a deep learning method that classifies focal liver lesions (FLLs) into cysts, hemangiomas, and metastases from portal phase abdominal CT images. We propose a synthetic data augmentation process to alleviate the class imbalance and the Lesion INformation Augmented (LINA) patch to improve the learning efficiency. A dataset of 502 portal phase CT scans of 1,290 FLLs was used. First, to alleviate the class imbalance and to diversify the training data patterns, we suggest synthetic training data augmentation using DCGAN-based lesion mask synthesis and pix2pix-based mask-to-image translation. Second, to improve the learning efficiency of convolutional neural networks (CNNs) for the small lesions, we propose a novel type of input patch termed the LINA patch to emphasize the lesion texture information while also maintaining the lesion boundary information in the patches. Third, we construct a multi-scale CNN through a model ensemble of ResNet-18 CNNs trained on LINA patches of various mini-patch sizes. The experiments demonstrate that (a) synthetic data augmentation method shows characteristics different but complementary to those in conventional real data augmentation in augmenting data distributions, (b) the proposed LINA patches improve classification performance compared to those by existing types of CNN input patches due to the enhanced texture and boundary information in the small lesions, and (c) through an ensemble of LINA patch-trained CNNs with different mini-patch sizes, the multi-scale CNN further improves overall classification performance. As a result, the proposed method achieved an accuracy of 87.30%, showing improvements of 10.81%p and 15.0%p compared to the conventional image patch-trained CNN and texture feature-trained SVM, respectively. The proposed synthetic data augmentation method shows promising results in improving the data diversity and class imbalance, and the proposed LINA patches enhance the learning efficiency compared to the existing input image patches.

Denoising of MR and CT images using cascaded multi-supervision convolutional neural networks with progressive training

As MR Rician noise and CT low-dose perfusion noise have a complicated distribution, it is still a challenging task to automatically and accurately remove the noises existing in MR and CT images. Recently, convolutional neural networks have shown favorable performance on image denoising tasks. However, existing methods ignored the hierarchical features extracted from multi-supervision inner layers and estimated the denoised image just by the last single layer, which can not adequately reserve the details of the image. In this paper, we propose a cascaded multi-supervision convolutional neural network named CMSNet to remove the low-dose perfusion noise in CT images and the Rician noise exist in MR images. The CMSNet consists of a multi-supervision network (MSNet) followed with a Refinement network. MSNet is presented to predict the noise constrained by the supervisions from last three convolution layers, which can help acquire more accurate noise prediction and thus obtain the noise-free image. Refinement network is introduced to relief the details lost problem caused by the denoising operation. We employ a progressive training strategy, i.e., MSNet is first trained independently to predict the preliminary noise and then jointly trained with Refinement network for more accurate noise estimating, which can boost the network performance. Experiments are conducted on clinic abdominal MR and CT images, and the results show that our proposed model achieved a promising performance in terms of unknown noise level, a specific noise level on peak signal to noise ratio (PSNR) and global structure similarity index measurement (SSIM).

Reducing reconstruction error of classified textural patches by integration of random forests and coupled dictionary nonlinear regressors: with applications to super-resolution of abdominal CT images.

Random forests and dictionary-based statistical regressions have common characteristics, including non-linear mapping and supervised learning. To reduce the reconstruction error of high-resolution images, we integrate random forests and coupled dictionary learning. Textural differences of image blocks are considered by the classification of patches using an Auto-Encoder network. The proposed algorithm partitions an input LR image by 5 × 5 blocks and classifies training patches into six categories. A single random forest regressor is then trained corresponding to each class. The output of an RF is considered as an initial estimate of the HR slice. If a slice's representation is sparse in the Discrete Cosine Transform domain, the initial reconstructed image is further improved by a coupled dictionary. In this study, we applied our method to abdominal CT scans and compared them to conventional and recent researches. We achieved an average improvement of 0.06 (2.37) using the SSIM (PSNR) index compared to the random forest + dictionary learning method. The low standard deviation of the results reveals the stability of the proposed method as well. The proposed algorithm depicts the effectiveness of classifying image patches and individual treatment of each class.

Multi-slice low-rank tensor decomposition based multi-atlas segmentation: Application to automatic pathological liver CT segmentation

Liver segmentation from abdominal CT images is an essential step for liver cancer computer-aided diagnosis and surgical planning. However, both the accuracy and robustness of existing liver segmentation methods cannot meet the requirements of clinical applications. In particular, for the common clinical cases where the liver tissue contains major pathology, current segmentation methods show poor performance. In this paper, we propose a novel low-rank tensor decomposition (LRTD) based multi-atlas segmentation (MAS) framework that achieves accurate and robust pathological liver segmentation of CT images. Firstly, we propose a multi-slice LRTD scheme to recover the underlying low-rank structure embedded in 3D medical images. It performs the LRTD on small image segments consisting of multiple consecutive image slices. Then, we present an LRTD-based atlas construction method to generate tumor-free liver atlases that mitigates the performance degradation of liver segmentation due to the presence of tumors. Finally, we introduce an LRTD-based MAS algorithm to derive patient-specific liver atlases for each test image, and to achieve accurate pairwise image registration and label propagation. Extensive experiments on three public databases of pathological liver cases validate the effectiveness of the proposed method. Both qualitative and quantitative results demonstrate that, in the presence of major pathology, the proposed method is more accurate and robust than state-of-the-art methods.

Deep Learning-Based Automatic Detection of Rectal Polyps Using Abdominal CT Images Guided by Cold Snare Polypectomy

The study drew attention to the therapeutic effects of cold snare polypectomy guided by a deep convolutional neural network- (CNN-) based abdominal CT and hot snare polypectomy (HSP) on colonic and rectal polyps. Specifically, 90 patients were enrolled into a blank group, a control (Ctrl) group, and an experimental group. The blank group accepted HSP, the Ctrl accepted cold snare polypectomy, and the experimental group accepted cold snare polypectomy guided by deep CNN-based CT images. It was found that the experimental group had the lowest false-positive rate (9.2%) in polyp detection in contrast with the Ctrl (21.4%) and the blank group (52.3%) P &lt; 0.05 . The complete resection rate of large polyps in the experimental group was the highest P &lt; 0.05 , and its operation time (2.91 ± 0.75 min) was obviously shorter versus the blank group (6.18 ± 1.19 min) P &lt; 0.05 . In conclusion, the cold snare polypectomy under the guidance of deep CNN-based CT has a relatively high complete resection rate and detection accuracy of polyps with a low complication rate, which can be adopted clinically.

Abdominal CT in patients with arms down positioning: Effect of virtual monoenergetic reconstruction on diagnostic image quality.

To evaluate effect of virtual monoenergetic reconstruction on contrast-enhanced abdominal CT obtained using a dual-layer detector CT (DLCT) in patients with arms-down positioning, in comparison with conventional 120-kVp image. A total of 162 consecutive patients who underwent abdominal CT scans using a DLCT with arms-down positioning were retrospectively enrolled. 120-kVp image and virtual monoenergetic images (VMIs) at 40-100keV were compared for arm-induced artifact reduction capability and image quality through objective and subjective assessments. Diagnostic performance was compared between VMI with the best quality and 120-kVp image. VMIs at 60-100keV showed significantly lower (p < 0.001) artifact index than 120-kVp image. Subjective severity of arm-induced streak artifacts was significantly (p < 0.001) lower in VMIs at 70-100keV than in 120-kVp image. CNRs measured in artifact and reference areas were superior (p < 0.05) in 40-70keV to those in 120 kVp. Overall image quality was superior (p < 0.05) in 60-90keV to 120 kVp. 70-keV VMI was chosen as the best image by two independent readers. In ROC analysis of lesion detection in the liver, spleen, and kidneys, the AUC was significantly higher for 70-keV VMI (reader 1, 0.962; p = 0.011; reader 2, 0.981; p = 0.012) than that for 120-kVp image (reader 1, 0.891; reader 2, 0.922). The70-keV VMI from DLCT could improve abdominal CT images by reducing streak artifacts and increasing image contrast in patients who could not elevated their arms. The 70-keV VMI showed better diagnostic performance than the conventional 120 kVp image.

Radiomics Analysis of Contrast-Enhanced CT Predicts Survival in Clear Cell Renal Cell Carcinoma.

PurposeTo develop and validate the radiomics nomogram that combines clinical factors and radiomics features to estimate overall survival (OS) in patients with clear cell renal cell carcinoma (ccRCC), and assess the incremental value of radiomics for OS estimation.Materials and MethodsOne hundred ninety-four ccRCC cases were included in the training cohort and 188 ccRCC patients from another hospital as the test cohort. Three-dimensional region-of-interest segmentation was manually segmented on multiphasic contrast-enhanced abdominal CT images. Radiomics score (Rad-score) was calculated from a formula generated via least absolute shrinkage and selection operator (LASSO) Cox regression, after which the association between the Rad-score and OS was explored. The radiomics nomogram (clinical factors + Rad-score) was developed to demonstrate the incremental value of the Rad-score to the clinical nomogram for individualized OS estimation, which was then evaluated in relation to calibration and discrimination.ResultsRad-score, calculated using a linear combination of the 11 screened features multiplied by their respective LASSO Cox coefficients, was significantly associated with OS. Calibration curves showed good agreement between the OS predicted by the nomograms and observed outcomes. The radiomics nomogram presented higher discrimination capability compared to clinical nomogram in the training (C-index: 0.884; 95% CI: 0.808–0.940 vs. 0.803; 95% CI: 0.705–0.899, P < 0.05) and test cohorts (C-index: 0.859; 95% CI: 0.800–0.921 vs. 0.846; 95% CI: 0.777–0.915, P < 0.05).ConclusionsThe radiomics nomogram may be used for predicting OS in patients with ccRCC, and radiomics is useful to assist quantitative and personalized treatment.

Small bowel obstruction due to internal herniation through a defect in the falciform ligament: a rare case report and review of the literature

BackgroundThe falciform ligament hernia is a type of internal herniation, caused by a defect in the falciform ligament of the liver. The etiology for this defect may be congenital or iatrogenic, created after the laparoscopic intervention.Case presentationA 45-year-old male was presented to our hospital, complaining of abdominal pain for 2 days. The physical examination revealed diffuse abdominal tenderness and guarding. The abdominal X-ray was suggestive of intestinal obstruction. On abdominal CT images, dilated proximal ileal loops between the left hepatic lobe and anterior abdominal wall, forming a closed-loop with mesenteric edema and fat stranding, were observed. There were findings of ileus in the jejunal loops proximal to this segment, but the ileal loops and the colon were completely collapsed, consistent with small bowel obstruction due to internal herniation through a defect in the falciform ligament. The patient underwent surgical intervention, reduction of the herniated bowel loops, and repairing the hernia defect. He had an uneventful recovery with a favorable outcome.ConclusionHerniation through a defect in the falciform ligament is extremely rare but should be considered in the differential diagnosis of acute abdomen. The clinical manifestations of falciform ligament hernia are nonspecific and may underestimate the diagnosis, leading to a delayed treatment that affects the management outcome. Computed tomography plays an important role in the timely diagnosis and planning of surgical intervention, precluding intestinal strangulation.

복부 CT 영상에서 췌장의 불확실성을 고려한 계층적 네트워크 기반 자동 췌장 분할

복부 CT 영상에서 췌장의 불확실성을 고려한 계층적 네트워크 기반 자동 췌장 분할

Abdominal multi-organ segmentation with cascaded convolutional and adversarial deep networks

Abdominal anatomy segmentation is crucial for numerous applications from computer-assisted diagnosis to image-guided surgery. In this context, we address fully-automated multi-organ segmentation from abdominal CT and MR images using deep learning. The proposed model extends standard conditional generative adversarial networks. Additionally to the discriminator which enforces the model to create realistic organ delineations, it embeds cascaded partially pre-trained convolutional encoder-decoders as generator. Encoder fine-tuning from a large amount of non-medical images alleviates data scarcity limitations. The network is trained end-to-end to benefit from simultaneous multi-level segmentation refinements using auto-context. Employed for healthy liver, kidneys and spleen segmentation, our pipeline provides promising results by outperforming state-of-the-art encoder-decoder schemes. Followed for the Combined Healthy Abdominal Organ Segmentation (CHAOS) challenge organized in conjunction with the IEEE International Symposium on Biomedical Imaging 2019, it gave us the first rank for three competition categories: liver CT, liver MR and multi-organ MR segmentation. Combining cascaded convolutional and adversarial networks strengthens the ability of deep learning pipelines to automatically delineate multiple abdominal organs, with good generalization capability. The comprehensive evaluation provided suggests that better guidance could be achieved to help clinicians in abdominal image interpretation and clinical decision making.