Liver Dataset Research Articles

Adaptive feature aggregation based multi-task learning for uncertainty-guided semi-supervised medical image segmentation

Automatic segmentation of medical images is a necessary prerequisite for diagnosing related diseases. Magnetic resonance imaging (MRI) is a widely used non-invasive method in clinical practice, but obtaining reliable labeled training data for deep learning models is time-consuming and challenging. In this study, we propose an adaptive feature aggregation-based multi-task learning for uncertainty-guided semi-supervised image segmentation, dubbed AFAM-Net. The framework leverages a reconstruction task to capture anatomical information from raw images and assist the segmentation task. Moreover, we propose an adaptive feature aggregation strategy that selectively transfers useful features while filtering out irrelevant information, considering the associations between the two tasks. To better utilize unlabeled data, we incorporate dual uncertainty-aware methods to improve the segmentation performance. We evaluate the proposed AFAM-Net on clinical liver, 3D heart, and Cine cardiac MRI datasets. According to the experimental results, our AFAM-Net significantly outperformed other state-of-the-art semi-supervised medical image segmentation algorithms.

Identification of a broadly fibrogenic macrophage subset induced by type 3 inflammation.

Macrophages are central orchestrators of the tissue response to injury, with distinct macrophage activation states playing key roles in fibrosis progression and resolution. Identifying key macrophage populations found in human fibrotic tissues could lead to new treatments for fibrosis. Here, we used human liver and lung single-cell RNA sequencing datasets to identify a subset of CD9+TREM2+ macrophages that express SPP1, GPNMB, FABP5, and CD63. In both human and murine hepatic and pulmonary fibrosis, these macrophages were enriched at the outside edges of scarring and adjacent to activated mesenchymal cells. Neutrophils expressing MMP9, which participates in the activation of TGF-β1, and the type 3 cytokines GM-CSF and IL-17A coclustered with these macrophages. In vitro, GM-CSF, IL-17A, and TGF-β1 drive the differentiation of human monocytes into macrophages expressing scar-associated markers. Such differentiated cells could degrade collagen IV but not collagen I and promote TGF-β1-induced collagen I deposition by activated mesenchymal cells. In murine models blocking GM-CSF, IL-17A or TGF-β1 reduced scar-associated macrophage expansion and hepatic or pulmonary fibrosis. Our work identifies a highly specific macrophage population to which we assign a profibrotic role across species and tissues. It further provides a strategy for unbiased discovery, triage, and preclinical validation of therapeutic targets based on this fibrogenic macrophage population.

Training Spiking Neural Networks with Metaheuristic Algorithms

Taking inspiration from the brain, spiking neural networks (SNNs) have been proposed to understand and diminish the gap between machine learning and neuromorphic computing. Supervised learning is the most commonly used learning algorithm in traditional ANNs. However, directly training SNNs with backpropagation-based supervised learning methods is challenging due to the discontinuous and non-differentiable nature of the spiking neuron. To overcome these problems, this paper proposes a novel metaheuristic-based supervised learning method for SNNs by adapting the temporal error function. We investigated seven well-known metaheuristic algorithms called Harmony Search (HS), Cuckoo Search (CS), Differential Evolution (DE), Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Artificial Bee Colony (ABC), and Grammatical Evolution (GE) as search methods for carrying out network training. Relative target firing times were used instead of fixed and predetermined ones, making the computation of the error function simpler. The performance of our proposed approach was evaluated using five benchmark databases collected in the UCI Machine Learning Repository. The experimental results showed that the proposed algorithm had a competitive advantage in solving the four classification benchmark datasets compared to the other experimental algorithms, with accuracy levels of 0.9858, 0.9768, 0.7752, and 0.6871 for iris, cancer, diabetes, and liver datasets, respectively. Among the seven metaheuristic algorithms, CS reported the best performance.

High-Resolution Swin Transformer for Automatic Medical Image Segmentation.

The resolution of feature maps is a critical factor for accurate medical image segmentation. Most of the existing Transformer-based networks for medical image segmentation adopt a U-Net-like architecture, which contains an encoder that converts the high-resolution input image into low-resolution feature maps using a sequence of Transformer blocks and a decoder that gradually generates high-resolution representations from low-resolution feature maps. However, the procedure of recovering high-resolution representations from low-resolution representations may harm the spatial precision of the generated segmentation masks. Unlike previous studies, in this study, we utilized the high-resolution network (HRNet) design style by replacing the convolutional layers with Transformer blocks, continuously exchanging feature map information with different resolutions generated by the Transformer blocks. The proposed Transformer-based network is named the high-resolution Swin Transformer network (HRSTNet). Extensive experiments demonstrated that the HRSTNet can achieve performance comparable with that of the state-of-the-art Transformer-based U-Net-like architecture on the 2021 Brain Tumor Segmentation dataset, the Medical Segmentation Decathlon's liver dataset, and the BTCV multi-organ segmentation dataset.

N-Net: an UNet architecture with dual encoder for medical image segmentation.

In order to assist physicians in diagnosis and treatment planning, accurate and automatic methods of organ segmentation are needed in clinical practice. UNet and its improved models, such as UNet + + and UNt3 + , have been powerful tools for medical image segmentation. In this paper, we focus on helping the encoder extract richer features and propose a N-Net for medical image segmentation. On the basis of UNet, we propose a dual encoder model to deepen the network depth and enhance the ability of feature extraction. In our implementation, the Squeeze-and-Excitation (SE) module is added to the dual encoder model to obtain channel-level global features. In addition, the introduction of full-scale skip connections promotes the integration of low-level details and high-level semantic information. The performance of our model is tested on the lung and liver datasets, and compared with UNet, UNet + + and UNet3 + in terms of quantitative evaluation with the Dice, Recall, Precision and F1 score and qualitative evaluation. Our experiments demonstrate that N-Net outperforms the work of UNet, UNet + + and UNet3 + in these three datasets. By visual comparison of the segmentation results, N-Net produces more coherent organ boundaries and finer details.

Phase-regularized and displacement-regularized compressed sensing for fast magnetic resonance elastography.

Liver magnetic resonance elastography (MRE) is a noninvasive stiffness measurement technique that captures the tissue displacement in the phase of the signal. To limit the scanning time to a single breath-hold, liver MRE usually involves advanced readout techniques such as simultaneous multislice (SMS) or multishot methods. Furthermore, all these readout techniques require additional in-plane acceleration using either parallel imaging capabilities, such as sensitivity encoding (SENSE), or -space undersampling, such as compressed sensing (CS). However, these methods apply a single regularization function on the complex image. This study aims to design and evaluate methods that use separate regularization on the magnitude and phase of MRE to exploit their distinct spatiotemporal characteristics. Specifically, we introduce two compressed sensing methods. The first method, termed phase-regularized compressed sensing (PRCS), applies a two-dimensional total variation (TV) prior to the magnitude and two-dimensional wavelet regularization to the phase. The second method, termed displacement-regularized compressed sensing (DRCS), exploits the spatiotemporal redundancy using 3D total variation on the magnitude. Additionally, DRCS includes a displacement fitting function to apply wavelet regularization to the displacement phasor. Both DRCS and PRCS were evaluated with different levels of compression factors in three datasets: an in silico abdomen dataset, an in vitro tissue-mimicking phantom, and an in vivo liver dataset. The reconstructed images were compared with the full sampled reconstruction, zero-filling reconstruction, wavelet-regularized compressed sensing, and a low rank plus sparse reconstruction. The metrics used for quantitative evaluation were the structural similarity index (SSIM) of magnitude (M-SSIM), displacement (D-SSIM), and shear modulus (S-SSIM), and mean shear modulus. Results from highly undersampled in silico and in vitro datasets demonstrate that the DRCS method provides higher reconstruction quality than the conventional compressed sensing method for a wide range of stiffness values. Notably, DRCS provides 24% and 22% increase in D-SSIM compared with CS for the in silico and in vitro datasets, respectively. Comparison with liver stiffness measured from full sampled data and highly undersampled data (CR=4) demonstrates that the DRCS method provided the strongest correlation ( =0.95), second-lowest mean bias (-0.18 kPa, lowest for CS with -0.16 kPa), and lowest coefficient of variation (CV=3.6%). Our results demonstrate the potential of using DRCS to improve the reconstruction quality of accelerated MRE.

Topologically preserved registration of 3D CT images with deep networks

Objective. Computed Tomography (CT) image registration makes fast and accurate imaging-based disease diagnosis possible. We aim to develop a framework which can perform accurate local registration of organs in 3D CT images while preserving the topology of transformation. Approach. In this framework, the Faster R-CNN method is first used to detect local areas containing organs from fixed and moving images whose results are then registered with a weakly supervised deep neural network. In this network, a novel 3D channel coordinate attention (CA) module is introduced to reduce the loss of position information. The image edge loss and the organ labelling loss are used to weakly supervise the training process of our deep network, which enables the network learning to focus on registering organs and image structures. An intuitive inverse module is also used to reduce the folding of deformation field. More specifically, the folding is suppressed directly by simultaneously maximizing forward and backward registration accuracy in the image domain rather than indirectly by measuring the consistency of forward and inverse deformation fields as usual. Main results. Our method achieves an average dice similarity coefficient (DSC) of 0.954 and an average Similarity (Sim) of 0.914 on publicly available liver datasets (LiTS for training and Sliver07 for testing) and achieves an average DSC of 0.914 and an average Sim of 0.947 on our home-built left ventricular myocardium (LVM) dataset. Significance. Experimental results show that our proposed method can significantly improve the registration accuracy of organs such as the liver and LVM. Moreover, our inverse module can intuitively improve the inherent topological preservation of transformations.

Histopathological carcinoma classification using parallel, cross‐concatenated and grouped convolutions deep neural network

AbstractCancer is more alarming in modern days due to its identification at later stages. Among cancers, lung, liver and colon cancers are the leading cause of untimely death. Manual cancer identification from histopathological images is time‐consuming and labour‐intensive. Thereby, computer‐aided decision support systems are desired. A deep learning model is proposed in this paper to accurately identify cancer. Convolutional neural networks have shown great ability to identify the significant patterns for cancer classification. The proposed Parallel, Cross Concatenated and Grouped Convolutions Deep Neural Network (PC2GCDN2) has been developed to obtain accurate patterns for classification. To prove the robustness of the model, it is evaluated on the KMC and TCGA‐LIHC liver dataset, LC25000 dataset for lung and colon cancer classification. The proposed PC2GCDN2 model outperforms states‐of‐the‐art methods. The model provides 5.5% improved accuracy compared to the LiverNet proposed by Aatresh et. al on the KMC dataset. On the LC25000 dataset, 2% improvement is observed compared to existing models. Performance evaluation metrics like Sensitivity, Specificity, Recall, F1‐Score and Intersection‐Over‐Union are used to evaluate the performance. To the best of our knowledge, PC2GCDN2 can be considered as gold standard for multiple histopathology image classification. PC2GCDN is able to classify the KMC and TCGA‐LIHC liver dataset with 96.4% and 98.6% accuracy, respectively, which are the best results obtained till now. The performance has been superior on LC25000 dataset with 99.5% and 100% classification accuracy on lung and colon dataset, by utilizing less than 0.5 million parameters.

Regularized Ultrasound Phantom-Free Local Attenuation Coefficient Slope (ACS) Imaging in Homogeneous and Heterogeneous Tissues.

Attenuation maps or measurements based on the local attenuation coefficient slope (ACS) in quantitative ultrasound (QUS) have shown potential for the diagnosis of liver steatosis. In liver cancers, tissue abnormalities and tumors detected using ACS are also of interest to provide new image contrast to clinicians. Current phantom-based approaches have the limitation of assuming a comparable speed of sound between the reference phantom and insonified tissues. Moreover, these methods present the inconvenience for operators to acquire data on phantoms and patients. The main goal was to alleviate these drawbacks by proposing a methodology for constructing phantom-free regularized (PF-R) local ACS maps and investigate the performance in both homogeneous and heterogeneous media. The proposed method was tested on two tissue-mimicking media with different ACS constructed as homogeneous phantoms, side-by-side and top-to-bottom phantoms, and inclusion phantoms with different attenuations. Moreover, an in vivo proof-of-concept was performed on healthy, steatotic, and cancerous human liver datasets. Modifications brought to previous works include: 1) a linear interpolation of the power spectrum in the log scale; 2) the relaxation of the underlying hypothesis on the diffraction factor; 3) a generalization to nonhomogeneous local ACS; and 4) an adaptive restriction of frequencies to a more reliable range than the usable frequency range. Regularization was formulated as a generalized least absolute shrinkage and selection operator (LASSO), and a variant of the Bayesian information criterion (BIC) was applied to estimate the Lagrangian multiplier on the LASSO constraint. In addition, we evaluated the proposed algorithm when applying median filtering before and after regularization. Tests conducted showed that the PF-R yielded robust results in all tested conditions, suggesting potential for additional validation as a diagnosis method.

Early detection and classification of liver diseases in ultrasound images using hybrid machine learning techniques

Liver disease is a form of sickness that starts in the ovaries and is particularly dangerous for women. As a result, aberrant cells develop that have the potential to spread to other parts of the body. Liver disease is a sort of risky improvement that impacts ovaries in ladies, and is hard to perceive at early phase because of which it stays as one of the guideline wellsprings of illness end. Unquestionable confirmation of intrinsic and typical parts is immense in making novel frameworks to perceive and ruin danger. Here, using hybrid machine learning approaches, we propose an early detection and classification of liver disease (EDCLD) in ultrasound images. This study is aimed at detecting and evaluating liver disease, which may be benign or malignant depending on the cancer type. To begin with, we use an improved ant swarm optimization algorithm for both high and low level feature extraction models. For liver disease classification, we used hybrid swallow swarm intelligent-deep neural network (HSSI-DNN) classifier to identify the disease stage. By using training and testing liver datasets, we are improving the accuracy of liver disease detection with several stages. Second, we present a search optimization for whales algorithm for cascaded and segmentation as a single fusion feature capable of representing both the texture patterns and semantic background distributed in the image. Third, the dividend yields are commitment to the HSSI-DNN classifier used to detect the “Normal” and “Diseases” from input image. The proposed HSSI-DNN classifier performance outperformed the existing classifier in terms of accuracy (0.98), specificity (0.98), and sensitivity (1). The proposed EDCLD technique implemented in MATLAB.

Data enhancement based on M2-Unet for liver segmentation in Computed Tomography

Liver cancer is one of the diseases with the highest mortality in the world. Accurate segmentation of liver regions from liver CT data is a necessary preprocessing stage of computer-aided detection and diagnosis algorithms. In view of the small number of existing medical image data sets and difficulty in acquisition, this paper expanded the effective data set based on an improved conditional generation adversarial network pix2pix. We added random noise to the input of the generator, and the last layer of the output of the discriminator. Moreover, the original fully connected layer was replaced with a probability matrix with a size of 16 × 16. This is to make the discriminator more delicate. We merged the block module in MobileNet-v2 and U-Net (M2-Unet) to achieve liver segmentation. In addition to the training and validation using the public competition data set LiTs, we also constructed a new liver data set. The Dice similarity coefficient of the algorithm in this paper is 88.7%. The proposed method has improved to the ordinary U-Net algorithm. In the test phase, each image only needs 0.072 s to complete the segmentation on average, which exceeds the segmentation speed of specialists. Experimental results show that algorithm in this paper can accurately segment the liver and meet the requirements of realtime segmentation.

Development of Novel Residual-Dense-Attention (RDA) U-Net Network Architecture for Hepatocellular Carcinoma Segmentation.

The research was based on the image recognition technology of artificial intelligence, which is expected to assist physicians in making correct decisions through deep learning. The liver dataset used in this study was derived from the open source website (LiTS) and the data provided by the Kaohsiung Chang Gung Memorial Hospital. CT images were used for organ recognition and lesion segmentation; the proposed Residual-Dense-Attention (RDA) U-Net can achieve high accuracy without the use of contrast. In this study, U-Net neural network was used to combine ResBlock in ResNet with Dense Block in DenseNet in the coder part, allowing the training to maintain the parameters while reducing the overall recognition computation time. The decoder was equipped with Attention Gates to suppress the irrelevant areas of the image while focusing on the significant features. The RDA model was used to identify and segment liver organs and lesions from CT images of the abdominal cavity, and excellent segmentation was achieved for the liver located on the left side, right side, near the heart, and near the lower abdomen with other organs. Better recognition was also achieved for large, small, and single and multiple lesions. The study was able to reduce the overall computation time by about 28% compared to other convolutions, and the accuracy of liver and lesion segmentation reached 96% and 94.8%, with IoU values of 89.5% and 87%, and AVGDIST of 0.28 and 0.80, respectively.

MTL-ABS3Net: Atlas-Based Semi-Supervised Organ Segmentation Network With Multi-Task Learning for Medical Images.

Organ segmentation is one of the most important step for various medical image analysis tasks. Recently, semi-supervised learning (SSL) has attracted much attentions by reducing labeling cost. However, most of the existing SSLs neglected the prior shape and position information specialized in the medical images, leading to unsatisfactory localization and non-smooth of objects. In this paper, we propose a novel atlas-based semi-supervised segmentation network with multi-task learning for medical organs, named MTL-ABS3Net, which incorporates the anatomical priors and makes full use of unlabeled data in a self-training and multi-task learning manner. The MTL-ABS3Net consists of two components: an Atlas-Based Semi-Supervised Segmentation Network (ABS3Net) and Reconstruction-Assisted Module (RAM). Specifically, the ABS3Net improves the existing SSLs by utilizing atlas prior, which generates credible pseudo labels in a self-training manner; while the RAM further assists the segmentation network by capturing the anatomical structures from the original images in a multi-task learning manner. Better reconstruction quality is achieved by using MS-SSIM loss function, which further improves the segmentation accuracy. Experimental results from the liver and spleen datasets demonstrated that the performance of our method was significantly improved compared to existing state-of-the-art methods.

Abstract 3759: Recurrent methylation heterogeneity and reduced expression of SPG20 is a common characteristic of different tumors

Abstract Background: The Spastic Paraplegia-20 (SPG20, also known as SPART), is a gene that encodes spartin protein. When SPG20 is mutated, spartin expression is downregulated and leads to a rare autosomal recessive disorder called Troyer Syndrome. We recently demonstrated that this gene is epigenetically silenced and downregulated in Diffuse Large B Cell Lymphoma (DLBCL) cell lines. In recent years, other authors demonstrated independently the epigenetic regulation of SPG20 in hepatocellular, gastric and colorectal cancers. In order to assess whether the differential methylation of SPG20 is a common characteristic of tumors of different histological origin, we analysed the methylation of 9 probes upstream of the transcription starting site (TSS) and the gene expression of SPG20 in tumor samples affecting 6 different anatomical sites. Methods: Illumina Infinium Human Methylation 450 BeadChip and RNAseq data extrapolated from The Cancer Genome Atlas (TCGA) database were used to evaluate the methylation levels and expression of SPG20. Bladder, Colon, Kidney, Liver, Lung and Prostate datasets were selected. At least 30 independent tumor and at least 5 normal tissue samples for each tumor site were analysed. The tumor and normal tissues were paired in each specific cohort. Transcriptome profiling was analysed only in datasets where both tumor and normal counterparts were available. Statistical analysis was performed by two-way ANOVA with a Sidak’s post hoc test for methylation analysis and multiple t-test with Holm-Sidak’s post hoc test for expression analysis. Results: SPG20 β-values ranging from 0.19 to 0.96 among the different tumor samples were observed in the promoter region covered by the 9 probes that we previously identified (Illumina 450k platform). In the normal group, the β-values were higher than 0.63. Statistically significant differences between the tumor and normal counterpart were observed in colon, liver and prostate and mainly in the probes closest the TSS. The heterogeneity in the methylation status of SPG20 was a common characteristic among the different tumor groups. The variability range of β-values between tumor and normal groups was significantly different with a p value <0.05. Differential expression was observed between tumor and normal counterpart in almost all sites. SPG20 was more expressed in normal than in tumor tissues and the differences were significant in bladder, lung and colon. Conclusion: Our analysis unveils that the methylation heterogeneity of SPG20 is a common hallmark in all the tumor samples, while the normal counterpart results homogenously distributed. In tumor samples there is an inverse correlation between SPG20 promoter methylation heterogeneity and gene expression. SPG20 is down-regulated in almost all the tumors of different histological origins. SPG20 could have a relavant but still undefined role in cancer as a tumor-suppressor gene. Citation Format: Vincenza Ylenia Cusenza, Raffaele Frazzi. Recurrent methylation heterogeneity and reduced expression of SPG20 is a common characteristic of different tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3759.

A dual stage AlexNet-HHO-DrpXLM archetype for an effective feature extraction, classification and prediction of liver cancer based on histopathology images

Nowadays, more number of people affected by the liver cancer globally. Hepatocellular carcinoma (HCC) is one types of liver cancer that mainly affects the adults and it leads to death. Hepatocellular carcinoma disease diagnosis in an early stage is very necessary in the medical industry for reducing the death rate. However, the existing approaches have more difficulties to obtain the accurate results due to its high error pruning, time consuming and provided very low accuracy for classifying the cancer. So this work proposed innovative dual stage AlexNet- Horse Herd Optimization- Dropout Extreme Learning Machine (AlexNet-HHO-DrpXLM) architecture for classifying the histopathological haematoxylin and eosin (H&E) images. Initially, the developed dual stage model is utilized to classify the histopathological images. Moreover, the histopathology images are obtained from the standard datasets, like KMC liver dataset, which are trained to the AlexNet model. Subsequently, the Dropout Extreme Learning Machine (DrpXLM) classifier is used for classifying the Hepatocellular carcinoma images and the performance is improved by Horse Herd Optimization (HHO) mechanism. The performance of this work is compared with recent existing methods in terms of accuracy, precision, sensitivity, specificity and recall. Hence, the proposed method attains 21.43%, 27.50%, 22.89%, 26.42%, 3.58% and 2.07%higher accuracy values for malignant classification and 25.70%, 37.48%, 9.99%, 33.32%, 25.70% and 7.31% higher accuracy values for benign. The comparison results prove that the efficiency of the proposed AlexNet-HHO-DrpXLM method is more efficient than the other existing methods.

Sex-Specific Metabolic Effects of Dietary Folate Withdrawal in Wild-Type and Aldh1l1 Knockout Mice.

ALDH1L1 (10-formyltetrahydrofolate dehydrogenase), an enzyme of folate metabolism, is highly expressed in the liver. It regulates the overall flux of folate-bound one-carbon groups by converting 10-formyltetrahydrofolate to tetrahydrofolate and CO2 in a NADP+-dependent reaction. Our previous study revealed that Aldh1l1 knockout (KO) mice have an altered liver metabotype with metabolic symptoms of folate deficiency when fed a standard chow diet containing 2 ppm folic acid. Here we performed untargeted metabolomic analysis of liver and plasma of KO and wild-type (WT) male and female mice fed for 16 weeks either standard or folate-deficient diet. OPLS-DA, a supervised multivariate technique that was applied to 6595 and 10,678 features for the liver and plasma datasets, respectively, indicated that genotype and diet, alone or in combination, gave distinct metabolic profiles in both types of biospecimens. A more detailed analysis of affected metabolic pathways based on most confidently identified metabolites in the liver and plasma (OL1 and OL2a ontology level) indicated that the dietary folate restriction itself does not fully recapitulate the metabolic effect of the KO. Of note, dietary folate withdrawal enhanced the metabolic perturbations linked to the ALDH1L1 loss only for a subset of metabolites. Importantly, both the ALDH1L1 loss and dietary folate deficiency produced sex-specific metabolic effects.

Empowering multi-class medical data classification by Group-of-Single-Class-predictors and transfer optimization: Cases of structured dataset by machine learning and radiological images by deep learning

In the medical domain, data are often collected over time, evolving from simple to refined categories. The data and the underlying structures of the medical data as to how they have grown to today’s complexity can be decomposed into crude forms when data collection starts. For instance, the cancer dataset is labeled either benign or malignant at its simplest or perhaps the earliest form. As medical knowledge advances and/or more data become available, the dataset progresses from binary class to multi-class, having more labels of sub-categories of the disease added. In machine learning, inducing a multi-class model requires more computational power. Model optimization is enforced over the multi-class models for the highest possible accuracy, which of course, is necessary for life-and-death decision making. This model optimization task consumes an extremely long model training time. In this paper, a novel strategy called Group-of-Single-Class prediction (GOSC) coupled with majority voting and model transfer is proposed for achieving maximum accuracy by using only a fraction of the model training time. The main advantage is the ability to achieve an optimized multi-class classification model that has the highest possible accuracy near to the absolute maximum, while the training time could be saved by up to 70%. Experiments on machine learning over liver dataset classification and deep learning over COVID19 lung CT images were tested. Preliminary results suggest the feasibility of this new approach.

Addressing concomitant gradient phase errors in time-interleaved chemical shift-encoded MRI fat fraction and R2 * mapping with a pass-specific phase fitting method.

Concomitant gradients induce phase errors that increase quadratically with distance from isocenter. This work proposes a complex-based fitting method that addresses concomitant gradient phase errors in chemical shift encoded (CSE) MRI estimation of proton density fat fraction (PDFF) and R2 * through joint estimation of pass-specific phase terms. This method is applicable to time-interleaved multi-echo gradient-echo acquisitions (i.e., multi-pass acquisitions) and does not require prior knowledge of gradient waveforms typically needed to address concomitant gradient phase errors. A CSE-MRI spoiled gradient echo signal model, with pass-specific phase terms, is introduced for non-linear least squares estimation of PDFF and R2 * in the presence of concomitant gradient phase errors. Cramér-Rao lower bound analysis was used to determine noise performance tradeoffs of the proposed fitting method, which was then validated in both phantom and in vivo experiments. The proposed fitting method removed PDFF and R2 * estimation errors up to 12% and 10 s-1 , respectively, at ±12 cm off isocenter (S/I) in a water phantom. In healthy volunteers, PDFF and R2 * bias was reduced by ~10% (12 cm off-isocenter) and ~30 s-1 (16 cm off-isocenter), respectively. An evaluation in 29 clinical liver datasets demonstrated reduced PDFF bias and variability (8.4% improvement in the coefficient of variation), even with the imaging volume centered at isocenter. Concomitant gradient induced phase errors in multi-pass CSE-MRI acquisitions can result in PDFF and R2 * estimation biases away from isocenter. The proposed fitting method enables accurate PDFF and R2 * quantification in the presence of concomitant gradient phase errors without knowledge of imaging gradient waveforms.

ELU-Net: An Efficient and Lightweight U-Net for Medical Image Segmentation

Recent years have witnessed a growing interest in the use of U-Net and its improvement. It is one of the classic semantic segmentation networks with an encoder-decoder architecture and is widely used in medical image segmentation. In the series versions of U-Net, U-Net++ has been developed as an improved U-Net by designing an architecture with nested and dense skip connections, and U-Net 3+ has been developed as an improved U-Net++ by taking advantage of full-scale skip connections and deep supervision on full-scale aggregated feature maps. Each network architecture has its own advantages in the use of the encoder and decoder. In this paper, we propose an efficient and lightweight U-Net (ELU-Net) with deep skip connections. The deep skip connections include same- and large-scale skip connections from the encoder to fully extract the features of the encoder. In addition, the proposed ELU-Net with different loss functions is discussed to improve the effect of brain tumor learning including WT (whole tumor), TC (tumor core) and ET (enhance tumor) and a new loss function DFK is designed. The effectiveness of the proposed method is demonstrated for a brain tumor dataset used in the BraTS 2018 Challenge and liver dataset used in the ISBI LiTS 2017 Challenge.

Gα12 overexpression in hepatocytes by ER stress exacerbates acute liver injury via ROCK1-mediated miR-15a and ALOX12 dysregulation.

Rationale: Liver injury must be further characterized to identify novel therapeutic approaches. Endoplasmic reticulum (ER) stress may cause hepatocyte death. Gα12 affects cell viability and its expression varies depending on physiological conditions. This study investigated whether hepatocyte-specific Gα12 overexpression affects acute liver injury, and if so, what the underlying mechanisms and treatment strategies are.Methods: All experiments were performed using human liver, hepatocytes, and toxicant injury models with Gna12 KO and/or hepatocyte-specific Gα12 overexpression. RNA-sequencing, immunoblotting, immunohistochemistry, reporter assays, and mutation assays were conducted.Results: Hepatic Gα12 was overexpressed in mice challenged with acetaminophen or other ER stress inducers or in patients with acute liver injury or fibrosis/cirrhosis. Several Gα12 and ER-associated pathways were identified using transcriptomic analysis. Acetaminophen intoxication was characterized by lipid peroxide-induced ferroptosis and was less severe in Gα12-deficient animals and cells. Conversely, Gα12 overexpression in wild-type or Gna12 KO hepatocytes increased hepatotoxicity, promoting lipid peroxidation, inflammation, and ferroptosis. IRE1α-dependent Xbp1 transactivated Gna12. Moreover, Gα12 overexpression enhanced the ability of acetaminophen to induce ALOX12, while downregulating GPX4. The level of miR-15a, herein identified as an ALOX12 inhibitor, was decreased. siRNA knockdown or pharmacological inhibition of ROCK1 prevented dysregulation of ALOX12 and GPX4, rescuing animals from toxicant-induced ferroptosis. These changes or correlations among the targets were confirmed in human liver specimens and datasets of livers exposed to other injurious medications.Conclusions: Gα12 overexpression by ER stress facilitates hepatocyte ferroptosis through ROCK1-mediated dysregulation of ALOX12, and miR-15a, supporting the concept that inhibition of Gα12 overexpression and/or ROCK1 axis may constitute a promising strategy for acute liver injury.