Segmentation of MR Brain Images Through Hidden Markov Random Field and Hybrid Metaheuristic Algorithm

Predicting the risk category of thymoma with machine learning-based computed tomography radiomics signatures and their between-imaging phase differences

The aim of this study was to develop a medical imaging and comprehensive stacked learning-based method for predicting high- and low-risk thymoma. A total of 126 patients with thymomas and 5 patients with thymic carcinoma treated at our institution, including 65 low-risk patients and 66 high-risk patients, were retrospectively recruited. Among them, 78 patients composed the training cohort, while the remaining 53 patients formed the validation cohort. We extracted 1702 features each from the patients' arterial-, venous-, and plain-phase images. Pairwise subtraction of these features yielded 1702 arterial-venous, arterial-plain, and venous-plain difference features each. The Mann‒Whitney U test and least absolute shrinkage and selection operator (LASSO) and SelectKBest methods were employed to select the best features from the training set. Six models were built with a stacked learning algorithm. By applying stacked ensemble learning, three machine learning algorithms (XGBoost, multilayer perceptron (MLP), and random forest) were combined by XGBoost to produce the the six basic imaging models. Then, the XGBoost algorithm was applied to the six basic imaging models to construct a combined radiomic model. Finally, the radiomic model was combined with clinical information to create a nomogram that could easily be used in clinical practice to predict the thymoma risk category. The areas under the curve (AUCs) of the combined radiomic model in the training and validation cohorts were 0.999 (95% CI 0.988-1.000) and 0.967 (95% CI 0.916-1.000), respectively, while those of the nomogram were 0.999 (95% CI 0.996-1.000) and 0.983 (95% CI 0.990-1.000). This study describes the application of CT-based radiomics in thymoma patients and proposes a nomogram for predicting the risk category for this disease, which could be advantageous for clinical decision-making for affected patients.

3D balanced SSFP UTE MRI for multiple contrasts whole brain imaging

PURPOSE

This study aimed to develop a new high-resolution MRI sequence for the imaging of the ultra-short transverse relaxation time (uT2) components in the brain, while simultaneously providing proton density (PD) contrast for reference and quantification.

THEORY

The sequence combines low flip angle balanced SSFP (bSSFP) and UTE techniques, together with a 3D dual-echo rosette k-space trajectory for readout.

METHODS

The expected image contrast was evaluated by simulations. A study cohort of six healthy volunteers and eight multiple sclerosis (MS) patients was recruited to test the proposed sequence. Subtraction between two TEs was performed to extract uT2 signals. In addition, conventional longitudinal relaxation time (T1) weighted, T2-weighted, and PD-weighted MRI sequences were also acquired for comparison.

RESULTS

Typical PD-contrast was found in the second TE images, while uT2 signals were selectively captured in the first TE images. The subtraction images presented signals primarily originating from uT2 components, but only if the first TE is short enough. Lesions in the MS subjects showed hyperintense signals in the second TE images but were hypointense signals in the subtraction images. The lesions had significantly lower signal intensity in subtraction images than normal white matter (WM), which indicated a reduction of uT2 components likely associated with myelin.

CONCLUSION

3D isotropic sub-millimeter (0.94 mm) spatial resolution images were acquired with the novel bSSFP UTE sequence within 3 min. It provided easy extraction of uT2 signals and PD-contrast for reference within a single acquisition.

Multi-Objective Unsupervised Band Selection Method for Hyperspectral Images Classification

With the increasing spectral dimension of hyperspectral images (HSI), how correctly choose bands based on band correlation and information has become more significant, but also complicated. Band selection is a combinatorial optimization problem, and intelligent optimization algorithms have been shown to be crucial in solving combinatorial optimization problems. However, major of them only use a single objective as the selection index, while neglecting the overall features of hyperspectral images, which may lead to inaccuracy in object detection. To tackle this, we propose a band selection method based on a multi-objective cuckoo search algorithm (MOCS) when constructing a multi-objective unsupervised band selection model based on the amount of information and correlation of the bands (MOCS-BS). Specifically, an adaptive strategy based on population crowding degree is first proposed to assist Lévy flight in overcoming the influence of the parameter constancy. Then, an information-sharing strategy based on grouping and crossover is designed to balance the search ability between global exploration and local exploitation, which can overcome the shortcomings caused by the lack of information interaction between individuals. Finally, the HSI classification experiments are performed by Random Forest and KNN classifiers based on the subset of bands selected by the proposed MOCS-BS method. The proposed method is compared with state-of-the-art algorithms including neighborhood grouping normalized matched filter (NGNMF) and multi-objective artificial bee colony with band selection (MABC-BS) on four HSI datasets. The experimental results demonstrate that MOCS-BS is more effective and robust than other methods.

Metabolic dysfunction-associated fatty liver disease directly related to liver fibrosis independent of insulin resistance, hyperlipidemia, and alcohol intake in morbidly obese patients

AIM

Hepatic fibrosis is associated with various factors, including metabolic dysfunction-associated fatty liver disease (MAFLD), insulin resistance, and alcohol intake in patients with morbid obesity. We investigated factors directly associated with hepatic fibrosis in patients with morbid obesity using a graphical model.

METHODS

We enrolled 134 consecutive patients with morbid obesity who underwent liver biopsy during sleeve gastrectomy (median age 43.5 years; MAFLD 78.4%; homeostasis model assessment of insulin resistance [HOMA-IR] 5.97; >20 g/day alcohol intake 14.2%). Patients were classified into none/mild (F0/1; n = 77) or significant/advanced fibrosis (F2/3; n = 57) groups, based on histology. Factors associated with F2/3 were analyzed using logistic regression analysis and a graphical model.

RESULTS

F2/3 was observed in 42.5% of the enrolled patients. The prevalence of MAFLD and HOMA-IR values were significantly higher in the F2/3 group than in the F0/1 group; however, no significant difference in alcohol intake was observed between the two groups. On logistic regression analysis, MAFLD, but not HOMA-IR or alcohol intake, was the only independent factor associated with F2/3 (odds ratio 7.555; 95% confidence interval 2.235-25.544; p = 0.0011). The graphical model revealed that F2/3 directly interacted with MAFLD, diabetes mellitus, HOMA-IR, and low-density lipoprotein cholesterol. Among these factors, MAFLD showed the strongest interaction with F2/3.

CONCLUSIONS

We determined that MAFLD was more directly associated with significant/advanced fibrosis than insulin resistance or hyperlipidemia, and alcohol intake was not directly associated with hepatic fibrosis. Metabolic dysfunction-associated fatty liver disease could be the most important factor for hepatic fibrosis in patients with morbid obesity.

Brain Image Segmentation in Recent Years: A Narrative Review

Brain image segmentation is one of the most time-consuming and challenging procedures in a clinical environment. Recently, a drastic increase in the number of brain disorders has been noted. This has indirectly led to an increased demand for automated brain segmentation solutions to assist medical experts in early diagnosis and treatment interventions. This paper aims to present a critical review of the recent trend in segmentation and classification methods for brain magnetic resonance images. Various segmentation methods ranging from simple intensity-based to high-level segmentation approaches such as machine learning, metaheuristic, deep learning, and hybridization are included in the present review. Common issues, advantages, and disadvantages of brain image segmentation methods are also discussed to provide a better understanding of the strengths and limitations of existing methods. From this review, it is found that deep learning-based and hybrid-based metaheuristic approaches are more efficient for the reliable segmentation of brain tumors. However, these methods fall behind in terms of computation and memory complexity.

An optimized initialization for LDPC decoding over GF(q) in impulsive noise environments

Modern navigation satellite communication has the characteristic of high transmitting rate. To avoid bit errors in data transmission, low density parity check (LDPC) codes are widely recognized as efficient ways for navigation communication. Conventionally, the LDPC decoding is applied for additive white Gaussian noise (AWGN) channel and degrades severely while facing the impulsive noise. However, navigation communication often suffers from impulsive interference due to the occurrence of high amplitude “spikes”. At this time, the conventional Gaussian noise assumption is inadequate. The impulsive component of interference has been found to be significant which influences the reliability of transmitted information. Therefore the LDPC decoding algorithms for AWGN channel are not suitable for impulsive noise environments. Consider that LDPC codes over GF(q) perform better than binary LDPC in resisting burst errors for current navigation system, it is necessary to conduct research on LDPC codes over GF(q). In this paper, an optimized initialization by calculating posterior probabilities of received symbols is proposed for non-binary LDPC decoding on additive white Class A noise (AWAN) channel. To verify the performance of the proposed initialization, extensive experiments are performed in terms of convergence, validity, and robustness. Preliminary results demonstrate that the decoding algorithm with the optimized initialization for non-binary LDPC codes performs better than the competing methods and that of binary LDPC codes on AWAN channel.

Discriminative Feature Learning for Thorax Disease Classification in Chest X-ray Images

This paper focuses on the thorax disease classification problem in chest X-ray (CXR) images. Different from the generic image classification task, a robust and stable CXR image analysis system should consider the unique characteristics of CXR images. Particularly, it should be able to: 1) automatically focus on the disease-critical regions, which usually are of small sizes; 2) adaptively capture the intrinsic relationships among different disease features and utilize them to boost the multi-label disease recognition rates jointly. In this paper, we propose to learn discriminative features with a two-branch architecture, named ConsultNet, to achieve those two purposes simultaneously. ConsultNet consists of two components. First, an information bottleneck constrained feature selector extracts critical disease-specific features according to the feature importance. Second, a spatial-and-channel encoding based feature integrator enhances the latent semantic dependencies in the feature space. ConsultNet fuses these discriminative features to improve the performance of thorax disease classification in CXRs. Experiments conducted on the ChestX-ray14 and CheXpert dataset demonstrate the effectiveness of the proposed method.