Individualized Texture Similarity Network in Schizophrenia

BACKGROUND

Structural covariance network disruption has been considered an important pathophysiological indicator for schizophrenia. Here, we introduced a novel individualized structural covariance network measure, referred to as a texture similarity network (TSN), and hypothesized that the TSN could reliably reveal unique intersubject heterogeneity and complex dysconnectivity patterns in schizophrenia.

METHODS

The TSN was constructed by measuring the covariance of 180 three-dimensional voxelwise gray-level co-occurrence matrix feature maps between brain areas in each participant. We first tested the validity and reproducibility of the TSN in characterizing the intersubject variability in 2 longitudinal test-retest healthy cohorts. The TSN was further applied to elucidate intersubject variability and dysconnectivity patterns in 10 schizophrenia case-control datasets (609 schizophrenia cases vs. 579 controls) as well as in a first-episode depression dataset (69 patients with depression vs. 69 control participants).

RESULTS

The test-retest analysis demonstrated higher TSN intersubject than intrasubject variability. Moreover, the TSN reliably revealed higher intersubject variability in both chronic and first-episode schizophrenia, but not in depression. The TSN also reproducibly detected coexistent increased and decreased TSN strength in widespread brain areas, increased global small-worldness, and the coexistence of both structural hyposynchronization in the central networks and hypersynchronization in peripheral networks in patients with schizophrenia but not in patients with depression. Finally, aberrant intersubject variability and covariance strength patterns revealed by the TSN showed a missing or weak correlation with other individualized structural covariance network measures, functional connectivity, and regional volume changes.

CONCLUSIONS

These findings support the reliability of a TSN in revealing unique structural heterogeneity and complex dysconnectivity in patients with schizophrenia.

Development and validation of a postoperative pulmonary infection prediction model for patients with primary hepatic carcinoma

BACKGROUND

There are factors that significantly increase the risk of postoperative pulmonary infections in patients with primary hepatic carcinoma (PHC). Previous reports have shown that over 10% of patients with PHC experience postoperative pulmonary infections. Thus, it is crucial to prioritize the prevention and treatment of postoperative pulmonary infections in patients with PHC.

AIM

To identify the risk factors for postoperative pulmonary infection in patients with PHC and develop a prediction model to aid in postoperative management.

METHODS

We retrospectively collected data from 505 patients who underwent hepatobiliary surgery between January 2015 and February 2023 in the Department of Hepatobiliary and Pancreaticospleen Surgery. Radiomics data were selected for statistical analysis, and clinical pathological parameters and imaging data were included in the screening database as candidate predictive variables. We then developed a pulmonary infection prediction model using three different models: An artificial neural network model; a random forest model; and a generalized linear regression model. Finally, we evaluated the accuracy and robustness of the prediction model using the receiver operating characteristic curve and decision curve analyses.

RESULTS

Among the 505 patients, 86 developed a postoperative pulmonary infection, resulting in an incidence rate of 17.03%. Based on the gray-level co-occurrence matrix, we identified 14 categories of radiomic data for variable screening of pulmonary infection prediction models. Among these, energy, contrast, the sum of squares (SOS), the inverse difference (IND), mean sum (MES), sum variance (SUV), sum entropy (SUE), and entropy were independent risk factors for pulmonary infection after hepatectomy and were listed as candidate variables of machine learning prediction models. The random forest model algorithm, in combination with IND, SOS, MES, SUE, SUV, and entropy, demonstrated the highest prediction efficiency in both the training and internal verification sets, with areas under the curve of 0.823 and 0.801 and a 95% confidence interval of 0.766-0.880 and 0.744-0.858, respectively. The other two types of prediction models had prediction efficiencies between areas under the curve of 0.734 and 0.815 and 95% confidence intervals of 0.677-0.791 and 0.766-0.864, respectively.

CONCLUSION

Postoperative pulmonary infection in patients undergoing hepatectomy may be related to risk factors such as IND, SOS, MES, SUE, SUV, energy, and entropy. The prediction model in this study based on diffusion-weighted images, especially the random forest model algorithm, can better predict and estimate the risk of pulmonary infection in patients undergoing hepatectomy, providing valuable guidance for postoperative management.

Machine learning-based gray-level co-occurrence matrix signature for predicting lymph node metastasis in undifferentiated-type early gastric cancer

BACKGROUND

The most important consideration in determining treatment strategies for undifferentiated early gastric cancer (UEGC) is the risk of lymph node metastasis (LNM). Therefore, identifying a potential biomarker that predicts LNM is quite useful in determining treatment.

AIM

To develop a machine learning (ML)-based integral procedure to construct the LNM gray-level co-occurrence matrix (GLCM) prediction model.

METHODS

We retrospectively selected 526 cases of UEGC confirmed through pathological examination after radical gastrectomy without endoscopic treatment in four tertiary hospitals between January 2015 to December 2021. We extracted GLCM-based features from grayscale images and applied ML to the classification of candidate predictive variables. The robustness and clinical utility of each model were evaluated based on the following factors: Receiver operating characteristic curve (ROC), decision curve analysis, and clinical impact curve.

RESULTS

GLCM-based feature extraction significantly correlated with LNM. The top 7 GLCM-based factors included inertia value 0° (IV_0), inertia value 45° (IV_45), inverse gap 0° (IG_0), inverse gap 45° (IG_45), inverse gap full angle (IG_all), Haralick 30° (Haralick_30), Haralick full angle (Haralick_all), and Entropy. The areas under the ROC curve (AUCs) of the random forest classifier (RFC) model, support vector machine, eXtreme gradient boosting, artificial neural network, and decision tree ranged from 0.805 [95% confidence interval (CI): 0.258-1.352] to 0.925 (95%CI: 0.378-1.472) in the training set and from 0.794 (95%CI: 0.237-1.351) to 0.912 (95%CI: 0.355-1.469) in the testing set, respectively. The RFC (training set: AUC: 0.925, 95%CI: 0.378-1.472; testing set: AUC: 0.912, 95%CI: 0.355-1.469) model that incorporates Entropy, Haralick_all, Haralick_30, IG_all, IG_45, IG_0, and IV_45 had the highest predictive accuracy.

CONCLUSION

The evaluation results indicate that the method of selecting radiological and textural features becomes more effective in the LNM discrimination against UEGC patients. Additionally, the ML-based prediction model developed using the RFC can be used to derive treatment options and identify LNM, which can hence improve clinical outcomes.

Quantitative Indicators of Retraction Phenomenon on an Automated Breast Volume Scanner: Initial Study in the Diagnosis and Prognostic Prediction of Breast Tumors

Retraction phenomenon is a unique sign on an automated breast volume scanner coronal plane image and has high specificity in differentiating benign lesions from malignant breast cancer. The purpose of this study was to quantify the retraction phenomenon by setting different rules to describe connected regions from different dimensions. In total, six quantitative indicators (F_Ω1,F_Γ,F_S,F_Ω2,F_Ω3and F_L) were obtained. F_Ω1, F_Ω2 and F_Ω3 represent the relative areas of the connected region under different rules. F_ΓandF_S represent the number ratio and absolute area of the connected region, respectively. F_L represents the ratio of edge numbers. Two hundred fourteen patients with 214 lesions (90 benign and 124 malignant) were enrolled in this study. All quantitative indicators in the malignant group were significantly higher than those in the benign group (all p values <0.001). The indicator F_Γ achieved the highest area under the receiver operating characteristic curve (AUC) (0.701, 95% confidence interval: 0.631-0.771). Both F_Γ and F_S had significant associations with axillary lymph node metastasis (p = 0.023 and 0.049). Compared with the classic texture feature gray-level co-occurrence matrix, retraction phenomenon quantization improved the AUC by 8.3%. The results indicate that retraction phenomenon quantitative indicators have certain value in distinguishing benign and malignant breast lesions and seem to be associated with axillary lymph node metastasis.

Adjacent cartilage tissue structure after successful transplantation: a quantitative MRI study using T2 mapping and texture analysis

OBJECTIVES

The aim of this study was to assess the texture of repair tissue and tissue adjacent to the repair site after matrix-associated chondrocyte transplantation (MACT) of the knee using gray-level co-occurrence matrix (GLCM) texture analysis of T₂ quantitative maps.

METHODS

Twenty patients derived from the MRI sub-study of multicenter, single-arm phase III study underwent examination on a 3 T MR scanner, including a T₂ mapping sequence 12 and 24 months after MACT. Changes between the time points in mean T₂ values and 20 GLCM features were assessed for repair tissue, adjacent tissue, and reference cartilage. Differences in T₂ values and selected GLCM features between the three cartilage sites at two time points were analyzed using linear mixed-effect models.

RESULTS

A significant decrease in T₂ values after MACT, between time points, was observed only in repair cartilage (p < 0.001). Models showed significant differences in GLCM features between repair tissue and reference cartilage, namely, autocorrelation (p < 0.001), correlation (p = 0.015), homogeneity (p = 0.002), contrast (p < 0.001), and difference entropy (p = 0.047). The effect of time was significant in a majority of models with regard to GLCM features (except autocorrelation) (p ≤ 0.001). Values in repair and adjacent tissue became similar to reference tissue over time.

CONCLUSIONS

GLCM is a useful add-on to T₂ mapping in the evaluation of knee cartilage after MACT by increasing the sensitivity to changes in cartilage structure. The results suggest that cartilage tissue adjacent to the repair site heals along with the cartilage implant.

KEY POINTS

• GLCM is a useful add-on to T₂ mapping in the evaluation of knee cartilage after MACT by increasing the sensitivity to changes in cartilage structure. • Repair and adjacent tissue became similar to reference tissue over time. • The results suggest that cartilage tissue adjacent to the repair site heals along with the cartilage implant.

Combined Features in Region of Interest for Brain Tumor Segmentation

Diagnosis of brain tumor gliomas is a challenging task in medical image analysis due to its complexity, the less regularity of tumor structures, and the diversity of tissue textures and shapes. Semantic segmentation approaches using deep learning have consistently outperformed the previous methods in this challenging task. However, deep learning is insufficient to provide the required local features related to tissue texture changes due to tumor growth. This paper designs a hybrid method arising from this need, which incorporates machine-learned and hand-crafted features. A semantic segmentation network (SegNet) is used to generate the machine-learned features, while the grey-level co-occurrence matrix (GLCM)-based texture features construct the hand-crafted features. In addition, the proposed approach only takes the region of interest (ROI), which represents the extension of the complete tumor structure, as input, and suppresses the intensity of other irrelevant area. A decision tree (DT) is used to classify the pixels of ROI MRI images into different parts of tumors, i.e. edema, necrosis and enhanced tumor. The method was evaluated on BRATS 2017 dataset. The results demonstrate that the proposed model provides promising segmentation in brain tumor structure. The F-measures for automatic brain tumor segmentation against ground truth are 0.98, 0.75 and 0.69 for whole tumor, core and enhanced tumor, respectively.

Glaucoma Detection from Retinal Images Using Statistical and Textural Wavelet Features

Glaucoma is a silent progressive eye disease that is among the leading causes of irreversible blindness. Early detection and proper treatment of glaucoma can limit severe vision impairments associated with advanced stages of the disease. Periodic automatic screening can help in the early detection of glaucoma while reducing the workload on expert ophthalmologists. In this work, a wavelet-based glaucoma detection algorithm is proposed for real-time screening systems. A combination of wavelet-based statistical and textural features computed from the detected optic disc region is used to determine whether a retinal image is healthy or glaucomatous. Two public datasets having different resolutions were considered in the performance analysis of the proposed algorithm. An accuracy of 96.7% and area under receiver operating curve (AUC) of 94.7% were achieved for the high-resolution dataset. Analysis of the wavelet-based statistical and textural features using three different methods showed their relevance for glaucoma detection. Furthermore, the proposed algorithm is shown to be suitable for real-time applications as it requires less than 3 s for processing the high-resolution retinal images.

Prediction of plant-level tomato biomass and yield using machine learning with unmanned aerial vehicle imagery

BACKGROUND

The objective of this study is twofold. First, ascertain the important variables that predict tomato yields from plant height (PH) and vegetation index (VI) maps. The maps were derived from images taken by unmanned aerial vehicles (UAVs). Second, examine the accuracy of predictions of tomato fresh shoot masses (SM), fruit weights (FW), and the number of fruits (FN) from multiple machine learning algorithms using selected variable sets. To realize our objective, ultra-high-resolution RGB and multispectral images were collected by a UAV on ten days in 2020's tomato growing season. From these images, 756 total variables, including first- (e.g., average, standard deviation, skewness, range, and maximum) and second-order (e.g., gray-level co-occurrence matrix features and growth rates of PH and VIs) statistics for each plant, were extracted. Several selection algorithms (i.e., Boruta, DALEX, genetic algorithm, least absolute shrinkage and selection operator, and recursive feature elimination) were used to select the variable sets useful for predicting SM, FW, and FN. Random forests, ridge regressions, and support vector machines were used to predict the yield using the top five selected variable sets.

RESULTS

First-order statistics of PH and VIs collected during the early to mid-fruit formation periods, about one month prior to harvest, were important variables for predicting SM. Similar to the case for SM, variables collected approximately one month prior to harvest were important for predicting FW and FN. Furthermore, variables related to PH were unimportant for prediction. Compared with predictions obtained using only first-order statistics, those obtained using the second-order statistics of VIs were more accurate for FW and FN. The prediction accuracy of SM, FW, and FN by models constructed from all variables (rRMSE = 8.8-28.1%) was better than that from first-order statistics (rRMSE = 10.0-50.1%).

CONCLUSIONS

In addition to basic statistics (e.g., average and standard deviation), we derived second-order statistics of PH and VIs at the plant level using the ultra-high resolution UAV images. Our findings indicated that our variable selection method reduced the number variables needed for tomato yield prediction, improving the efficiency of phenotypic data collection and assisting with the selection of high-yield lines within breeding programs.

Prediction of plant-level tomato biomass and yield using machine learning with unmanned aerial vehicle imagery

BACKGROUND

The objective of this study is twofold. First, ascertain the important variables that predict tomato yields from plant height (PH) and vegetation index (VI) maps. The maps were derived from images taken by unmanned aerial vehicles (UAVs). Second, examine the accuracy of predictions of tomato fresh shoot masses (SM), fruit weights (FW), and the number of fruits (FN) from multiple machine learning algorithms using selected variable sets. To realize our objective, ultra-high-resolution RGB and multispectral images were collected by a UAV on ten days in 2020's tomato growing season. From these images, 756 total variables, including first- (e.g., average, standard deviation, skewness, range, and maximum) and second-order (e.g., gray-level co-occurrence matrix features and growth rates of PH and VIs) statistics for each plant, were extracted. Several selection algorithms (i.e., Boruta, DALEX, genetic algorithm, least absolute shrinkage and selection operator, and recursive feature elimination) were used to select the variable sets useful for predicting SM, FW, and FN. Random forests, ridge regressions, and support vector machines were used to predict the yield using the top five selected variable sets.

RESULTS

First-order statistics of PH and VIs collected during the early to mid-fruit formation periods, about one month prior to harvest, were important variables for predicting SM. Similar to the case for SM, variables collected approximately one month prior to harvest were important for predicting FW and FN. Furthermore, variables related to PH were unimportant for prediction. Compared with predictions obtained using only first-order statistics, those obtained using the second-order statistics of VIs were more accurate for FW and FN. The prediction accuracy of SM, FW, and FN by models constructed from all variables (rRMSE = 8.8-28.1%) was better than that from first-order statistics (rRMSE = 10.0-50.1%).

CONCLUSIONS

In addition to basic statistics (e.g., average and standard deviation), we derived second-order statistics of PH and VIs at the plant level using the ultra-high resolution UAV images. Our findings indicated that our variable selection method reduced the number variables needed for tomato yield prediction, improving the efficiency of phenotypic data collection and assisting with the selection of high-yield lines within breeding programs.

Magnetic Resonance Imaging Texture of Medial Pulvinar in Dementia with Lewy Bodies

INTRODUCTION

Executive dysfunction is common in dementia with Lewy bodies (DLB). The pulvinar nucleus plays a role in executive control and synchronizes with cortical regions in the salience network that are vulnerable to Lewy pathology.

OBJECTIVE

We investigated the pulvinar subregions in patients with mild DLB and their associations with executive function.

METHODS

The sample consisted of 38 DLB patients and 38 age- and sex-matched normal controls. We evaluated cognitive function using the Consortium to Establish a Registry for Alzheimer's Disease Assessment Packet. We obtained four pulvinar nuclei using preprocessed T1-weighted magnetic resonance images. We compared volumes and textures of the DLB patients and the normal controls for each nucleus. We used a linear regression to determine the association of textures and neuropsychological test scores.

RESULTS

The DLB patients showed comparable volumes to the normal controls in all pulvinar nuclei. However, the DLB patients showed different texture of the left medial pulvinar (PuM) from the normal controls. The entropy, contrast, and cluster shade were lower but autocorrelation of left PuM was higher in the DLB patients compared to the normal controls. These texture features of the left PuM were associated with the set-shifting performance measured by the Trail Making Test.

CONCLUSIONS

In DLB, the left PuM may be altered from early stage, which may contribute to the development of executive dysfunction.

Nipple Localization in Automated Whole Breast Ultrasound Coronal Scans Using Ensemble Learning

Nipple is a vital landmark in the breast lesion diagnosis. Although there are advanced computer-aided detection (CADe) systems for nipple detection in breast mediolateral oblique (MLO) views of mammogram images, few academic works address the coronal views of breast ultrasound (BUS) images. This paper addresses a novel CADe system to locate the Nipple Shadow Area (NSA) in ultrasound images. Here the Hu Moments and Gray-level Co-occurrence Matrix (GLCM) were calculated through an iterative sliding window for the extraction of shape and texture features. These features are then concatenated and fed into an Artificial Neural Network (ANN) to obtain probable NSA's. Later, contour features, such as shape complexity through fractal dimension, edge distance from the periphery and contour area, were computed and passed into a Support Vector Machine (SVM) to identify the accurate NSA in each case. The coronal plane BUS dataset is built upon our own, which consists of 64 images from 13 patients. The test results show that the proposed CADe system achieves 91.99% accuracy, 97.55% specificity, 82.46% sensitivity and 88% F-score on our dataset.

Discrimination of Low-Energy Acetabular Fractures from Controls Using Computed Tomography-Based Bone Characteristics

The incidence of low-energy acetabular fractures has increased. However, the structural factors for these fractures remain unclear. The objective of this study was to extract trabecular bone architecture and proximal femur geometry (PFG) measures from clinical computed tomography (CT) images to (1) identify possible structural risk factors of acetabular fractures, and (2) to discriminate fracture cases from controls using machine learning methods. CT images of 107 acetabular fracture subjects (25 females, 82 males) and 107 age-gender matched controls were examined. Three volumes of interest, one at the acetabulum and two at the femoral head, were extracted to calculate bone volume fraction (BV/TV), gray-level co-occurrence matrix and histogram of the gray values (GV). The PFG was defined by neck shaft angle and femoral neck axis length. Relationships between the variables were assessed by statistical mean comparisons and correlation analyses. Bayesian logistic regression and Elastic net machine learning models were implemented for classification. We found lower BV/TV at the femoral head (0.51 vs. 0.55, p = 0.012) and lower mean GV at both the acetabulum (98.81 vs. 115.33, p < 0.001) and femoral head (150.63 vs. 163.47, p = 0.005) of fracture subjects when compared to their matched controls. The trabeculae within the femoral heads of the acetabular fracture sides differed in structure, density and texture from the corresponding control sides of the fracture subjects. Moreover, the PFG and trabecular architectural variables, alone and in combination, were able to discriminate fracture cases from controls (area under the receiver operating characteristics curve 0.70 to 0.79). In conclusion, lower density in the acetabulum and femoral head with abnormal trabecular structure and texture at the femoral head, appear to be risk factors for low-energy acetabular fractures.

Combination of spectra and texture data of hyperspectral imaging for prediction and visualization of palmitic acid and oleic acid contents in lamb meat

The feasibility of combining spectral and textural information from hyperspectral imaging to improve the prediction of the C16:0 and C18:1 n9 contents for lamb was explored. 29 and 22 optimal wavelengths were selected for the C16:0 and C18:1 n9 contents, respectively, by conducting the variable combination population analysis-iteratively retaining informative variables (VCPA-IRIV) algorithm. To extract the textural features of images, a gray-level co-occurrence matrix (GLCM) analysis was implemented in the first principal component image. The least squares support vector machine (LSSVM) model and the partial least squares regression (PLSR) model were developed to predict the C16:0 and C18:1 n9 contents from the spectra and the fusion data. The distribution map was visualized using the best model with the imaging process. The results showed that the combination of the spectral and textural information of hyperspectral imaging coupled with the VCPA-IRIV algorithm had strong potential for the prediction and visualization of the C16:0 and C18:1 n9 contents of lamb.

An Analysis of Human Dorsal Hand Skin Texture Using Hyperspectral Imaging Technique for Assessing the Skin Aging Process

Skin texture has become an important issue in recent research with applications in the cosmetic industry and medicine. In this paper, we analyzed the dependence of skin texture features on wavelength as well as on different parameters (age and gender) of human participants using grey-level co-occurrence matrix and hyperspectral imaging technique for a more accurate quantitative assessment of the aging process. A total of 42 healthy participants (men and women; age range, 20-70 years) was enrolled in this study. A region of interest was selected from the hyperspectral images. The results were analyzed in terms of texture using the gray-level co-occurrence matrix which generated four features (homogeneity, contrast, entropy, and correlation). The results showed that most of these features displayed variations with wavelength (the exception was entropy), with higher variations in women. Only correlation in both sexes and contrast in men proved to vary statistically significant with age, making them the targeted variables in future attempts to characterize aging skin using the complex method of hyperspectral imaging. In conclusion, by using hyperspectral imaging some measure of the degree of damage or the aging process of the hand skin can be obtained, mainly in terms of correlation values. At the present time, reasonable explanations that can link the process of skin aging and the above mentioned features could not be found, but deeper investigations are on the way.

Brain tumour classification and abnormality detection using neuro-fuzzy technique and Otsu thresholding

Brain tumour is one of the main causes for an increase in transience among children and adults. This paper proposes an improved method based on Magnetic Resonance Imaging (MRI) brain image classification and image segmentation approach. Automated classification is encouraged by the need of high accuracy when dealing with a human life. The detection of the brain tumour is a challenging problem, due to high diversity in tumour appearance and ambiguous tumour boundaries. MRI images are chosen for detection of brain tumours, as they are used in soft tissue determinations. First of all, image pre-processing is used to enhance the image quality. Second, dual-tree complex wavelet transform multi-scale decomposition is used to analyse texture of an image. Feature extraction extracts features from an image using gray-level co-occurrence matrix (GLCM). Then, the Neuro-Fuzzy technique is used to classify the stages of brain tumour as benign, malignant or normal based on texture features. Finally, tumour location is detected using Otsu thresholding. The classifier performance is evaluated based on classification accuracies. The simulated results show that the proposed classifier provides better accuracy than previous method.