Three-fold Cross-validation Research Articles

Convolutional Neural Network for Depression and Schizophrenia Detection

Background/Objectives: This study presents a Convolutional Neural Network (CNN) approach for detecting depression and schizophrenia using motor activity patterns represented as images. Participants’ motor activity data were captured and transformed into visual representations, enabling advanced computer vision techniques for the classification of these mental disorders. The model’s performance was evaluated using a three-fold cross-validation, achieving an average accuracy of 95%, demonstrating the effectiveness of the proposed approach in accurately identifying mental health conditions. The objective of the study is to develop a model capable of identifying different mental disorders by processing motor data using CNN in order to provide a support tool to areas specialized in the diagnosis and efficient treatment of these psychological conditions. Methods: The methodology involved segmenting and transforming motor activity data into images, followed by a CNN training and testing phase on these visual representations. This innovative approach enables the identification of subtle motor behavior patterns, potentially indicative of specific mental states, without invasive interventions or self-reporting. Results: The results suggest that CNNs can capture discriminative features in motor activity to differentiate between individuals with depression, schizophrenia, and those without mental health diagnoses. Conclusions: These findings underscore the potential of computer vision and deep neural network techniques to contribute to early, non-invasive mental health disorder diagnosis, with significant implications for developing mental health support tools.

Construction of a Compound Model to Enhance the Accuracy of Hepatic Fat Fraction Estimation with Quantitative Ultrasound.

Background: we evaluated regression models based on quantitative ultrasound (QUS) parameters and compared them with a vendor-provided method for calculating the ultrasound fat fraction (USFF) in metabolic dysfunction-associated steatotic liver disease (MASLD). Methods: We measured the attenuation coefficient (AC) and the backscatter-distribution coefficient (BSC-D) and determined the USFF during a liver ultrasound and calculated the magnetic resonance imaging proton-density fat fraction (MRI-PDFF) and steatosis grade (S0-S4) in a combined retrospective-prospective cohort. We trained multiple models using single or various QUS parameters as independent variables to forecast MRI-PDFF. Linear and nonlinear models were trained during five-time repeated three-fold cross-validation in a retrospectively collected dataset of 60 MASLD cases. We calculated the models' Pearson correlation (r) and the intraclass correlation coefficient (ICC) in a prospectively collected test set of 57 MASLD cases. Results: The linear multivariable model (r = 0.602, ICC = 0.529) and USFF (r = 0.576, ICC = 0.54) were more reliable in S0- and S1-grade steatosis than the nonlinear multivariable model (r = 0.492, ICC = 0.461). In S2 and S3 grades, the nonlinear multivariable (r = 0.377, ICC = 0.32) and AC-only (r = 0.375, ICC = 0.313) models' approximated correlation and agreement surpassed that of the multivariable linear model (r = 0.394, ICC = 0.265). We searched a QUS parameter grid to find the optimal thresholds (AC ≥ 0.84 dB/cm/MHz, BSC-D ≥ 105), above which switching from a linear (r = 0.752, ICC = 0.715) to a nonlinear multivariable (r = 0.719, ICC = 0.641) model could improve the overall fit (r = 0.775, ICC = 0.718). Conclusions: The USFF and linear multivariable models are robust in diagnosing low-grade steatosis. Switching to a nonlinear model could enhance the fit to MRI-PDFF in advanced steatosis.

Adaptive Evolutionary Optimization of Deep Learning Architectures for Focused Liver Ultrasound Image Segmentation.

Background: Liver ultrasound segmentation is challenging due to low image quality and variability. While deep learning (DL) models have been widely applied for medical segmentation, generic pre-configured models may not meet the specific requirements for targeted areas in liver ultrasound. Quantitative ultrasound (QUS) is emerging as a promising tool for liver fat measurement; however, accurately segmenting regions of interest within liver ultrasound images remains a challenge. Methods: We introduce a generalizable framework using an adaptive evolutionary genetic algorithm to optimize deep learning models, specifically U-Net, for focused liver segmentation. The algorithm simultaneously adjusts the depth (number of layers) and width (neurons per layer) of the network, dropout, and skip connections. Various architecture configurations are evaluated based on segmentation performance to find the optimal model for liver ultrasound images. Results: The model with a depth of 4 and filter sizes of [16, 64, 128, 256] achieved the highest mean adjusted Dice score of 0.921, outperforming the other configurations, using three-fold cross-validation with early stoppage. Conclusions: Adaptive evolutionary optimization enhances the deep learning architecture for liver ultrasound segmentation. Future work may extend this optimization to other imaging modalities and deep learning architectures.

The analysis of gene co-expression network and immune infiltration revealed biomarkers between triple-negative and non-triple negative breast cancer.

Triple-negative breast cancer (TNBC) is a heterogeneous disease with a worse prognosis. Despite ongoing efforts, existing therapeutic approaches show limited success in improving early recurrence and survival outcomes for TNBC patients. Therefore, there is an urgent need to discover novel and targeted therapeutic strategies, particularly those focusing on the immune infiltrate in TNBC, to enhance diagnosis and prognosis for affected individuals. The gene co-expression network and gene ontology analyses were used to identify the differential modules and their functions based on the GEO dataset of GSE76275. The Weighted Gene Co-Expression Network Analysis (WGCNA) was used to describe the correlation patterns among genes across multiple samples. Subsequently, we identified key genes in TNBC by assessing genes with an absolute correlation coefficient greater than 0.80 within the eigengene of the enriched module that were significantly associated with breast cancer subtypes. The diagnostic potential of these key genes was evaluated using receiver operating characteristic (ROC) curve analysis with three-fold cross-validation. Furthermore, to gain insights into the prognostic implications of these key genes, we performed relapse-free survival (RFS) analysis using the Kaplan-Meier plotter online tool. CIBERSORT analysis was used to characterize the composition of immune cells within complex tissues based on gene expression data, typically derived from bulk RNA sequencing or microarray datasets. Therefore, we explored the immune microenvironment differences between TNBC and non-TNBC by leveraging the CIBERSORT algorithm. This enabled us to estimate the immune cell compositions in the breast cancer tissue of the two subtypes. Lastly, we identified key transcription factors involved in macrophage infiltration and polarization in breast cancer using transcription factor enrichment analysis integrated with orthogonal omics. The gene co-expression network and gene ontology analyses revealed 19 modules identified using the dataset GSE76275. Of these, modules 5, 11, and 12 showed significant differences between in breast cancer tissue between TNBC and non-TNBC. Notably, module 11 showed significant enrichment in the WNT signaling pathway, while module 12 demonstrated enrichment in lipid/fatty acid metabolism pathways. Subsequently, we identified SHC4/KCNK5 and ABCC11/ABCA12 as key genes in module 11 and module 12, respectively. These key genes proved to be crucial in accurately distinguishing between TNBC and non-TNBC, as evidenced by the promising average AUC value of 0.963 obtained from the logistic regression model based on their combinations. Furthermore, we found compelling evidence indicating the prognostic significance of three key genes, KCNK5, ABCC11, and ABCA12, in TNBC. Finally, we also identified the immune cell compositions in breast cancer tissue between TNBC and non-TNBC. Our findings revealed a notable increase in M0 and M1 macrophages in TNBC compared to non-TNBC, while M2 macrophages exhibited a significant reduction in TNBC. Particularly intriguing discovery emerged with respect to the transcription factor FOXM1, which demonstrated a significant regulatory role in genes positively correlated with the proportions of M0 and M1 macrophages, while displaying a negative correlation with the proportion of M2 macrophages in breast cancer tissue. Our research provides new insight into the biomarkers and immune infiltration of TNBC, which could be useful for clinical diagnosis of TNBC.

A feature-based approach for atlas selection in automatic pelvic segmentation.

Accurate and efficient automatic segmentation is essential for various clinical tasks such as radiotherapy treatment planning. However, atlas-based segmentation still faces challenges due to the lack of representative atlas dataset and the computational limitations of deformation algorithms. In this work, we have proposed an atlas selection procedure (subset atlas grouping approach, MAS-SAGA) which utilized both image similarity and volume features for selecting the best-fitting atlases for contour propagation. A dataset of anonymized female pelvic Computed Tomography (CT) images demonstrated that MAS-SAGA significantly outperforms conventional multi-atlas-based segmentation (cMAS) in terms of Dice Similarity Coefficient (DSC) and 95th Percentile Hausdorff Distance (95HD) for bladder and rectum segmentation using a three-fold cross-validation strategy. The proposed procedure also reduced computation time compared to cMAS, making it a promising tool for medical image analysis applications. In addition, we have evaluated two distinct atlas selection methods: the Feature-based Atlas Selection Approach (MAS-FASA) and the Similarity-based Atlas Selection Approach (MAS-SIM). We investigate the differences between these two methods in terms of their ability to select the best fitting atlases. The findings demonstrated that MAS-FASA selected different atlases than MAS-SIM, resulting in improved segmentation performance overall. It highlighted the potential of feature-based subgrouping techniques in enhancing the efficacy of MAS algorithms in the field of medical image segmentation.

An interpretable machine learning model for precise prediction of biomarkers for intermittent fasting pattern

Intermittent fasting is currently a highly sought-after dietary pattern. To explore the potential biomarkers of intermittent fasting, untargeted metabolomics analysis of fecal metabolites in two groups of mice, intermittent fasting and normal feeding, was conducted using UPLC-HRMS. The data was further analyzed through interpretable machine learning (ML) to data mine the biomarkers for two dietary patterns. We developed five machine learning models and results showed that under three-fold cross-validation, Random Forest model was the most suitable for distinguishing the two dietary patterns. Finally, Shapely Additive exPlanations (SHAP) were explored to perform a weighted explanatory analysis on the Random Forest model, and the contribution of each metabolite to the model was calculated. Results indicated that Ganoderenic Acid C is the potential biomarkers to distinguish the two dietary patterns. Our work provides new insights for metabolic biomarker analysis and lays a theoretical foundation for the selection of a healthieir dietary lifestyle.

Modeling the fasting blood glucose response to basal insulin adjustment in type 2 diabetes: An explainable machine learning approach on real-world data.

Optimal basal insulin titration for people with type 2 diabetes is vital to effectively reducing the risk of complications. However, a sizeable proportion of people (30-50%) remain in suboptimal glycemic control six months post-initiation of basal insulin. This indicates a clear need for novel titration methods that account for individual patient variability in real-world settings. This study aims to investigate the use of real-world data and explainable machine learning in modeling fasting glucose responses to basal insulin adjustments, focusing on identifying factors influencing fasting glucose variability. A three-step explanatory approach was used to develop models using multiple linear regression, forward feature selection, and three-fold cross-validation. The models were built progressively, starting with a baseline model incorporating fasting blood glucose and insulin dose adjustments, followed by iterative models that in turn included biometric data, social factors, and biochemistry data, and lastly, a comprehensive model without constraints on the feature pool. The baseline model yielded an average root mean squared error (RMSE) of 1.52 [95% CI: 1.33-1.71]. The iterative models resulted in an average RMSE of 1.49 [95% CI: 1.35-1.62] (biometric data), 1.47 [95% CI: 1.36-1.58] (social factors), and 1.52 [95% CI: 1.34-1.70] (biochemistry data). The comprehensive model yielded an average RMSE of 1.44 [95% CI: 1.41-1.48]. Developing explainable machine learning models using real-world data is possible for basal insulin titration. However, model performance is influenced by data's ability to capture everyday behavior, underscoring the need for incorporating more detailed behavioral and social data to optimize future titration models.

Multi-parameter MRI radiomics model in predicting postoperative progressive cerebral edema and hemorrhage after resection of meningioma

BackgroundPostoperative progressive cerebral edema and hemorrhage (PPCEH) are major complications after meningioma resection, yet their preoperative predictive studies are limited. The aim is to develop and validate a multiparametric MRI machine learning model to predict PPCEH after meningioma resection.MethodsThis retrospective study included 148 patients with meningioma. A stratified three-fold cross-validation was used to split the dataset into training and validation sets. Radiomics features from the tumor enhancement (TE) and peritumoral brain edema (PTBE) regions were extracted from T1WI, T2WI, and ADC maps. Support vector machine constructed different radiomics models, and logistic regression explored clinical risk factors. Prediction models, integrating clinical and radiomics features, were evaluated using the area under the curve (AUC), visualized in a nomogram.ResultsThe radiomics model based on TE and PTBE regions (training set mean AUC: 0.85 (95% CI: 0.78–0.93), validation set mean AUC: 0.77 (95%CI: 0.63–0.90)) outperformed the model with TE region solely (training set mean AUC: 0.83 (95% CI: 0.76–0.91), validation set mean AUC: 0.73 (95% CI: 0.58–0.87)). Furthermore, the combined model incorporating radiomics features, and clinical features of preoperative peritumoral edema and tumor boundary adhesion, had the best predictive performance, with AUC values of 0.87 (95% CI: 0.80–0.94) and 0.84 (95% CI: 0.72–0.95) for the training and validation set.ConclusionsWe developed a novel model based on clinical characteristics and multiparametric radiomics features derived from TE and PTBE regions, which can accurately and non-invasively predict PPCEH after meningioma resection. Additionally, our findings suggest the crucial role of PTBE radiomics features in understanding the potential mechanisms of PPCEH.

Cross-Domain Mutual-Assistance Learning Framework for Fully Automated Diagnosis of Primary Tumor in Nasopharyngeal Carcinoma.

Accurate T-staging of nasopharyngeal carcinoma (NPC) holds paramount importance in guiding treatment decisions and prognosticating outcomes for distinct risk groups. Regrettably, the landscape of deep learning-based techniques for T-staging in NPC remains sparse, and existing methodologies often exhibit suboptimal performance due to their neglect of crucial domain-specific knowledge pertinent to primary tumor diagnosis. To address these issues, we propose a new cross-domain mutual-assistance learning framework for fully automated diagnosis of primary tumor using H&N MR images. Specifically, we tackle primary tumor diagnosis task with the convolutional neural network consisting of a 3D cross-domain knowledge perception network (CKP net) for excavated cross-domain-invariant features emphasizing tumor intensity variations and internal tumor heterogeneity, and a multi-domain mutual-information sharing fusion network (M2SF net), comprising a dual-pathway domain-specific representation module and a mutual information fusion module, for intelligently gauging and amalgamating multi-domain, multi-scale T-stage diagnosis-oriented features. The proposed 3D cross-domain mutual-assistance learning framework not only embraces task-specific multi-domain diagnostic knowledge but also automates the entire process of primary tumor diagnosis. We evaluate our model on an internal and an external MR images dataset in a three-fold cross-validation paradigm. Exhaustive experimental results demonstrate that our method outperforms the other algorithms, and obtains promising performance for tumor segmentation and T-staging. These findings underscore its potential for clinical application, offering valuable assistance to clinicians in treatment decision-making and prognostication for various risk groups.

Autism Identification Based on the Intelligent Analysis of Facial Behaviors: An Approach Combining Coarse- and Fine-Grained Analysis.

Facial behavior has emerged as a crucial biomarker for autism identification. However, heterogeneity among individuals with autism poses a significant obstacle to traditional feature extraction methods, which often lack the necessary discriminative power. While deep-learning methods hold promise, they are often criticized for their lack of interpretability. To address these challenges, we developed an innovative facial behavior characterization model that integrates coarse- and fine-grained analyses for intelligent autism identification. The coarse-grained analysis provides a holistic view by computing statistical measures related to facial behavior characteristics. In contrast, the fine-grained component uncovers subtle temporal fluctuations by employing a long short-term memory (LSTM) model to capture the temporal dynamics of head pose, facial expression intensity, and expression types. To fully harness the strengths of both analyses, we implemented a feature-level attention mechanism. This not only enhances the model's interpretability but also provides valuable insights by highlighting the most influential features through attention weights. Upon evaluation using three-fold cross-validation on a self-constructed autism dataset, our integrated approach achieved an average recognition accuracy of 88.74%, surpassing the standalone coarse-grained analysis by 8.49%. This experimental result underscores the improved generalizability of facial behavior features and effectively mitigates the complexities stemming from the pronounced intragroup variability of those with autism, thereby contributing to more accurate and interpretable autism identification.

Developing a logistic regression model to predict spontaneous preterm birth from maternal socio-demographic and obstetric history at initial pregnancy registration

BackgroundCurrent predictive machine learning techniques for spontaneous preterm birth heavily rely on a history of previous preterm birth and/or costly techniques such as fetal fibronectin and ultrasound measurement of cervical length to the disadvantage of those considered at low risk and/or those who have no access to more expensive screening tools.Aims and objectivesWe aimed to develop a predictive model for spontaneous preterm delivery < 37 weeks using socio-demographic and clinical data readily available at booking -an approach which could be suitable for all women regardless of their previous obstetric history.MethodsWe developed a logistic regression model using seven feature variables derived from maternal socio-demographic and obstetric history from a preterm birth (n = 917) and a matched full-term (n = 100) cohort in 2018 and 2020 at a tertiary obstetric unit in the UK. A three-fold cross-validation technique was applied with subsets for data training and testing in Python® (version 3.8) using the most predictive factors. The model performance was then compared to the previously published predictive algorithms.ResultsThe retrospective model showed good predictive accuracy with an AUC of 0.76 (95% CI: 0.71–0.83) for spontaneous preterm birth, with a sensitivity and specificity of 0.71 (95% CI: 0.66–0.76) and 0.78 (95% CI: 0.63–0.88) respectively based on seven variables: maternal age, BMI, ethnicity, smoking, gestational type, substance misuse and parity/obstetric history.ConclusionPending further validation, our observations suggest that key maternal demographic features, incorporated into a traditional mathematical model, have promising predictive utility for spontaneous preterm birth in pregnant women in our region without the need for cervical length and/or fetal fibronectin.

Dry age-related macular degeneration classification from optical coherence tomography images based on ensemble deep learning architecture.

Dry age-related macular degeneration (AMD) is a retinal disease, which has been the third leading cause of vision loss. But current AMD classification technologies did not focus on the classification of early stage. This study aimed to develop a deep learning architecture to improve the classification accuracy of dry AMD, through the analysis of optical coherence tomography (OCT) images. We put forward an ensemble deep learning architecture which integrated four different convolution neural networks including ResNet50, EfficientNetB4, MobileNetV3 and Xception. All networks were pre-trained and fine-tuned. Then diverse convolution neural networks were combined. To classify OCT images, the proposed architecture was trained on the dataset from Shenyang Aier Excellence Hospital. The number of original images was 4,096 from 1,310 patients. After rotation and flipping operations, the dataset consisting of 16,384 retinal OCT images could be established. Evaluation and comparison obtained from three-fold cross-validation were used to show the advantage of the proposed architecture. Four metrics were applied to compare the performance of each base model. Moreover, different combination strategies were also compared to validate the merit of the proposed architecture. The results demonstrated that the proposed architecture could categorize various stages of AMD. Moreover, the proposed network could improve the classification performance of nascent geographic atrophy (nGA). In this article, an ensemble deep learning was proposed to classify dry AMD progression stages. The performance of the proposed architecture produced promising classification results which showed its advantage to provide global diagnosis for early AMD screening. The classification performance demonstrated its potential for individualized treatment plans for patients with AMD.

B-280 Nmr-based biomaker measurement determines high-dose methotrexate elimination rates in lymphoma and leukemia treatment

Abstract Background The cytotoxic methotrexate (MTX) is used in high doses in lymphoma and acute lymphocytic leukemia (ALL) treatment. A therapeutic plasma dose range is crucial to minimize MTX accumulation and prolonged exposure, which can lead to myelosuppression and other adverse effects. Determining MTX elimination prior to drug administration is difficult, given its non-linear clearance. Here we present a model to predict MTX elimination rates based on metabolite measurement in patients receiving high-dose methotrexate, using nuclear-magnetic resonance (NMR) spectroscopy. Methods A series of 26 residual sera from patients receiving MTX infusions as per treatment protocols for CNS lymphoma (MATRIX protocol, 3.5 g/m2 over four hours) and ALL (GMALL protocol, 1.5 g/m2 over 24 hours) were collected from patients at the University Hospital Regensburg at four timepoints (t0, t24, t42, and t48). Serum was NMR-measured in five replicates. NMR spectra for each sample were characterized and relevant parameters extracted. First order elimination constants were modeled for two MTX dosage regimes (&amp;lt;4 g, ≥4 g) using characterized NMR features from samples measured at t0. Results Metabolomic and lipidomic biomarkers measured at baseline (t0) were able to accurately predict MTX elimination rate constant k, determined by fitting MTX levels at a t24, t42, and t48, using first order c(t)∼c0*e(-kt) kinetic function (&amp;gt;4 g: R2 = 0.475, p-value &amp;lt; 0.05; ≤4 g: R2 = 0.918, p-value &amp;lt; 0.01) (Fig. A). Results were cross-validated with five-time repeated three-fold cross-validation. Performance values are given as means of the folds. Conclusions NMR-measured biomarkers can predict dose-dependent MTX elimination rates in patients receiving high-dose MTX, presenting a novel method to predict MTX clearance, allowing for personalized dose-monitoring and minimization of adverse events. Validation with a larger cohort is required.

MSDAFL: molecular substructure-based dual attention feature learning framework for predicting drug-drug interactions.

Drug-drug interactions (DDIs) can cause unexpected adverse drug reactions, affecting treatment efficacy and patient safety. The need for computational methods to predict DDIs has been growing due to the necessity of identifying potential risks associated with drug combinations in advance. Although several deep learning methods have been recently proposed to predict DDIs, many overlook feature learning based on interactions between the substructures of drug pairs. In this work, we introduce a molecular Substructure-based Dual Attention Feature Learning framework (MSDAFL), designed to fully utilize the information between substructures of drug pairs to enhance the performance of DDI prediction. We employ a self-attention module to obtain a set number of self-attention vectors, which are associated with various substructural patterns of the drug molecule itself, while also extracting interaction vectors representing inter-substructure interactions between drugs through an interactive attention module. Subsequently, an interaction module based on cosine similarity is used to further capture the interactive characteristics between the self-attention vectors of drug pairs. We also perform normalization after the interaction feature extraction to mitigate overfitting. After applying three-fold cross-validation, the MSDAFL model achieved average precision scores of 0.9707, 0.9991, and 0.9987, and area under the receiver operating characteristic curve scores of 0.9874, 0.9934, and 0.9974 on three datasets, respectively. In addition, the experiment results of five-fold cross-validation and cross-datum study also indicate that MSDAFL performs well in predicting DDIs. Data and source codes are available at https://github.com/27167199/MSDAFL.

Deep learning approaches for assessing pediatric sleep apnea severity through SpO2 signals

Pediatric Sleep Apnea–Hypopnea (SAH) presents a significant health challenge, particularly in diagnostic contexts, where conventional Polysomnography (PSG) testing, although effective, can be distressing for children. Addressing this, our research proposes a less invasive method to assess pediatric SAH severity by analyzing blood oxygen saturation (SpO2) signals. We adopted two advanced deep learning architectures, namely ResNet-based and attention-augmented hybrid CNN-BiGRU models, to process SpO2 signals in a one-dimensional (1D) format for Apnea–Hypopnea Index (AHI) estimation in pediatric subjects. Employing the CHAT dataset, which includes 844 SpO2 signals, the data was partitioned into training (60%), testing (30%), and validation (10%) sets. A predefined validation subset was randomly selected to ensure the models' robustness via a threefold cross-validation approach. Comparative analysis revealed that while the ResNet model attained an average accuracy of 72.9% across four SAH severity categories with a kappa score of 0.57, the CNN-BiGRU-Attention model demonstrated superior performance, achieving an average accuracy of 75.95% and a kappa score of 0.63. This distinction underscores our method's efficacy in both estimating AHI and categorizing SAH severity levels with notable precision. Further, to evaluate diagnostic capabilities, the models were benchmarked against common AHI thresholds (1, 5, and 10 events/hour) in each test fold, affirming their effectiveness in identifying pediatric SAH. This study marks a significant advance in the field, offering a non-invasive, child-friendly alternative for pediatric SAH diagnosis. Although challenges persist in accurately estimating AHI, particularly in severe cases, our findings represent a critical stride towards improving diagnostic processes in pediatric SAH.

Reproducible and Interpretable Machine Learning-Based Radiomic Analysis for Overall Survival Prediction in Glioblastoma Multiforme

Purpose: To develop and validate an MRI-based radiomic model for predicting overall survival (OS) in patients diagnosed with glioblastoma multiforme (GBM), utilizing a retrospective dataset from multiple institutions. Materials and Methods: Pre-treatment MRI images of 289 GBM patients were collected. From each patient’s tumor volume, 660 radiomic features (RFs) were extracted and subjected to robustness analysis. The initial prognostic model with minimum RFs was subsequently enhanced by including clinical variables. The final clinical–radiomic model was derived through repeated three-fold cross-validation on the training dataset. Performance evaluation included assessment of concordance index (C-Index), integrated area under curve (iAUC) alongside patient stratification into low and high-risk groups for overall survival (OS). Results: The final prognostic model, which has the highest level of interpretability, utilized primary gross tumor volume (GTV) and one MRI modality (T2-FLAIR) as a predictor and integrated the age variable with two independent, robust RFs, achieving moderately good discriminatory performance (C-Index [95% confidence interval]: 0.69 [0.62–0.75]) with significant patient stratification (p = 7 × 10−5) on the validation cohort. Furthermore, the trained model exhibited the highest iAUC at 11 months (0.81) in the literature. Conclusion: We identified and validated a clinical–radiomic model for stratification of patients into low and high-risk groups based on OS in patients with GBM using a multicenter retrospective dataset. Future work will focus on the use of deep learning-based features, with recently standardized convolutional filters on OS tasks.

Study on Transportation Carbon Emissions in Tibet: Measurement, Prediction Model Development, and Analysis

In recent years, the socio-economic development in the Tibet region of China has experienced substantial growth. However, transportation increasingly strains the region’s fragile ecological environment. Most studies overlook the accurate measurement and analysis of factors influencing traffic carbon emissions in Tibet due to data scarcity. To address this, this paper applies an improved traffic carbon emissions model, using transportation turnover data to estimate emissions in Tibet from 2008 to 2020. Simultaneously, the estimated traffic carbon emissions in Tibet served as the predicted variable, and various machine learning algorithms, including Radial Basis Function Support Vector Machine (RBF-SVM), eXtreme Gradient Boosting (XGBoost), Random Forest, and Gradient Boosting Decision Tree (GBDT) are employed to conduct an initial comparison of the constructed prediction models using three-fold cross-validation and multiple evaluation metrics. The best-performing model undergoes further optimization using Grid Search (GS) and Real-coded Genetic Algorithm (RGA). Finally, the central difference method and Local Interpretable Model-Agnostic Explanation (LIME) algorithm are used for local sensitivity and interpretability analyses on twelve core variables. The results assess each variable’s contribution to the model’s output, enabling a comprehensive analysis of their impact on Tibet’s traffic carbon emissions. The findings demonstrate a significant upward trend in Tibet’s traffic carbon emissions, with road transportation and civil aviation being the main contributors. The RBF-SVM algorithm is most suitable for predicting traffic carbon emissions in this region. After GS optimization, the model’s R2 value exceeded 0.99, indicating high predictive accuracy and stability. Key factors influencing traffic carbon emissions in Tibet include civilian vehicle numbers, transportation land-use area, transportation output value, urban green coverage areas, per capita GDP, and built-up area. This paper provides a systematic framework and empirical support for measuring, predicting, and analyzing factors influencing traffic carbon emissions in Tibet. It employs innovative measurement methods, optimized machine learning models, and detailed sensitivity and interpretability analyses. The results can guide regional carbon reduction targets and promote green sustainable development.

An integrated computer-aided diagnosis BCanD model for detection, segmentation and classification of breast cancer

A fully integrated Computer-Aided Diagnosis (CAD) system involves the integration of detection, segmentation, and classification, which makes it very useful for medical applications, particularly while dealing with the detection of breast mass and its classification into malignant and benign. The carried-out research work is intended to propose a Breast Cancer Detection (BCanD) model that is an integrated CAD system, where the system is capable enough for mass detection, its segmentation, and for the classification using mammograms. The proposed integrated system utilizes deep learning based YOLO model to detect the abnormality (mass) in the mammogram, where U-net is used for segmentation of the mass, as it has the capability to produce pixel level segmentation map, and at last stage that is the classification stage deep CNN is used for the classification. The proposed system is evaluated on open-source MIAS database. For the performance evaluation of the proposed BCanD, a three-fold cross-validation test was utilized. The mass detection accuracy of the BCanD is 98.99%, MCC is 97.96%, and F1-score is 98.87%. The model is evaluated with and without automated mass segmentation to study the impact of segmentation on the suggested CAD system. The best results was observed with the segmentation with the overall accuracy of 94.20%, F1-score (Dice) of 93.60%, MCC of 88.33%, and Jaccard of 88.08%. The proposed BCanD model surpasses the latest existing deep learning-based methodologies like fuzzy classifier, CNNI-BCC etc Hence, the proposed CAD system can be implemented and used by radiologists for all the stages from detection to diagnosis of breast mass.

Unraveling gender-specific lipids and flavor volatiles in giant salamander (Andrias davidianus) livers via lipidomics and GC-IMS

To uncover the relationships between lipid components and flavor volatiles, distinctness in lipid components and odor substances in giant salamander livers of different genders were comparatively characterized through UPLC-Q Exactive-MS lipidomics and gas chromatography-ion migration spectrometry (GC-IMS). A total of 2171 and 974 lipid metabolites were detected in positive and negative ion modes, respectively. Triglycerides (TG) and phosphatidylcholines (PC) are the most abundant types of lipids. TG level in male livers was higher than that in female livers (P < 0.05), whereas PC level showed no marked variation (P > 0.05). Additionally, a total of 51 volatile components were detected through GC-IMS. Ketones (42.18 % ∼ 45.44 %) and alcohols (24.19 % ∼ 26.50 %) were the predominant categories, and their relative contents were higher in female livers. Finally, 30 differential lipid metabolites and 12 differential odor substances were screened and could be used as distinguishing labels in giant salamander livers of different genders. Correlation analysis indicated that PS(36:2e), TG(48:13), ZyE(37:6), and ZyE(33:6) correlated positively with 3-methyl butanal, 3-hydroxy-2-butanone, and 2-methyl-1-propanol (P < 0.05), but adversely linked with 1-penten-3-one, and 1-octen-3-one (P < 0.01). By three-fold cross-validation, prediction accuracies of these differential lipids and volatile compounds for gender recognition based on random forest model were 100 % and 92 %, respectively. These findings might not only add knowledge on lipid and volatile profiles in giant salamander livers as affected by genders, but also provide clues for their gender recognition.

Unraveling druggable cancer-driving proteins and targeted drugs using artificial intelligence and multi-omics analyses

The druggable proteome refers to proteins that can bind to small molecules with appropriate chemical affinity, inducing a favorable clinical response. Predicting druggable proteins through screening and in silico modeling is imperative for drug design. To contribute to this field, we developed an accurate predictive classifier for druggable cancer-driving proteins using amino acid composition descriptors of protein sequences and 13 machine learning linear and non-linear classifiers. The optimal classifier was achieved with the support vector machine method, utilizing 200 tri-amino acid composition descriptors. The high performance of the model is evident from an area under the receiver operating characteristics (AUROC) of 0.975 ± 0.003 and an accuracy of 0.929 ± 0.006 (threefold cross-validation). The machine learning prediction model was enhanced with multi-omics approaches, including the target-disease evidence score, the shortest pathways to cancer hallmarks, structure-based ligandability assessment, unfavorable prognostic protein analysis, and the oncogenic variome. Additionally, we performed a drug repurposing analysis to identify drugs with the highest affinity capable of targeting the best predicted proteins. As a result, we identified 79 key druggable cancer-driving proteins with the highest ligandability, and 23 of them demonstrated unfavorable prognostic significance across 16 TCGA PanCancer types: CDKN2A, BCL10, ACVR1, CASP8, JAG1, TSC1, NBN, PREX2, PPP2R1A, DNM2, VAV1, ASXL1, TPR, HRAS, BUB1B, ATG7, MARK3, SETD2, CCNE1, MUTYH, CDKN2C, RB1, and SMARCA4. Moreover, we prioritized 11 clinically relevant drugs targeting these proteins. This strategy effectively predicts and prioritizes biomarkers, therapeutic targets, and drugs for in-depth studies in clinical trials. Scripts are available at https://github.com/muntisa/machine-learning-for-druggable-proteins.