Training Set Research Articles

Intratumoral and peritumoral radiomics for preoperative prediction of neoadjuvant chemotherapy effect in breast cancer based on 18F-FDG PET/CT

ObjectiveTo investigate the value of 18F-FDG PET/CT-based intratumoral and peritumoral radiomics in predicting the efficacy of neoadjuvant chemotherapy (NAC) for breast cancer.Methods190 patients who met the inclusion and exclusion criteria from 2017 to 2022 were studied. Features were extracted from the PET/CT intratumoral and peritumoral regions, feature selection was performed through the correlation analysis, t-tests, and least absolute shrinkage and selection operator regression (LASSO). Four classifiers, support vector machine (SVM), k-nearest neighbor (KNN), logistic regression (LR), and naive bayes (NB) were used to build the prediction models. The receiver operating characteristic (ROC) curves were plotted to measure the predictive performance of the models. Concurrent stratified analysis was conducted to establish subtype-specific features for each molecular subtype.ResultsCompared to intratumoral features alone, intratumoral + peritumoral features achieved higher AUC values in each classifier. The SVM model constructed with intratumoral + peritumoral features achieved the highest AUC values in both the train and test set (train set: 0.95 and test set: 0.83). Subtype-specific features improve performance in predicting the efficacy of NAC (luminal group: 0.90; HER2 + group: 0.86; triple negative group: 0.92).ConclusionIntratumoral and peritumoral radiomics models based on 18F-FDG PET/CT can reliably forecast the efficacy of NAC, thereby assisting clinical decision-making.

Magnetic 3D macroporous MOF oriented urinary exosome metabolomics for early diagnosis of bladder cancer

Bladder cancer (BCa) exhibits the escalating incidence and mortality due to the untimely and inaccurate early diagnosis. Urinary exosome metabolites, carrying critical tumor cell information and directly related to bladder, emerge as promising non-invasive diagnostic biomarkers of BCa. Herein, the magnetic 3D ordered macroporous zeolitic imidazolate framework-8 (magMZIF-8) is synthesized and used for efficient urinary exosome isolation. Notably, beyond retaining the single crystals and micropores of conventional ZIF-8, MZIF-8 is further enhanced with highly oriented and ordered macropores (150 nm) and the large specific surface area (973 m2·g–1), which could enable the high purity and yield separation of exosomes via leveraging the combination of size exclusion, affinity, and electrostatic interactions between magMZIF-8 and the surfaces of exosome. Furthermore, the magnetic and hydrophilic properties of magMZIF-8 will further simplify the process and enhance the efficiency of separation. After conditional optimization, a 50 mL of urine is sufficient for exosome metabolomics analysis, and the time for isolating exosomes from 42 urine samples was 2 hours only. Incorporating machine learning algorithms with LC-MS/MS analysis of the metabolic patterns obtained from isolated exosomes, early-stage BCa patients were differentiated from healthy controls, with area under the curve (AUC) value of 0.844–0.9970 in the training set and 0.875-1.00 in the test set, signifying its potential as a reliable diagnostic tool. This study offers a promising approach for the non-invasive and efficient diagnosis of BCa on a large scale via exosome metabolomics.Graphical

Five-gene prognostic model based on autophagy-dependent cell death for predicting prognosis in lung adenocarcinoma

Non-small cell lung adenocarcinoma (LUAD) is the predominant form of lung cancer originating from lung epithelial cells, making it the most prevalent pathological type. Currently, reliable indicators for predicting treatment efficacy and disease prognosis are lacking. Despite extensive validation of autophagy-dependent cell death (ADCD) in solid tumor studies and its correlation with immunotherapy effectiveness and cancer prognosis, systematic research on ADCD-related genes in LUAD is limited. We utilized AddModuleScore, ssGSEA, and WGCNA to identify genes associated with ADCD across single-cell and bulk transcriptome datasets. The TCGA dataset, comprising 598 cases, was randomly divided into training and validation sets to develop an ADCD-related LUAD prediction model. Internal validation was performed using the TCGA validation set. For external validation, datasets GSE13213 (119 LUAD samples), GSE26939 (115 LUAD samples), GSE29016 (39 LUAD samples), and GSE30219 (86 LUAD samples) were employed. We evaluated the model’s accuracy and effectiveness in predicting prognostic risk. Additionally, CIBERSORT, ESTIMATE, and ssGSEA techniques were used to explore immunological characteristics, drug response, and gene expression in LUAD. Real-time RT-PCR was conducted to assess variations in mRNA expression levels of the gene XCR1 between cancerous and normal tissues in 10 lung cancer patients. We identified 249 genes associated with autophagy-dependent cell death (ADCD) at both single-cell and bulk transcriptome levels. Univariate COX regression analysis revealed that 18 genes were significantly associated with overall survival (OS). Using LASSO-Cox analysis, we developed an ADCD signature based on five genes (BIRC3, TAP1, SLAMF1, XCR1, and HLA-DMB) and created the ADCD-related risk scoring system (ADCDRS). Validation of this model demonstrated its ability to predict disease prognosis and its correlation with clinical characteristics, immune cell infiltration, and the tumor microenvironment. To enhance clinical applicability, we integrated an ADCDRS nomogram. Furthermore, we identified potential drugs targeting specific risk subgroups. We successfully identified a model based on five ADCD genes to predict disease prognosis and treatment efficacy in LUAD, as well as to assess the tumor immune microenvironment. An efficient and practical ADCDRS nomogram was designed.

A Comprehensive Nomogram Integrating Phonocardiogram and Echocardiogram Features for the Diagnosis of Heart Failure With Preserved Ejection Fraction.

Heart failure with preserved ejection fraction (HFpEF) is associated with high hospitalization and mortality rates, representing a significant healthcare burden. This study aims to utilize various information including echocardiogram and phonocardiogram to construct and validate a nomogram, assisting in clinical decision-making. This study analyzed 204 patients (68 HFpEF and 136 non-HFpEF) from the First Affiliated Hospital of Chongqing Medical University. A total of 49 features were integrated and used, including phonocardiogram, echocardiogram features, and clinical parameters. The least absolute shrinkage and selection operator (LASSO) regression was used to select the optimal matching factors, and a stepwise logistic regression was employed to determine independent risk factors and develop a nomogram. Model performance was evaluated by the area under receiver operating characteristic (ROC) curve (AUC), calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC). The nomogram was constructed using five significant indicators, including NT-proBNP (OR = 4.689, p = 0.015), E/e' (OR = 1.219, p = 0.032), LAVI (OR = 1.088, p < 0.01), D/S (OR = 0.014, p < 0.01), and QM1 (OR = 1.058, p < 0.01), and showed a better AUC of 0.945 (95% CI = 0.908-0.982) in the training set and 0.933 (95% CI = 0.873-0.992) in the testing set compared to conventional nomogram without phonocardiogram features. The calibration curve and Hosmer-Lemeshow test demonstrated no statistical significance in the training and testing sets (p = 0.814 and p = 0.736), indicating the nomogram was well-calibrated. The DCA and CIC results confirmed favorable clinical usefulness. The nomogram, integrating phonocardiogram and echocardiogram features, enhances HFpEF diagnostic efficiency, offering a valuable tool for clinical decision-making.

Machine learning prediction of stalk lignin content using Fourier transform infrared spectroscopy in large scale maize germplasm

Lignin has been recognized as a major factor contributing to lignocellulosic recalcitrance in biofuel production and attracted attentions as a high-value product in the biorefinery field. As the traditional wet chemical methods for detecting lignin content are labor-intensive, time-consuming and environment-toxic, it is an urgent need to develop high-throughput and environment-friendly techniques for large-scale crop germplasms screening. In this study, we conducted a Fourier transform infrared (FTIR) assay on 150 maize germplasms with a diverse lignin composition to build predictive models for lignin content in maize stalk. Principal component analysis (PCA) was applied to the FTIR spectra for use as model inputs. Classification and advanced gradient boosting machine (GBM) algorithms demonstrated higher predictive accuracy (0.82–0.96) compared to traditional linear and regularization algorithms (0.03–0.04) in the training set. Notably, two optimal models, built using the extreme gradient boosting (XGBoost) and light gradient boosting machine (LightGBM) algorithms, achieved R2 values of over 0.91 in the training set and over 0.82 in the test set. Overall, the combination of FTIR and machine learning (ML) algorithms offers a high-throughput and efficient method for predicting lignin content. This approach holds significant potential for genetic breeding and the effective utilization of maize in industrial production.

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.

Automated Diagnosis of Knee Osteoarthritis Using ResNet101 on a DEEP:PHI: Leveraging a No-Code AI Platform for Efficient and Accurate Medical Image Analysis

Background: Knee osteoarthritis (OA) is a prevalent degenerative joint disease significantly impacting global health. Early and accurate diagnosis is crucial for effective management, but traditional methods often rely on subjective assessments. This study evaluates the efficacy of a deep learning model implemented through a no-code AI platform for diagnosing and grading knee OA from plain radiographs. Methods: We utilized the Osteoarthritis Initiative (OAI) dataset, comprising knee X-ray data from 1526 patients. The data were split into training (47.0%), validation (26.5%), and test (26.5%) sets. We employed a ResNet101 model on the DEEP:PHI no-code AI platform for image analysis. The model was trained to classify knee OA into five grades (0–4) based on the Kellgren–Lawrence scale. Results: Our AI model demonstrated high accuracy in distinguishing between different OA grades, with particular strength in early-stage detection. The model achieved optimal performance at 20 epochs, suggesting efficient learning dynamics. Grad-CAM visualizations were used to enhance the interpretability of the model’s decision-making process. Conclusions: This study demonstrates the potential of AI, implemented through a no-code platform, to accurately diagnose and grade knee OA from radiographs. The use of a no-code AI platform such as DEEP:PHI represents a step towards democratizing AI in healthcare, enabling the rapid development and deployment of sophisticated medical AI applications without extensive coding expertise. This approach could significantly enhance the early detection and management of knee OA, potentially improving patient outcomes and streamlining clinical workflows.

Non-destructive estimation for Kyoho grape shelf-life using Vis/NIR hyperspectral imaging and deep learning algorithm

Grape shelf-life estimation is a substantial challenge for the grape industry. The objective of this study is to investigate the potential of grape shelf-life estimation using HSI technique and a deep learning algorithm. The visible and near-infrared (400.68–1001.61 nm) hyperspectral reflectance images data of grape samples was acquired and preprocessed with different spectral preprocessing methods. Additionally, a stacked denoising autoencoder (SDAE)-based deep learning algorithm was developed to extract deep features from pixel-level hyperspectral data of grapes, and then these features were used as inputs to establish support vector machine (SVM) models for estimating grape shelf-life. Furthermore, SVM, one-dimensional convolutional neural network (1D CNN) and long short-term memory (LSTM) models were used as traditional machine learning and end to end models for comparison. The results demonstrated that the SDAE-SVM model achieved reasonable recognition accuracy of 100 % and 98.125 % for the shelf-life of grapes in the training and test sets, respectively. The overall results suggested that SDAE-based deep learning method can be used as a powerful tool to deal with large-scale hyperspectral data as well as this research confirms the feasibility of non-destructive estimation for grapes shelf-life by the combination of HSI technique and deep learning method, which would provide a valuable guidance for shelf-life estimation of other postharvest fruit.

Research and performance analysis of random forest-based feature selection algorithm in sports effectiveness evaluation

The rapid progress in fields such as data mining and machine learning, as well as the explosive growth of sports big data, have posed new challenges to the research of sports big data. Most of the available sports data mining techniques concentrates on extracting and constructing effective features for basic sports data, which cannot be achieved simply by using data statistics. Especially in the targeted mining of sports data, traditional mining techniques still have shortcomings such as low classification accuracy and insufficient refinement. In order to solve the problem of low accuracy in traditional mining methods, the study combines the random forest algorithm with the artificial raindrop algorithm, and adopts a sports data mining method based on feature selection to achieve effective analysis of sports big data. This study is based on the evaluation method of motion effects using random forests, and uses feature extraction algorithms to study the motion effect impacts. It uses the information gain index to rank the importance of features and accurately gain the degree of influence of exercise on various indicators of the human body. Through simulation verification, the algorithm proposed by the research institute performs the best in accuracy and FI scores on the training and testing sets, with accuracies of 0.849 ± 0.021 and 0.819 ± 0.022, respectively, and F1 scores of 0.837 ± 0.020 and 0.864 ± 0.021, respectively. This indicates that the algorithm proposed by the research institute has high classification accuracy and performance proves that the Random Forest-based feature selection algorithm established in this study is superior to the existing traditional feature extraction and extraction methods in terms of both performance and accuracy. The proposal of this data analysis method has achieved accurate and efficient utilization of sports big data, which is of great significance for the development of the sports education industry.

Integrating Multisource Data and Machine Learning for Supraglacial Lake Detection: Implications for Environmental Management and Sustainable Development Goals in High Mountainous Regions

The accurate detection and monitoring of supraglacial lakes in high mountainous regions are crucial for understanding their dynamic nature and implications for environmental management and sustainable development goals. In this study, we propose a novel approach that integrates multisource data and machine learning techniques for supra-glacial lake detection in the Passu Batura glacier of the Hunza Basin, Pakistan. We extract pertinent features or parameters by leveraging multisource datasets such as radar backscatter intensity VH and VV parameters from Sentinel-1 Ground Range Detected (GRD) data, near-infrared (NIR), NDWI_green, NDWI_blue parameters from Sentinel-2 Multi-spectral Instrument(MSI) data, and surface slope, aspect, and elevation parameters from topographic data. The entire dataset is partitioned into training and testing sets, with machine learning models including the artificial neural network (ANN), the support vector machine (SVM), logistic regression (LR), random forest (RF), and K-nearest neighbour (KNN) trained on the training data (70%). Accuracy assessment employs testing data and involves the evaluation of metrics such as ROC curves and confusion matrices. The best-performing model, ANN, is validated against manually digitized lake polygons derived from Sentinel-2 and Google Earth Pro imagery. Furthermore, the digitized lake polygons are used to analyze glacial lake dynamics from 2016 to 2022. Key findings of this research presented that the NDWI_green, Sigma0_VH, and elevation are the most significant predictors in detecting supra-glacial lakes. Among the various trained and evaluated models, the Artificial Neural Network (ANN) achieved the highest performance (accuracy: 95%, AUC: 0.99) and accurately mapped supra-glacial lakes regardless of their small size. The findings have significant implications for understanding glacial lake behavior in the context of climate change and informing future research and monitoring efforts.

Radiomics nomogram combined with clinical factors for predicting pathological complete response in resectable esophageal squamous cell carcinoma

IntroductionPredicting the efficacy of neoadjuvant immunochemotherapy (NICT) for esophageal squamous cell carcinoma (ESSC) prior to surgery can minimize unnecessary surgical interventions and facilitate personalized treatment strategies. Our goal is to develop and validate an image-based radiomic model using preoperative computed tomography (CT) scans and clinical data to predict pathological complete response (pCR) in resectable ESSC following neoadjuvant immunotherapy.MethodsWe retrospectively collected data from patients diagnosed with ESCC at the First Affiliated Hospital of Soochow University between January 2018 and May 2023, who received preoperative neoadjuvant immunochemotherapy. Eligible patients were randomly divided into training and validation sets. Radiomic features extracted from preprocessed CT images were used to develop a radiomic model, incorporating Radiomic score (Rad-score) and clinical factors through multivariate logistic regression analysis. The model’s performance was assessed for calibration, discrimination, and clinical utility in an independent validation cohort.ResultsWe enrolled a total of 105 eligible participants who were randomly divided into two groups: a training set (N=74) and a validation set (N=31). After data dimension reduction and feature selection, we identified 11 radiomic features, which collectively formed the Rad-score. Rad-score had an area under the curve (AUC) of 0.83 (95% CI 0.72-0.93) in the training set and 0.78 (95% CI 0.60-0.95) in the validation set. Multivariate analysis revealed that radiological response and Neutrophil–Lymphocyte Ratio (NLR) were independent predictors of pCR, with p-values of 0.0026 and 0.0414, respectively. We developed and validated a nomogram combining Rad-score and clinical features, achieving AUCs of 0.90 (95% CI 0.82-0.98) in the training set and 0.85 (95% CI 0.70-0.99) in the validation set. The Delong test confirmed the nomogram’s superiority over pure radiomic and clinical models. Decision curve analysis (DCA) and integrated discrimination improvement (IDI) assessment supported the clinical value and superiority of the combined model.ConclusionThe nomogram, which integrates Rad-score and clinical features, offers a precise and reliable method for predicting pCR status in ESCC patients who have undergone neoadjuvant immunochemotherapy. This tool aids in tailoring treatment strategies to individual patients.

Construction of a combined prognostic model for pancreatic ductal adenocarcinoma based on deep learning and digital pathology images

BackgroundDeep learning has made significant advancements in the field of digital pathology, and the integration of multiple models has further improved accuracy. In this study, we aimed to construct a combined prognostic model using deep learning-extracted features from digital pathology images of pancreatic ductal adenocarcinoma (PDAC) alongside clinical predictive indicators and to explore its prognostic value.MethodsA retrospective analysis was conducted on 142 postoperative pathologically confirmed PDAC cases. These cases were divided into training (n = 114) and testing sets (n = 28) at an 8:2 ratio. Tumor whole-slide imaging features were extracted and screened to construct a pathological risk model based on a pre-trained deep learning model. Clinical and pathological data from the training set were used to select independent predictive factors for PDAC and establish a clinical risk model using LASSO, univariate, and multivariate Cox regression analyses. Based on the pathological and clinical risk models, a combined model was developed. The Harrell concordance index (C-index) was computed to assess the predictive performance of each model for PDAC survival prognosis.ResultsFor the training and testing sets, the C-index values for the clinical risk model were 0.76 and 0.75, respectively; for the pathological risk model, they were 0.82 and 0.73, respectively; and for the combined model, they were 0.86 and 0.77, respectively. The combined model exhibited appropriate calibration at 1-, 3-, and 5-year time points, as well as a superior area under the curve of the receiver operating characteristic curve and clinical net benefit compared to the single models.ConclusionsIntegrating the pathological and clinical risk models may provide a higher predictive value for survival prognosis.

Prediction of femoral head collapse in osteonecrosis using deep learning segmentation and radiomics texture analysis of MRI

BackgroundFemoral head collapse is a critical pathological change and is regarded as turning point in disease progression in osteonecrosis of the femoral head (ONFH). In this study, we aim to build an automatic femoral head collapse prediction pipeline for ONFH based on magnetic resonance imaging (MRI) radiomics.MethodsIn the segmentation model development dataset, T1-weighted MRI of 222 hips from two hospitals were retrospectively collected and randomly split into training (n = 190) and test (n = 32) sets. In the prognosis prediction model development dataset, 206 hips were also retrospectively collected from two hospitals and divided into training set (n = 155) and external test set (n = 51) according to data source. A deep learning model for automatic lesion segmentation was trained with nnU-Net, from which three-dimensional regions of interest were segmented and a total of 107 radiomics features were extracted. After intra-class correlation coefficients screening, feature correlation coefficient screening and Least Absolute Shrinkage and Selection Operator regression feature selection, a machine learning model for ONFH prognosis prediction was trained with Logistic Regression (LR) and Light Gradient Boosting Machine (LightGBM) algorithm.ResultsThe segmentation model achieved an average dice similarity coefficient of 0.848 and an average 95% Hausdorff distance of 3.794 in the test set, compared to the manual segmentation results. After feature selection, nine radiomics features were included in the prognosis prediction model. External test showed that the LightGBM model exhibited acceptable predictive performance. The area under the curve (AUC) of the prediction model was 0.851 (95% CI: 0.7268–0.9752), with an accuracy of 0.765, sensitivity of 0.833, and specificity of 0.727. Decision curve analysis showed that the LightGBM model exhibited favorable clinical utility.ConclusionThis study presents an automated pipeline for predicting femoral head collapse in ONFH with acceptable performance. Further research is necessary to determine the clinical applicability of this radiomics-based approach and to assess its potential to assist in treatment decision-making for ONFH.

Enhancing the differential diagnosis of small pulmonary nodules: a comprehensive model integrating plasma methylation, protein biomarkers, and LDCT imaging features

BackgroundAccurate differentiation between malignant and benign pulmonary nodules, especially those measuring 5–10 mm in diameter, continues to pose a significant diagnostic challenge. This study introduces a novel, precise approach by integrating circulating cell-free DNA (cfDNA) methylation patterns, protein profiling, and computed tomography (CT) imaging features to enhance the classification of pulmonary nodules.MethodsBlood samples were collected from 419 participants diagnosed with pulmonary nodules ranging from 5 to 30 mm in size, before any disease-altering procedures such as treatment or surgical intervention. High-throughput bisulfite sequencing was used to conduct DNA methylation profiling, while protein profiling was performed utilizing the Olink proximity extension assay. The dataset was divided into a training set and an independent test set. The training set included 162 matched cases of benign and malignant nodules, balanced for sex and age. In contrast, the test set consisted of 46 benign and 49 malignant nodules. By effectively integrating both molecular (DNA methylation and protein profiling) and CT imaging parameters, a sophisticated deep learning-based classifier was developed to accurately distinguish between benign and malignant pulmonary nodules.ResultsOur results demonstrate that the integrated model is both accurate and robust in distinguishing between benign and malignant pulmonary nodules. It achieved an AUC score 0.925 (sensitivity = 83.7%, specificity = 82.6%) in classifying test set. The performance of the integrated model was significantly higher than that of individual methylation (AUC = 0.799, P = 0.004), protein (AUC = 0.846, P = 0.009), and imaging models (AUC = 0.866, P = 0.01). Importantly, the integrated model achieved a higher AUC of 0.951 (sensitivity = 83.9%, specificity = 89.7%) in 5–10 mm small nodules. These results collectively confirm the accuracy and robustness of our model in detecting malignant nodules from benign ones.ConclusionsOur study presents a promising noninvasive approach to distinguish the malignancy of pulmonary nodules using multiple molecular and imaging features, which has the potential to assist in clinical decision-making.Trial registration: This study was registered on ClinicalTrials.gov on 01/01/2020 (NCT05432128). https://classic.clinicaltrials.gov/ct2/show/NCT05432128.

Construction and Validation of a Prognostic Prediction Model for Gastric Mucinous Adenocarcinoma Based on the SEER Database

This study focused on the development and validation of a prognostic model to predict overall survival (OS) in patients with gastric mucinous adenocarcinoma (GMA), utilizing data from the SEER database. A cohort of 537 GMA patients was analyzed to identify factors affecting survival, using both univariate and multivariate Cox regression analyses on demographic and clinicopathological variables. The analysis led to the creation of a nomogram model. Findings revealed that TNM staging and the administration of radiotherapy or chemotherapy were significant factors impacting survival outcomes. Specifically, patients at stages T4, N3, and M1 faced a notably higher mortality risk, while those who received radiotherapy and chemotherapy experienced a significant reduction in mortality risk. The model's performance, as indicated by C-indices of 0.73 in the training set and 0.66 in the validation set, suggests robust predictive capability. The area under the curve (AUC) analysis confirmed high accuracy in predicting 1-year, 3-year, and 5-year survival rates, and calibration curves displayed strong agreement between predicted and actual outcomes. Decision curve analysis (DCA) demonstrated that the model exhibited significant clinical utility across various thresholds. Overall, this study presents a reliable prognostic tool for individualized risk stratification and clinical treatment in GMA patients. Future work should focus on enhancing the model by incorporating molecular biological data to increase its predictive accuracy and applicability.

Innovative label-free lymphoma diagnosis using infrared spectroscopy and machine learning on tissue sections

The diagnosis of lymphomas is challenging due to their diverse histological presentations and clinical manifestations. There is a need for inexpensive tools that require minimal expertise and are accessible for routine laboratories. Contrastingly, current conventional diagnostic methods are often found only in specialized environments. Attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectroscopy offers a nondestructive and user-friendly approach in the analysis of a wide range of samples. In this paper, we determined whether the technique coupled with machine learning can detect and differentiate lymphoma within lymphoid tissue samples. Tissue sections from 295 individuals diagnosed with lymphoma and 389 individuals without the disease were analyzed using ATR-FTIR spectroscopy. The resulting spectral dataset was split using a 70:30 train-test split. Partial least Squares Discriminant Analysis (PLS-DA) models were trained to distinguish non-malignant lymphoid tissue from lymphoma samples and to differentiate between subtypes. On the training set (n = 478), significant spectral differences were mainly identified in the 1800–900 cm–1 region, attributed to fundamental biochemical constituents like proteins, lipids, carbohydrates, and nucleic acids. On the independent test set (n = 206), the trained PLS-DA model achieved a promising AUC of 0.882 (95% CI: 0.881–0.884) in the differentiation between lymphoma and non-malignant lymphoid tissue. In addition, comparative analyses revealed spectral distinctions and notable clustering between the different lymphoma subtypes. This study provides valuable insights into the application of ATR-FTIR spectroscopy and machine learning in the field of lymphoma diagnosis as a non-destructive, rapid and inexpensive tool with the potential to be easily implemented in non-specialized laboratories.

Development of a Practical Prediction Model for Adverse Neonatal Outcomes at the Start of the Second Stage of Labor.

To develop a prediction model for adverse neonatal outcomes using electronic fetal monitoring (EFM) interpretation data and other relevant clinical information known at the start of the second stage of labor. This was a retrospective cohort study of individuals who labored and delivered at two academic medical centers between July 2016 and June 2020. Individuals were included if they had a singleton gestation at term (more than 37 weeks of gestation), a vertex-presenting, nonanomalous fetus, and planned vaginal delivery and reached the start of the second stage of labor. The primary outcome was a composite of severe adverse neonatal outcomes. We developed and compared three modeling approaches to predict the primary outcome using factors related to EFM data (as interpreted and entered in structured data fields in the electronic health record by the bedside nurse), maternal comorbidities, and labor characteristics: traditional logistic regression, LASSO (least absolute shrinkage and selection operator), and extreme gradient boosting. Model discrimination and calibration were compared. Predicted probabilities were stratified into risk groups to facilitate clinical interpretation, and positive predictive values for adverse neonatal outcomes were calculated for each. A total of 22,454 patients were included: 14,820 in the training set and 7,634 in the test set. The composite adverse neonatal outcome occurred in 3.2% of deliveries. Of the three modeling methods compared, the logistic regression model had the highest discrimination (0.690, 95% CI, 0.656-0.724) and was well calibrated. When stratified into risk groups (no increased risk, higher risk, and highest risk), the rates of the composite adverse neonatal outcome were 2.6% (95% CI, 2.3-3.1%), 6.7% (95% CI, 4.6-9.6%), and 10.3% (95% CI, 7.6-13.8%), respectively. Factors with the strongest associations with the composite adverse neonatal outcome included the presence of meconium (adjusted odds ratio [aOR] 2.10, 95% CI, 1.68-2.62), fetal tachycardia within the 2 hours preceding the start of the second stage (aOR 1.94, 95% CI, 1.03-3.65), and number of prior deliveries (aOR 0.77, 95% CI, 0.60-0.99).

A large scale multi institutional study for radiomics driven machine learning for meningioma grading

This study aims to develop and evaluate radiomics-based machine learning (ML) models for predicting meningioma grades using multiparametric magnetic resonance imaging (MRI). The study utilized the BraTS-MEN dataset’s training split, including 698 patients (524 with grade 1 and 174 with grade 2–3 meningiomas). We extracted 4872 radiomic features from T1, T1 with contrast, T2, and FLAIR MRI sequences using PyRadiomics. LASSO regression reduced features to 176. The data was split into training (60%), validation (20%), and test (20%) sets. Five ML algorithms (TabPFN, XGBoost, LightGBM, CatBoost, and Random Forest) were employed to build models differentiating low-grade (grade 1) from high-grade (grade 2–3) meningiomas. Hyperparameter tuning was performed using Optuna, optimizing model-specific parameters and feature selection. The CatBoost model demonstrated the best performance, achieving an area under the receiver operating characteristic curve (AUROC) of 0.838 [95% confidence interval (CI): 0.689–0.935], precision of 0.492 (95% CI: 0.371–0.623), recall of 0.838 (95% CI: 0.689–0.935), F1 score of 0.620 (95% CI: 0.495–0.722), accuracy of 0.729 (95% CI: 0.650–0.800), an area under the precision-recall curve (AUPRC) of 0.620 (95% CI: 0.433–0.753), and Brier score of 0.156 (95% CI: 0.122-0.200). Other models showed comparable performance, with mean AUROCs ranging from 0.752 to 0.784. The radiomics-based ML approach presented in this study showcases the potential for non-invasive and pre-operative grading of meningiomas using multiparametric MRI. Further validation on larger and independent datasets is necessary to establish the robustness and generalizability of these findings.

Circular rubber aggregate CFST stub columns under axial compression: prediction and reliability analysis

Extensive studies support using steel tubes to enhance the structural integrity of rubber aggregate concrete (RBAC), namely RBAC-filled steel tubes (RCFST). However, current design codes for assessing the axial compressive behaviour of circular stub RCFST (CS-RCFST) columns are limited. Furthermore, there is a scarcity of studies focused on ensuring the structural safety of these columns. Based on an extensive experimental database comprising 145 columns, this study explores machine learning (ML) capabilities for predicting the axial strength of CS-RCFST columns, using six typical machine-learning models, i.e., symbolic regression (SR), XGBoost, CatBoost, random forest, LightGBM, and Gaussian process regression models. The hyperparameter tuning of the introduced ML models is performed using the Bayesian Optimization technique. The comparison results show that the CatBoost model is the most reliable and accurate ML model (R2 = 0.999 and 0.993 for the training and testing sets, respectively). In addition, a simple and practical design expression for CS-RCFST columns has been developed with acceptable accuracy based on the SR model (an average test-to-prediction ratio of 0.99 and CoV of 0.132). Meanwhile, the axial strength predicted by ML models was compared with two prominent practice codes (i.e., AISC360 and EC4). The comparison results indicated that the ML models could introduce a highly reliable and accurate approach over current design standards for strength prediction. Furthermore, a reliability analysis is conducted on two different ML models to evaluate the reliability of utilising ML models in practical design applications. This assessment involves identifying the statistical properties associated with the compressive strength of RBAC, as well as introducing the required resistance design factors aligned with the target reliability recommended by code standards.

Clinical information extraction for lower-resource languages and domains with few-shot learning using pretrained language models and prompting

Abstract A vast amount of clinical data are still stored in unstructured text. Automatic extraction of medical information from these data poses several challenges: high costs of clinical expertise, restricted computational resources, strict privacy regulations, and limited interpretability of model predictions. Recent domain adaptation and prompting methods using lightweight masked language models showed promising results with minimal training data and allow for application of well-established interpretability methods. We are first to present a systematic evaluation of advanced domain-adaptation and prompting methods in a lower-resource medical domain task, performing multi-class section classification on German doctor’s letters. We evaluate a variety of models, model sizes (further-pre)training and task settings, and conduct extensive class-wise evaluations supported by Shapley values to validate the quality of small-scale training data and to ensure interpretability of model predictions. We show that in few-shot learning scenarios, a lightweight, domain-adapted pretrained language model, prompted with just 20 shots per section class, outperforms a traditional classification model, by increasing accuracy from $48.6\%$ to $79.1\%$ . By using Shapley values for model selection and training data optimization, we could further increase accuracy up to $84.3\%$ . Our analyses reveal that pretraining of masked language models on general-language data is important to support successful domain-transfer to medical language, so that further-pretraining of general-language models on domain-specific documents can outperform models pretrained on domain-specific data only. Our evaluations show that applying prompting based on general-language pretrained masked language models combined with further-pretraining on medical-domain data achieves significant improvements in accuracy beyond traditional models with minimal training data. Further performance improvements and interpretability of results can be achieved, using interpretability methods such as Shapley values. Our findings highlight the feasibility of deploying powerful machine learning methods in clinical settings and can serve as a process-oriented guideline for lower-resource languages and domains such as clinical information extraction projects.