Assessment of shape-based features ability to predict the ascending aortic aneurysm growth

Hypoxemia is a common complication associated with anesthesia in painless gastroscopy. With the aging of the social population, the number of cases of hypoxemia among middle-aged and elderly patients is increasing. However, tools for predicting hypoxemia in middle-aged and elderly patients are lacking. In this study, we investigated the risk factors for hypoxemia in middle-aged and elderly outpatients undergoing painless gastroscopy based on machine learning and constructed a risk prediction model. In this retrospective study, we included the data on 1,348 outpatients undergoing painless gastroscopy. In total, 26 characteristic variables, including demographic information, past medical history, and clinical data of the patients were included, and BorutaShap was used for feature selection. Five machine learning algorithm models, including logistic regression (LR), support vector machine (SVM), random forest (RF), extreme gradient boosting (XGB), and light gradient boosting machine (LightGBM), were selected. The best models were selected based on the area under the receiver operating characteristic curve (AUROC). Model feature importance was explained and analyzed using Shapley Additive Explanations (SHAP). The endpoint event of this study was considered to be hypoxemia during the procedure, defined as at least one occurrence of pulse oxygen saturation below 90% without probe misalignment or interference from the beginning of anesthesia induction to the end of painless gastroscopy. In the final cohort of 984 patients, 11% of patients (108/984) experienced hypoxemia during the painless gastroscopy procedure. The AUROCs of the five models were as follows: Logistic Regression (AUROC = 0.893, 95CI: 0.881–0.899), SVM (AUROC = 0.855, 95CI: 0.812–0.884), Random Forest (AUROC = 0.914, 95CI: 0.889–0.924), XGB (AUROC = 0.902, 95CI: 0.865–0.919), and LightGBM (AUROC = 0.891, 95CI: 0.847–0.917). Regarding the explanation of the importance of SHAP features, preoperative variables (baseline SpO2, body mass index, and micrognathia) and intraoperative variables (operating time of gastroscopy, induction dose of etomidate and propofol mixture, append anesthetic, cough, and repeated pharyngeal irritation) significantly contributed to the model. We identified eight potential risk factors related to the occurrence of hypoxemia in middle-aged and elderly patients undergoing painless gastroscopy, based on machine learning feature engineering. Among the five machine learning algorithms, RF exhibited the best predictive performance in the internal test set and had a certain degree of generalization ability in the external validation set, which indicated that the RF model was more suitable for the data framework of this study. This model was more likely to enhance the accuracy of hypoxemia prediction in middle-aged and elderly patients undergoing painless gastroscopy, and thus, it is suitable for assisting anesthesiologists in clinical decision-making.

Unruptured intracranial aneurysms (UIAs) affect 3.2% of the general population, and approximately 85% of subarachnoid hemorrhages result from their rupture. Despite their classification as low risk by prediction tools such as PHASES (population, hypertension, age, size of aneurysm, earlier subarachnoid hemorrhage from another aneurysm, and site of aneurysm) and the Unruptured Cerebral Aneurysm Study (UCAS), UIAs less than 10 mm are susceptible to rupture. To develop and externally validate a machine-learning model (MLM) predicting rupture risk of UIAs. This retrospective multicenter prognostic study analyzed UIAs from 4 institutions across 3 continents from January 2003 to November 2022. Each UIA was characterized by 29 clinical and 18 morphological variables. For model development, patients with UIAs were drawn from a large institutional cohort. Statistical analysis was performed from April 2024 to March 2025. An MLM based on the Light Gradient Boosting Machine algorithm was trained, and performance was assessed for validation externally. The primary outcome was aneurysm rupture within 2 years after risk evaluation. Model performance was assessed using sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), negative likelihood ratio (NLR), and the area under the receiver operating characteristic curve (AUROC) with 95% CIs. Drawing from 11 579 UIAs from multiple institutions of 8846 patients, there were 2750 patients with 3312 UIAs in the development cohort (median [IQR] age, 65.3 [54.9-73.6] years; 1856 females [67.5%]) and 1153 patients with 1501 UIAs in the external cohort (median [IQR] age, 63.6 [53.9-70.9] years; 828 females [71.8%]), for whom the MLM demonstrated a robust performance in risk estimation. In the development cohort, 71 UIAs (2.1%) ruptured during 8.5 years' median follow-up (IQR, 5.1-11.6 years), and in the external cohort, 48 UIAs (3.2%) ruptured during a median (IQR) follow-up of 5.4 (3.7-8.7) years. With a threshold of 0.52 for the MLM, performance in the development cohort for sensitivity was 0.78 (95% CI, 0.67-0.86); specificity, 0.82 (95% CI, 0.80-0.83); PPV, 0.09 (95% CI, 0.07-0.11); NPV, 0.99 (95% CI, 0.99-1.00); PLR, 4.23 (95% CI, 3.66-4.89); NLR, 0.28 (95% CI, 0.18-0.42); and AUROC, 0.88 (95% CI, 0.85-0.92) and in the external cohort, sensitivity was 0.90 (95% CI, 0.78-0.95); specificity, 0.70 (95% CI, 0.67-0.72); PPV, 0.09 (95% CI, 0.07-0.12); NPV, 1.00 (95% CI, 0.99-1.00); PLR, 2.94 (95% CI, 2.60-3.33); NLR, 0.15 (95% CI, 0.07-0.34); and AUROC, 0.90 (95% CI, 0.85-0.94). The MLM performed consistently in UIAs less than 10 mm (AUROC, 0.88 [95% CI, 0.82-0.94]), suggesting potential complementary value to PHASES and the UCAS. In this prognostic study, the MLM consistently identified features of UIAs as significantly associated with rupture across different cohorts. These results support the MLM's potential to assist physicians and patients in UIA treatment decisions.

Magnetic resonance imaging (MRI) paired with appropriate disease-specific machine learning holds promise for the clinical differentiation of Parkinson disease (PD), multiple system atrophy (MSA) parkinsonian variant, and progressive supranuclear palsy (PSP). A prospective study is needed to test whether the approach meets primary end points to be considered in a diagnostic workup. To assess the discriminative performance of Automated Imaging Differentiation for Parkinsonism (AIDP) using 3-T diffusion MRI and support vector machine (SVM) learning. This was a prospective, multicenter cohort study conducted from July 2021 to January 2024 across 21 Parkinson Study Group sites (US/Canada). Included were patients with PD, MSA, and PSP with established criteria and unanimous agreement in the clinical diagnosis among 3 independent, blinded neurologists who specialize in movement disorders. Patients were assigned to a training set or an independent testing set. MRI. Area under the receiver operating characteristic curve (AUROC) in the testing set for primary model end points of PD vs atypical parkinsonism, MSA vs PSP, PD vs MSA, and PD vs PSP. AIDP was also paired with antemortem MRI to test against postmortem neuropathology in a subset of autopsy cases. A total of 316 patients were screened and 249 patients (mean [SD] age, 67.8 [7.7] years; 155 male [62.2%]) met inclusion criteria. Of these patients, 99 had PD, 53 had MSA, and 97 had PSP. A retrospective cohort of 396 patients (mean [SD] age, 65.8 [8.9] years; 234 male [59.1%]) was also included. Of these patients, 211 had PD, 98 had MSA, and 87 had PSP. Patients were assigned to the training set (78%; 104 prospective, 396 retrospective) or independent testing set, which included 145 (22%; 60 PD, 27 MSA, 58 PSP) prospective patients (mean age, 67.4 [SD 7.7] years; 95 male [65.5%]). The model was robust in differentiating PD vs atypical parkinsonism (AUROC, 0.96; 95% CI, 0.93-0.99; positive predictive value [PPV], 0.91; negative predictive value [NPV], 0.83), MSA vs PSP (AUROC, 0.98; 95% CI, 0.96-1.00; PPV, 0.98; NPV, 0.81), PD vs MSA (AUROC, 0.98; 95% CI, 0.96-1.00; PPV, 0.97; NPV, 0.97), and PD vs PSP (AUROC, 0.98; 95% CI, 0.96-1.00; PPV, 0.92; NPV, 0.98). AIDP predictions were confirmed neuropathologically in 46 of 49 brains (93.9%). This prospective multicenter cohort study of AIDP met its primary end points. Results suggest using AIDP in the diagnostic workup for common parkinsonian syndromes.

Autism spectrum disorder (ASD) affects social interaction, communication, and behavior. Early diagnosis is important as it enables timely intervention that can significantly improve long-term outcomes, but current diagnostic, which rely heavily on behavioral observations and clinical interviews, are often subjective and time-consuming. This study introduces an AI-based approach that uses T1-weighted structural MRI (sMRI) scans, network analysis, and vision transformers to automatically diagnose ASD. sMRI data from 79 ASD patients and 105 healthy controls were obtained from the Autism Brain Imaging Data Exchange (ABIDE) database. Complex network analysis (CNA) features and ViT (Vision Transformer) features were developed for predicting ASD. Five models were developed for each type of features: logistic regression, support vector machine (SVM), gradient boosting (GB), K-nearest neighbors (KNN), and neural network (NN). 25 models were further developed by federating the two sets of 5 models. Model performance was evaluated using accuracy, area under the receiver operating characteristic curve (AUC-ROC), sensitivity, and specificity via fivefold cross-validation. The federate model CNA(KNN)-ViT(NN) achieved highest performance, with accuracy 0.951 ± 0.067, AUC-ROC 0.980 ± 0.020, sensitivity 0.963 ± 0.050, and specificity 0.943 ± 0.047. The performance of the ViT-based models exceeds that of the complex network-based models on 80% of the performance metrics. By federating CNA models, the ViT models can achieve better performance. This study demonstrates the feasibility of using CNA and ViT models for the automated diagnosis of ASD. The proposed CNA(KNN)-ViT(NN) model achieved better accuracy in ASD classification based solely on T1 sMRI images. The proposed method's reliance on widely available T1 sMRI scans highlights its potential for integration into routine clinical examinations, facilitating more efficient and accessible ASD screening.

BackgroundIschemic stroke remains one of the leading causes of morbidity and mortality worldwide. NT-proBNP is emerging as a promising biomarker for risk stratification in acute ischemic stroke. This review aimed to assess its prognostic accuracy for mortality in this patient population.MethodsWe conducted a comprehensive literature search across databases (PubMed, Embase, Web of Science, Scopus and Cochrane Library), retrieving 4832 records. Following duplicate removal and full-text screening, 11 prospective observational studies meeting the inclusion criteria were included. Data on study characteristics, risk of bias, and diagnostic performance (sensitivity, specificity, likelihood ratios, diagnostic odds ratio, and area under the ROC curve [AUROC]) were extracted and pooled using a bivariate random-effects model.ResultsThe 11 studies included 2994 patients (12.19% mortality). Pooled prognostic performance of NT‐proBNP for mortality showed a sensitivity of 0.83 (95% CI, 0.73–0.89) and specificity of 0.77 (95% CI, 0.67–0.84), with an area under the receiver operating characteristic curve (AUROC) of 0.87 (95% CI, 0.84–0.90). In subgroup analyses, NT‐proBNP demonstrated a sensitivity of 0.84 and specificity of 0.70 (AUROC = 0.86) for predicting in-hospital mortality (5 studies), and sensitivity of 0.81 and specificity of 0.87 (AUROC = 0.91) for predicting 3‐month mortality (4 studies). The positive likelihood ratio was 3.5, and the negative likelihood ratio was 0.23, corresponding to a diagnostic odds ratio of 16.ConclusionsNT-proBNP demonstrates moderate to high accuracy for predicting mortality in acute ischemic stroke, supporting its potential role in clinical risk stratification. Further standardized, multicenter research is warranted to confirm optimal threshold values and to evaluate its impact on patient management.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12916-025-04299-1.

Acute kidney injury (AKI) after liver transplantation has been reported to be associated with increased mortality. Recently, machine learning approaches were reported to have better predictive ability than the classic statistical analysis. We compared the performance of machine learning approaches with that of logistic regression analysis to predict AKI after liver transplantation. We reviewed 1211 patients and preoperative and intraoperative anesthesia and surgery-related variables were obtained. The primary outcome was postoperative AKI defined by acute kidney injury network criteria. The following machine learning techniques were used: decision tree, random forest, gradient boosting machine, support vector machine, naïve Bayes, multilayer perceptron, and deep belief networks. These techniques were compared with logistic regression analysis regarding the area under the receiver-operating characteristic curve (AUROC). AKI developed in 365 patients (30.1%). The performance in terms of AUROC was best in gradient boosting machine among all analyses to predict AKI of all stages (0.90, 95% confidence interval [CI] 0.86–0.93) or stage 2 or 3 AKI. The AUROC of logistic regression analysis was 0.61 (95% CI 0.56–0.66). Decision tree and random forest techniques showed moderate performance (AUROC 0.86 and 0.85, respectively). The AUROC of support the vector machine, naïve Bayes, neural network, and deep belief network was smaller than that of the other models. In our comparison of seven machine learning approaches with logistic regression analysis, the gradient boosting machine showed the best performance with the highest AUROC. An internet-based risk estimator was developed based on our model of gradient boosting. However, prospective studies are required to validate our results.

BackgroundBacterial infection has been challenging to diagnose in patients with COVID-19. We developed and evaluated supervised machine learning algorithms to support the diagnosis of secondary bacterial infection in hospitalized patients during the COVID-19 pandemic.MethodsInpatient data at three London hospitals for the first COVD-19 wave in March and April 2020 were extracted. Demographic, blood test and microbiology data for individuals with and without SARS-CoV-2-positive PCR were obtained. A Gaussian Naive Bayes, Support Vector Machine (SVM) and Artificial Neural Network were trained and compared using the area under the receiver operating characteristic curve (AUCROC). The best performing algorithm (SVM with 21 blood test variables) was prospectively piloted in July 2020. AUCROC was calculated for the prediction of a positive microbiological sample within 48 h of admission.ResultsA total of 15 599 daily blood profiles for 1186 individual patients were identified to train the algorithms; 771/1186 (65%) individuals were SARS-CoV-2 PCR positive. Clinically significant microbiology results were present for 166/1186 (14%) patients during admission. An SVM algorithm trained with 21 routine blood test variables and over 8000 individual profiles had the best performance. AUCROC was 0.913, sensitivity 0.801 and specificity 0.890. Prospective testing on 54 patients on admission (28/54, 52% SARS-CoV-2 PCR positive) demonstrated an AUCROC of 0.960 (95% CI: 0.90–1.00).ConclusionsAn SVM using 21 routine blood test variables had excellent performance at inferring the likelihood of positive microbiology. Further prospective evaluation of the algorithms ability to support decision making for the diagnosis of bacterial infection in COVID-19 cohorts is underway.

Patients with non-small cell lung cancer (NSCLC) and patients with NSCLC combined with chronic obstructive pulmonary disease (COPD) have similar physiological conditions in early stages, and the latter have shorter survival times and higher mortality rates. The purpose of this study was to develop and compare machine learning models to identify future diagnoses of COPD combined with NSCLC patients based on the patient's disease and routine clinical data. Data were obtained from 237 patients with COPD combined with NSCLC as well as NSCLC admitted to Ningxia Hui Autonomous Region People's Hospital from October 2013 to July 2022. Six machine learning algorithms (K-nearest neighbor, logistic regression, eXtreme gradient boosting, support vector machine, naïve Bayes, and artificial neural network) were used to develop prediction models for NSCLC combined with COPD. Sensitivity, specificity, positive predictive value, negative predictive value, accuracy, F1 score, Mathews correlation coefficient (MCC), Kappa, area under the receiver operating characteristic curve (AUROC)and area under the precision-recall curve (AUPRC) were used as performance indicators to evaluate the performance of the models. 135 patients with NSCLC combined with COPD, 102 patients with NSCLC were included in the study. The results showed that pulmonary function and emphysema were important risk factors and that the support vector machine-based identification model showed optimal performance with accuracy:0.946, recall:0.940, specificity:0.955, precision:0.972, npv:0.920, F1 score:0.954, MCC:0.893, Kappa:0.888, AUROC:0.975, AUPRC:0.987. The use of machine learning tools combining clinical symptoms and routine examination data features is suitable for identifying the risk of concurrent NSCLC in COPD patients.

Objective To investigate the predictive value of cardial troponin I for right ventricular dysfunction (RVD) in cancer and non-cancer patients with acute pulmonary embolism. Methods The clinical data of 196 patients with acute pulmonary embolism admitted in recent years were retrospectively analyzed. The patients were divided into cancer group(45 cases) and non-cancer groups(151 cases). Receiver operating characteristic (ROC) curve and area under the ROC curve (AUC) for cTnI predicting RVD were analyzed and related cutoff levels for both groups were intercepted. Results In the cancer group, analysis of the ROC curves showed an AUC of 0.82 (95%CI 0.70-0.96) for cTnI predicting RVD with an optimal cutoff value of 0.047 ng/ml, the sensitivity and specificity of the test was 79% and 86%, the negative likelihood ratio was 0.244. In the non-cancer group, AUC was 0.85(95%CI 0.79-0.91) for cTnI predicting RVD with an optimal cutoff value of 0.024 ng/ml, the sensitivity and specificity of the test was 86% and 72%, the negative likelihood ratio was 0.194. Conclusions cTnI is effective for predicting RVD in acute pulmonary embolism patients. However, the effectiveness of cTnI to predict RVD is higher in acute pulmonary embolism patients with non-cancer than that in patients with cancer. Key words: Acute pulmonary embolism; Cancer; Cardial troponin I; Right ventricular dysfunction

<title>Abstract</title> Objective The aims of this study were: ① to evaluate the diagnostic efficacy of six mainstream TI-RADS (Thyroid Imaging Reporting and Data System) classification systems (C-TIRADS, ACR-TIRADS, etc.) in the Northwestern Chinese population; and ② to identify risk factors for malignant thyroid nodules (TNs) using logistic regression based on clinical and ultrasound features, construct a quantifiable scoring Nomogram model, enable rapid and objective risk assessment, and assist in clinical decision-making. Methods A total of 2,047 patients with TNs (1,433 malignant and 614 benign) were enrolled from January 2018 to January 2024 at Shaanxi Provincial People’s Hospital. The nodules were divided into a training group (1,435 nodules) and a validation group (612 nodules) in a 7:3 ratio. Twelve characteristics were collected, including age, nodule size, margin, calcification, and the presence of suspicious lymph nodes. Independent risk factors were identified through univariate and multivariate logistic regression analyses to construct a Nomogram model. The model’s performance was evaluated using receiver operating characteristic (ROC) curves, accuracy, and other metrics, and compared with the six traditional TI-RADS systems. Results Ten independent risk factors were identified, including age, nodule size, and irregular margins. In the validation group, the Nomogram model achieved an accuracy of 78.4%, a sensitivity of 81.6%, a specificity of 71.7%, and an area under the ROC curve (AUC) of 0.849. The sensitivities of the six TI-RADS systems (C-TIRADS, ACR-TIRADS, EU-TIRADS, ATA Guidelines, Kwak-TIRADS, and AACE) for distinguishing benign and malignant nodules were 86.0%, 93.2%, 96.9%, 98.3%, 84.4%, and 98.1%, respectively; specificities were 55.6%, 34.8%, 25.3%, 22.2%, 57.1%, and 21.7%, respectively; accuracies were 76.1%, 74.3%, 73.7%, 73.7%, 75.8%, and 73.4%, respectively; and AUCs were 0.752, 0.661, 0.628, 0.617, 0.757, and 0.616, respectively, with no statistically significant differences among them. The Nomogram model significantly outperformed the traditional systems in measures such as AUC, Net Reclassification Improvement (NRI), Integrated Discrimination Improvement (IDI), Positive Likelihood Ratio (PLR), and Negative Likelihood Ratio (NLR) (P &lt; 0.001). Conclusion The six traditional TI-RADS systems demonstrate similar but overall limited diagnostic efficacy in the Northwestern Chinese population. The Nomogram model, by integrating multidimensional features and applying a quantitative scoring approach, improves the accuracy and objectivity of malignancy risk assessment. Compared to traditional models, it offers better clinical utility, supports optimized decision-making, and helps reduce unnecessary invasive procedures.

Machine-learning-derived online prediction models of outcomes for patients with cholelithiasis-induced acute cholangitis: development and validation in two retrospective cohorts

PurposeThe aims of the study were to develop and evaluate a machine learning model with which to predict postnatal growth failure (PGF) among very low birth weight (VLBW) infants.Materials and MethodsOf 10425 VLBW infants registered in the Korean Neonatal Network between 2013 and 2017, 7954 infants were included. PGF was defined as a decrease in Z score >1.28 at discharge, compared to that at birth. Six metrics [area under the receiver operating characteristic curve (AUROC), accuracy, precision, sensitivity, specificity, and F1 score] were obtained at five time points (at birth, 7 days, 14 days, 28 days after birth, and at discharge). Machine learning models were built using four different techniques [extreme gradient boosting (XGB), random forest, support vector machine, and convolutional neural network] to compare against the conventional multiple logistic regression (MLR) model.ResultsThe XGB algorithm showed the best performance with all six metrics across the board. When compared with MLR, XGB showed a significantly higher AUROC (p=0.03) for Day 7, which was the primary performance metric. Using optimal cut-off points, for Day 7, XGB still showed better performances in terms of AUROC (0.74), accuracy (0.68), and F1 score (0.67). AUROC values seemed to increase slightly from birth to 7 days after birth with significance, almost reaching a plateau after 7 days after birth.ConclusionWe have shown the possibility of predicting PGF through machine learning algorithms, especially XGB. Such models may help neonatologists in the early diagnosis of high-risk infants for PGF for early intervention.

High-resolution NMR metabolomics of patients with subjective cognitive decline plus: Perturbations in the metabolism of glucose and branched-chain amino acids

Despite the significance of predicting the prognosis of idiopathic sudden sensorineural hearing loss (ISSNHL), no predictive models have been established. This study used artificial intelligence to develop prognosis models to predict recovery from ISSNHL. We retrospectively reviewed the medical data of 453 patients with ISSNHL (men, 220; women, 233; mean age, 50.3 years) who underwent treatment at a tertiary hospital between January 2021 and December 2019 and were followed up after 1 month. According to Siegel’s criteria, 203 patients recovered in 1 month. Demographic characteristics, clinical and laboratory data, and pure-tone audiometry were analyzed. Logistic regression (baseline), a support vector machine, extreme gradient boosting, a light gradient boosting machine, and multilayer perceptron were used. The outcomes were the area under the receiver operating characteristic curve (AUROC) primarily, area under the precision-recall curve, Brier score, balanced accuracy, and F1 score. The light gradient boosting machine model had the best AUROC and balanced accuracy. Together with multilayer perceptron, it was also significantly superior to logistic regression in terms of AUROC. Using the SHapley Additive exPlanation method, we found that the initial audiogram shape is the most important prognostic factor. Machine/deep learning methods were successfully established to predict the prognosis of ISSNHL.

ObjectivesTo investigate the added diagnostic value of arterial spin labelling (ASL) and diffusion tensor imaging (DTI) to structural MRI for computer-aided classification of Alzheimer's disease (AD), frontotemporal dementia (FTD), and controls.MethodsThis retrospective study used MRI data from 24 early-onset AD and 33 early-onset FTD patients and 34 controls (CN). Classification was based on voxel-wise feature maps derived from structural MRI, ASL, and DTI. Support vector machines (SVMs) were trained to classify AD versus CN (AD-CN), FTD-CN, AD-FTD, and AD-FTD-CN (multi-class). Classification performance was assessed by the area under the receiver-operating-characteristic curve (AUC) and accuracy. Using SVM significance maps, we analysed contributions of brain regions.ResultsCombining ASL and DTI with structural MRI resulted in higher classification performance for differential diagnosis of AD and FTD (AUC = 84%; p = 0.05) than using structural MRI by itself (AUC = 72%). The performance of ASL and DTI themselves did not improve over structural MRI. The classifications were driven by different brain regions for ASL and DTI than for structural MRI, suggesting complementary information.ConclusionsASL and DTI are promising additions to structural MRI for classification of early-onset AD, early-onset FTD, and controls, and may improve the computer-aided differential diagnosis on a single-subject level.Key points• Multiparametric MRI is promising for computer-aided diagnosis of early-onset AD and FTD.• Diagnosis is driven by different brain regions when using different MRI methods.• Combining structural MRI, ASL, and DTI may improve differential diagnosis of dementia.