Precision-recall Curve Research Articles

Enhancing Breast Cancer Detection through Advanced AI-Driven Ultrasound Technology: A Comprehensive Evaluation of Vis-BUS.

This study aims to enhance breast cancer detection accuracy through an AI-driven ultrasound tool, Vis-BUS, developed by Barreleye Inc., Seoul, South Korea. Vis-BUS incorporates Lesion Detection AI (LD-AI) and Lesion Analysis AI (LA-AI), along with a Cancer Probability Score (CPS), to differentiate between benign and malignant breast lesions. A retrospective analysis was conducted on 258 breast ultrasound examinations to evaluate Vis-BUS's performance. The primary methods included the application of LD-AI and LA-AI to b-mode ultrasound images and the generation of CPS for each lesion. Diagnostic accuracy was assessed using metrics such as the Area Under the Receiver Operating Characteristic curve (AUROC) and the Area Under the Precision-Recall curve (AUPRC). The study found that Vis-BUS achieved high diagnostic accuracy, with an AUROC of 0.964 and an AUPRC of 0.967, indicating its effectiveness in distinguishing between benign and malignant lesions. Logistic regression analysis identified that 'Fatty' lesion density had an extremely high odds ratio (OR) of 27.7781, suggesting potential convergence issues. The 'Unknown' density category had an OR of 0.3185, indicating a lower likelihood of correct classification. Medium and large lesion sizes were associated with lower likelihoods of correct classification, with ORs of 0.7891 and 0.8014, respectively. The presence of microcalcifications showed an OR of 1.360. Among Breast Imaging-Reporting and Data System categories, category C5 had a significantly higher OR of 10.173, reflecting a higher likelihood of correct classification. Vis-BUS significantly improves diagnostic precision and supports clinical decision-making in breast cancer screening. However, further refinement is needed in areas like lesion density characterization and calcification detection to optimize its performance.

Recognition of dispersed organic matter macerals using YOLOv5m model with convolutional block attention module

The identification of dispersed organic matter (DOM) macerals plays a crucial role in petroleum geology. However, the conventional manual statistical methods are both time-consuming and laborious. With advancements in deep learning, automatic computerized identification of macerals has become the prevailing trend. In this study, we present the first application of the YOLOv5m lightweight semantic segmentation model with convolutional block attention module (CBAM) for classifying DOM macerals into five distinct categories, namely vitrinite, inertinite, solid bitumen, sporinite, and alginite. The results demonstrate that the YOLOv5m + CBAM model exhibits a superior convergence rate, as evidenced by the loss and precision-recall curves. It also achieves a 2 % increase in mean average precision at 0.5 intersection over union threshold over the YOLOv5m model. Moreover, the YOLOv5m + CBAM model shows robustness and efficiency in DOM macerals segmentation, achieving a mean intersection over union of 53.4 % and a mean pixel accuracy of 72.2 % (excluding the background category), while maintaining similar parameters and floating-point operations comparable to those of YOLOv5m. Comparative analysis against models such as DeepLab v3 + and U-net reveals that YOLOv5m + CBAM achieves superior accuracy with fewer parameters, particularly on smaller datasets. Statistical analysis confirms its consistency with manual labeling, resulting in a mean absolute percentage error of 7.34 %. Consequently, the proposed YOLOv5m + CBAM framework represents an ideal approach for pixel segmentation of DOM macerals.

Multimodal fusion learning for long QT syndrome pathogenic genotypes in a racially diverse population

Congenital long QT syndrome (LQTS) diagnosis is complicated by limited genetic testing at scale, low prevalence, and normal QT corrected interval in patients with high-risk genotypes. We developed a deep learning approach combining electrocardiogram (ECG) waveform and electronic health record data to assess whether patients had pathogenic variants causing LQTS. We defined patients with high-risk genotypes as having ≥1 pathogenic variant in one of the LQTS-susceptibility genes. We trained the model using data from United Kingdom Biobank (UKBB) and then fine-tuned in a racially/ethnically diverse cohort using Mount Sinai BioMe Biobank. Following group-stratified 5-fold splitting, the fine-tuned model achieved area under the precision-recall curve of 0.29 (95% confidence interval [CI] 0.28–0.29) and area under the receiver operating curve of 0.83 (0.82–0.83) on independent testing data from BioMe. Multimodal fusion learning has promise to identify individuals with pathogenic genetic mutations to enable patient prioritization for further work up.

ChatGPT-Assisted Classification of Postoperative Bleeding Following Microinvasive Glaucoma Surgery Using Electronic Health Record Data

PurposeTo evaluate the performance of a large language model (LLM) in classifying electronic health record (EHR) text, and to use this classification to evaluate the type and resolution of hemorrhagic events (HE) following micro-invasive glaucoma surgery (MIGS). DesignRetrospective cohort study. ParticipantsEyes from the Bascom Palmer Glaucoma Repository. MethodsEyes that underwent MIGS between July 1, 2014 and February 1, 2022 were analyzed. ChatGPT was used to classify deidentified EHR anterior chamber examination text into HE categories (no hyphema, microhyphema, clot, and hyphema). Agreement between classifications by ChatGPT and a glaucoma specialist was evaluated using Cohen’s Kappa and precision-recall (PR) curve. Time to resolution of HEs was assessed using Cox proportional-hazards models. Goniotomy HE resolution was evaluated by degree of angle treatment (90-179º, 180-269º, 270-360º). Logistic regression was used to identified HE risk factors. Main Outcome MeasuresAccuracy of ChatGPT HE classification and incidence and resolution of HEs. ResultsThe study included 434 goniotomy eyes (368 patients) and 528 Schlemm’s Canal Stent (SCS) eyes (390 patients). ChatGPT facilitated excellent HE classification (Cohen’s kappa 0.93, area under PR curve 0.968). Using ChatGPT classifications, at postoperative day 1, HEs occurred in 67.8% of goniotomy and 25.2% of SCS eyes (p<0.001). The 270-360º goniotomy group had the highest HE rate (84.0%, p<0.001). At postoperative week 1, HEs were observed in 43.4% and 11.3% of goniotomy and SCS eyes respectively (p<0.001). By postoperative month 1, HE rates were 13.3% and 1.3% among goniotomy and SCS eyes respectively (p<0.001). Time to HE resolution differed between the goniotomy angle groups (log-rank p=0.034); median time to resolution was 10, 10, and 15 days for the 90-179º, 180-269º, and 270-360º groups respectively. Risk factor analysis demonstrated greater goniotomy angle was the only significant predictor of HEs (OR for 270-360º: 4.08, p<0.001). ConclusionsLLMs can be effectively used to classify longitudinal EHR free-text exam data with high accuracy, highlighting a promising direction for future LLM-assisted research and clinical decision support. HEs are relatively common, self-resolving complications that occur more often in goniotomy cases and with larger goniotomy treatments. Time to HE resolution differs significantly between goniotomy groups.

ShapeBench: A new approach to benchmarking local 3D shape descriptors

The ShapeBench evaluation methodology is proposed as an extension to the popular Area Under Precision-Recall Curve (PRC/AUC) for measuring the matching performance of local 3D shape descriptors. It is observed that the PRC inadequately accounts for other similar surfaces in the same or different objects when determining whether a candidate match is a true positive. The novel Descriptor Distance Index (DDI) metric is introduced to address this limitation. In contrast to previous evaluation methodologies, which identify entire objects in a given scene, the DDI metric measures descriptor performance by analysing point-to-point distances. The ShapeBench methodology is also more scalable than previous approaches, by using procedural generation. The benchmark is used to evaluate both old and new descriptors. The results produced by the implementation of the benchmark are fully replicable, and are made publicly available.

Deep Learning and Minimally Invasive Endoscopy: Panendoscopic Detection of Pleomorphic Lesions

Introduction: Capsule endoscopy (CE) is a minimally invasive exam suitable of panendoscopic evaluation of the gastrointestinal (GI) tract. Nevertheless, CE is time-consuming with suboptimal diagnostic yield in the upper GI tract. Convolutional neural networks (CNN) are human brain architecture-based models suitable for image analysis. However, there is no study about their role in capsule panendoscopy. Methods: Our group developed an artificial intelligence (AI) model for panendoscopic automatic detection of pleomorphic lesions (namely vascular lesions, protuberant lesions, hematic residues, ulcers, and erosions). 355,110 images (6,977 esophageal, 12,918 gastric, 258,443 small bowel, 76,772 colonic) from eight different CE and colon CE (CCE) devices were divided into a training and validation dataset in a patient split design. The model classification was compared to three CE experts’ classification. The model’s performance was evaluated by its sensitivity, specificity, accuracy, positive predictive value, negative predictive value, and area under the precision-recall curve. Results: The binary esophagus CNN had a diagnostic accuracy for pleomorphic lesions of 83.6%. The binary gastric CNN identified pleomorphic lesions with a 96.6% accuracy. The undenary small bowel CNN distinguished pleomorphic lesions with different hemorrhagic potentials with 97.6% accuracy. The trinary colonic CNN (detection and differentiation of normal mucosa, pleomorphic lesions, and hematic residues) had 94.9% global accuracy. Discussion/Conclusion: We developed the first AI model for panendoscopic automatic detection of pleomorphic lesions in both CE and CCE from multiple brands, solving a critical interoperability technological challenge. Deep learning-based tools may change the landscape of minimally invasive capsule panendoscopy.

Enhancing corporate bankruptcy prediction via a hybrid genetic algorithm and domain adaptation learning architecture

In the contemporary business landscape, accurately evaluating a company’s financial health is essential for stakeholders to mitigate risks and avert bankruptcy. This study presents an innovative approach to improving business bankruptcy prediction through the hybrid integration of Domain Adaptation Learning (DAL) and Genetic Algorithm (GA) techniques. The hybrid model harnesses DAL to address distributional changes in real-world scenarios and utilises GA’s proficiency in feature selection. Six machine learning models are rigorously evaluated against the proposed hybrid model: Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LR), Gradient Boosting (GB), k-Nearest Neighbours (k-NN), and Stacking Ensemble (SE). Our hybrid model performs well on imbalanced target datasets using the Area Under the Precision–Recall Curve metric: 0.93 (RF), 0.93 (SVM), 0.89 (LR), 0.91 (GB), 0.88 (k-NN), and 0.92 (SE). These findings highlight the model’s ability to overcome the limitations of traditional approaches, offering a more reliable predictive framework for stakeholders to make informed decisions and proactively manage financial stability. Future research directions may explore the applicability of this hybrid model across different industries and the integration of additional techniques to further enhance its performance.

Prediction of vasopressor needs in hypotensive emergency department patients using serial arterial blood pressure data with deep learning

AbstractBackgroundShock is a life‐threatening condition that is associated with high mortality and morbidity. Therefore, the timely identification and management of this condition are important. We aimed to develop a prediction model for vasopressor use based on concise serial arterial blood pressure data.MethodsWe collected continuous arterial blood pressure from patients admitted to the emergency department (ED) resuscitation room. Patients with an initial systolic blood pressure lower than 90 mmHg were included in the study. We developed prediction models using convolutional neural networks (CNNs) and long short‐term memory (LSTM) networks. Discrimination performance was assessed using the area under the receiver operating characteristic curve (AUROC) and the area under the precision–recall curve (AUPRC).ResultsA total of 120 patients were enrolled in the study. The CNN and LSTM models yielded AUROCs ranging from 0.731 to 0.834 for predicting the need for vasopressor infusion within different time frames (30 min, 1 h, and 6 h). LSTM outperformed the CNN in terms of predicting vasopressor infusion within 30 min and 1 h (p value &lt; 0.05). The AUPRC values ranged from 0.200 to 0.252, and sensitivity–specificity analyses indicated that the models have potential for clinical applications.ConclusionThis study used serial arterial blood pressure data to construct a promising prediction model for the need for vasopressors in hypotensive ED patients. The simplicity and accuracy of the model provide valuable insights for developing a clinical decision support tool for resuscitation in emergency settings.

Development of a deep learning model for detecting lumbar vertebral fractures on CT images: An external validation

ObjectiveTo develop and externally validate a binary classification model for lumbar vertebral body fractures based on CT images using deep learning methods. MethodsThis study involved data collection from two hospitals for AI model training and external validation. In Cohort A from Hospital 1, CT images from 248 patients, comprising 1508 vertebrae, revealed that 20.9% had fractures (315 vertebrae) and 79.1% were non-fractured (1193 vertebrae). In Cohort B from Hospital 2, CT images from 148 patients, comprising 887 vertebrae, indicated that 14.8% had fractures (131 vertebrae) and 85.2% were non-fractured (756 vertebrae). The AI model for lumbar spine fractures underwent two stages: vertebral body segmentation and fracture classification. The first stage utilized a 3D V-Net convolutional deep neural network, which produced a 3D segmentation map. From this map, region of each vertebra body were extracted and then input into the second stage of the algorithm. The second stage employed a 3D ResNet convolutional deep neural network to classify each proposed region as positive (fractured) or negative (not fractured). ResultsThe AI model’s accuracy for detecting vertebral fractures in Cohort A’s training set (n = 1199), validation set (n = 157), and test set (n = 152) was 100.0 %, 96.2 %, and 97.4 %, respectively. For Cohort B (n = 148), the accuracy was 96.3 %. The area under the receiver operating characteristic curve (AUC-ROC) values for the training, validation, and test sets of Cohort A, as well as Cohort B, and their 95 % confidence intervals (CIs) were as follows: 1.000 (1.000, 1.000), 0.978 (0.944, 1.000), 0.986 (0.969, 1.000), and 0.981 (0.970, 0.992). The area under the precision-recall curve (AUC-PR) values were 1.000 (0.996, 1.000), 0.964 (0.927, 0.985), 0.907 (0.924, 0.984), and 0.890 (0.846, 0.971), respectively. According to the DeLong test, there was no significant difference in the AUC-ROC values between the test set of Cohort A and Cohort B, both for the overall data and for each specific vertebral location (all P>0.05). ConclusionThe developed model demonstrates promising diagnostic accuracy and applicability for detecting lumbar vertebral fractures.

Predicting postoperative delirium assessed by the Nursing Screening Delirium Scale in the recovery room for non-cardiac surgeries without craniotomy: A retrospective study using a machine learning approach.

Postoperative delirium (POD) contributes to severe outcomes such as death or development of dementia. Thus, it is desirable to identify vulnerable patients in advance during the perioperative phase. Previous studies mainly investigated risk factors for delirium during hospitalization and further used a linear logistic regression (LR) approach with time-invariant data. Studies have not investigated patients' fluctuating conditions to support POD precautions. In this single-center study, we aimed to predict POD in a recovery room setting with a non-linear machine learning (ML) technique using pre-, intra-, and postoperative data. The target variable POD was defined with the Nursing Screening Delirium Scale (Nu-DESC) ≥ 1. Feature selection was conducted based on robust univariate test statistics and L1 regularization. Non-linear multi-layer perceptron (MLP) as well as tree-based models were trained and evaluated-with the receiver operating characteristics curve (AUROC), the area under precision recall curve (AUPRC), and additional metrics-against LR and published models on bootstrapped testing data. The prevalence of POD was 8.2% in a sample of 73,181 surgeries performed between 2017 and 2020. Significant univariate impact factors were the preoperative ASA status (American Society of Anesthesiologists physical status classification system), the intraoperative amount of given remifentanil, and the postoperative Aldrete score. The best model used pre-, intra-, and postoperative data. The non-linear boosted trees model achieved a mean AUROC of 0.854 and a mean AUPRC of 0.418 outperforming linear LR, well as best applied and retrained baseline models. Overall, non-linear machine learning models using data from multiple perioperative time phases were superior to traditional ones in predicting POD in the recovery room. Class imbalance was seen as a main impediment for model application in clinical practice.

Predicting prognostic factors in kidney transplantation using a machine learning approach to enhance outcome predictions: a retrospective cohort study.

Accurate forecasting of clinical outcomes after kidney transplantation is essential for improving patient care and increasing the success rates of transplants. Our study employs advanced machine learning (ML) algorithms to identify crucial prognostic indicators for kidney transplantation. By analyzing complex datasets with ML models, we aim to enhance prediction accuracy and provide valuable insights to support clinical decision-making. Analyzing data from 4077 KT patients (June 1990 - May 2015) at a single center, this research included 27 features encompassing recipient/donor traits and peri-transplant data. The dataset was divided into training (80%) and testing (20%) sets. Four ML models-eXtreme Gradient Boosting (XGBoost), Feedforward Neural Network, Logistic Regression, and Support Vector Machine-were trained on carefully selected features to predict the success of graft survival. Performance was assessed by precision, sensitivity, F1 score, Area Under the Receiver Operating Characteristic (AUROC), and Area Under the Precision-Recall Curve. XGBoost emerged as the best model, with an AUROC of 0.828, identifying key survival predictors like T-cell flow crossmatch positivity, creatinine levels two years post-transplant and human leukocyte antigen mismatch. The study also examined the prognostic importance of histological features identified by the Banff criteria for renal biopsy, emphasizing the significance of intimal arteritis, interstitial inflammation, and chronic glomerulopathy. The study developed ML models that pinpoint clinical factors crucial for KT graft survival, aiding clinicians in making informed post-transplant care decisions. Incorporating these findings with the Banff classification could improve renal pathology diagnosis and treatment, offering a data-driven approach to prioritizing pathology scores.

Ex-ante expected changes in ESG and future stock returns based on machine learning

This study has two primary objectives. Firstly, it enhances the reliability and transparency of machine-learning-based models for predicting future changes in environmental, social and governance (ESG) performance. Secondly, it explores the relationship between ex-ante expected changes in ESG and future stock returns. This study collects 3258 STOXX Europe 600 firm-year observations. In the ESG prediction phase, two machine learning algorithms (logistic regression and random forest) are utilised to develop ESG forecasting models. Hyperparameter optimisation and walk-forward validation techniques are employed to address issues of underestimation and information leakage. The machine-learning-based ESG forecasting models are evaluated using three metrics: accuracy, area under the curve (AUC) and area under the precision-recall curve (AUPR). Subsequently, this study investigates the relationship between ex-ante expected ESG changes and future stock returns using the predicted ESG changes. A positive correlation is found between ex-ante expected ESG changes and future stock returns. The supplementary tests also reveal that this positive relationship is highly and statistically significant among large firms and after the COVID-19 pandemic. Moreover, this study introduces a robust and transparent approach for constructing effective machine-learning-based ESG forecasting models using hyperparameter optimisation and walk-forward validation. Additionally, traditional regression analyses are modernised by incorporating machine-learning-predicted independent variables. Furthermore, the findings provide empirical support for stakeholder, agency and resource-based theories. Finally, practical insights are provided to facilitate ESG-focused investment portfolio decision making.

The digital signature of emergent tremor in Parkinson’s disease

Emergent tremor in Parkinson’s disease (PD) can occur during sustained postures or movements that are different from action tremor. Tremor can contaminate the clinical rating of bradykinesia during finger tapping. Currently, there is no reliable way of isolating emergent tremor and measuring the cardinal motor symptoms based on voluntary movements only. In this study, we investigated whether emergent tremor during repetitive alternating finger tapping (RAFT) on a quantitative digitography (QDG) device could be reliably identified and distinguished from voluntary tapping. Ninety-six individuals with PD and forty-two healthy controls performed a thirty-second QDG-RAFT task and the Movement Disorders Society – Unified Parkinson’s Disease Rating Scale Part III (MDS-UPDRS III). Visual identification of tremor during QDG-RAFT was labeled by an experienced movement disorders specialist. Two methods of identifying tremor were investigated: 1) physiologically informed temporal thresholds 2) XGBoost model using temporal and amplitude features of tapping. The XGBoost model showed high accuracy for identifying tremor (area under the precision-recall curve of 0.981) and outperformed temporal-based thresholds. Percent time duration of classifier-identified tremor showed significant correlations with MDS-UPDRS III tremor subscores (r = 0.50, p < 0.0001). There was a significant change in QDG metrics for bradykinesia, rigidity, and arrhythmicity after tremor strikes were excluded (p < 0.01). The results demonstrate that emergent tremor during QDG-RAFT has a unique digital signature and the duration of tremor correlated with the MDS-UPDRS III tremor items. When involuntary tremor strikes were excluded, the QDG metrics of bradykinesia and rigidity were significantly worse, demonstrating the importance of distinguishing tremor from voluntary movement when rating bradykinesia.

A foundational transformer leveraging full night, multichannel sleep study data accurately classifies sleep stages.

To investigate whether a foundational transformer model using 8-hour, multichannel data from polysomnograms can outperform existing artificial intelligence (AI) methods for sleep stage classification. We utilized the Sleep Heart Health Study (SHHS) visits 1 and 2 for training and validation and the Multi-Ethnic Study of Atherosclerosis (MESA) for testing of our model. We trained a self-supervised foundational transformer (called PFTSleep) that encodes 8-hour long sleep studies at 125 Hz with 7 signals including brain, movement, cardiac, oxygen, and respiratory channels. These encodings are used as input for training of an additional model to classify sleep stages, without adjusting the weights of the foundational transformer. We compared our results to existing AI methods that did not utilize 8-hour data or the full set of signals but did report evaluation metrics for the SHHS dataset. We trained and validated a model with 8,444 sleep studies with 7 signals including brain, movement, cardiac, oxygen, and respiratory channels and tested on an additional 2,055 studies. In total, we trained and tested 587,944 hours of sleep study signal data. Area under the precision recall curve (AUPRC) scores were 0.82, 0.40, 0.53, 0.75, and 0.82 and area under the receiving operating characteristics curve (AUROC) scores were 0.99, 0.95, 0.96, 0.98, and 0.99 for wake, N1, N2, N3, and REM, respectively, on the SHHS validation set. For MESA, the AUPRC scores were 0.56, 0.16, 0.40, 0.45, and 0.65 and AUROC scores were 0.94, 0.77, 0.87, 0.91, and 0.96, respectively. Our model was compared to the longest context window state-of-the-art model and showed increases in macro evaluation scores, notably sensitivity (3.7% increase) and multi-class REM (3.39% increase) and wake (0.97% increase) F1 scores. Utilizing full night, multi-channel PSG data encodings derived from a foundational transformer improve sleep stage classification over existing methods.

A two-stage ensemble learning based prediction and grading model for PD-1/PD-L1 inhibitor-related cardiac adverse events: A multicenter retrospective study

BackgroundImmune-related cardiac adverse events (ircAEs) caused by programmed cell death protein-1 (PD-1) and programmed death-ligand-1 (PD-L1) inhibitors can lead to fulminant and even fatal consequences. This study aims to develop a prediction and grading model for ircAEs, enabling graded management of patients. MethodsThis study utilized medical record systems from two medical institutions to develop a prediction and grading model for ircAEs using ten machine learning algorithms and two variable screening methods. The model was developed based on a two-stage ensemble learning framework. In the first stage, the ircAEs and non-ircAEs cases were classified. In the second stage, ircAEs cases were grouped into grades 1–2 and 3–5. The experiments were evaluated using five-fold cross-validation. The model's prediction performance was assessed using accuracy, precision, recall, F1 value, Brier score, receiver operating characteristic curve area (AUC), and area under the precision-recall curve (AUPR). Results615 patients were included in the study. 147 experienced ircAEs, and 44 experienced grade 3–5 ircAEs. The soft voting classifier trained using the variables screened by feature importance ranking performed better than other classifiers in both stages. The average AUC for the first and second stages is 84.18 % and 85.13 %, respectively. In the first stage, the three most important variables are N-terminal B-type natriuretic peptide (NT-proBNP), interleukin-2 (IL-2), and C-reactive protein (CRP). In the second stage, the patient's age, NT-proBNP, and left ventricular ejection fraction (LVEF) are the three most critical variables. ConclusionsThe prediction and grading model of ircAEs based on two-stage ensemble learning established in this study has good performance and potential clinical application.

Deep Learning-Based Electrocardiogram Analysis Predicts Biventricular Dysfunction and Dilation in Congenital Heart Disease

BackgroundArtificial intelligence–enhanced electrocardiogram (AI-ECG) analysis shows promise to detect biventricular pathophysiology. However, AI-ECG analysis remains underexplored in congenital heart disease (CHD). ObjectivesThe purpose of this study was to develop and externally validate an AI-ECG model to predict cardiovascular magnetic resonance (CMR)-defined biventricular dysfunction/dilation in patients with CHD. MethodsWe trained (80%) and tested (20%) a convolutional neural network on paired ECG-CMRs (≤30 days apart) from patients with and without CHD to detect left ventricular (LV) dysfunction (ejection fraction ≤40%), RV dysfunction (ejection fraction ≤35%), and LV and RV dilation (end-diastolic volume z-score ≥4). Performance was assessed during internal testing and external validation on an outside health care system using area under receiver-operating curve (AUROC) and area under precision recall curve. ResultsThe internal and external cohorts comprised 8,584 ECG-CMR pairs (n = 4,941; median CMR age 20.7 years) and 909 ECG-CMR pairs (n = 746; median CMR age 25.4 years), respectively. Model performance was similar for internal testing (AUROC: LV dysfunction 0.87; LV dilation 0.86; RV dysfunction 0.88; RV dilation 0.81) and external validation (AUROC: LV dysfunction 0.89; LV dilation 0.83; RV dysfunction 0.82; RV dilation 0.80). Model performance was lowest in functionally single ventricle patients. Tetralogy of Fallot patients predicted to be at high risk of ventricular dysfunction had lower survival (P < 0.001). Model explainability via saliency mapping revealed that lateral precordial leads influence all outcome predictions, with high-risk features including QRS widening and T-wave inversions for RV dysfunction/dilation. ConclusionsAI-ECG shows promise to predict biventricular dysfunction/dilation, which may help inform CMR timing in CHD.

Pediatric septic shock estimation using deep learning and electronic medical records.

Diagnosing pediatric septic shock is difficult due to the complex and often impractical traditional criteria, such as systemic inflammatory response syndrome (SIRS), which result in delays and higher risks. This study aims to develop a deep learning-based model using SIRS data for early diagnosis in pediatric septic shock cases. The study analyzed data from pediatric patients (<18 years old) admitted to a tertiary hospital from January 2010 to July 2023. Vital signs, lab tests, and clinical information were collected. Septic shock cases were identified using SIRS criteria and inotrope use. A deep learning model was trained and evaluated using the area under the receiver operating characteristics curve (AUROC) and area under the precision-recall curve (AUPRC). Variable contributions were analyzed using the Shapley additive explanation value. The analysis, involving 9,616,115 measurements, identified 34,696 septic shock cases (0.4%). Oxygen supply was crucial for 41.5% of the control group and 20.8% of the septic shock group. The final model showed strong performance, with an AUROC of 0.927 and AUPRC of 0.879. Key influencers were age, oxygen supply, sex, and partial pressure of carbon dioxide, while body temperature had minimal impact on estimation. The proposed deep learning model simplifies early septic shock diagnosis in pediatric patients, reducing the diagnostic workload. Its high accuracy allows timely treatment, but external validation through prospective studies is needed.

A machine learning model to predict liver-related outcomes after the functional cure of chronic hepatitis B

The risk of hepatocellular carcinoma (HCC) and hepatic decompensation persists after hepatitis B surface antigen (HBsAg) seroclearance. This study aimed to develop and validate a machine learning model to predict the risk of liver-related outcomes (LROs) following HBsAg seroclearance. A total of 4,787 consecutive patients who achieved HBsAg seroclearance between 2000 and 2022 were enrolled from six centers in South Korea and a territory-wide database in Hong Kong, comprising the training (n= 944), internal validation (n= 1,102), and external validation (n= 2,741) cohorts. Three machine learning-based models were developed and compared in each cohort. The primary outcome was the development of any LRO, including HCC, decompensation, and liver-related death. During a median follow-up of 55.2 (IQR 30.1-92.3) months, 123 LROs were confirmed (1.1%/person-year) in the Korean cohort. The model with the best predictive performance in the training cohort was selected as the final model (designated as PLAN-B-CURE), which was constructed using a gradient boosting algorithm and seven variables (age, sex, diabetes, alcohol consumption, cirrhosis, albumin, and platelet count). Compared to previous HCC prediction models, PLAN-B-CURE showed significantly superior accuracy in the training cohort (c-index: 0.82 vs. 0.63-0.70, all p <0.001; area under the receiver-operating characteristic curve: 0.86 vs. 0.62-0.72, all p <0.01; area under the precision-recall curve: 0.53 vs. 0.13-0.29, all p <0.01). PLAN-B-CURE showed a reliable calibration function (Hosmer-Lemeshow test p >0.05) and these results were reproduced in the internal and external validation cohorts. This novel machine learning model consisting of seven variables provides reliable risk prediction of LROs after HBsAg seroclearance that can be used for personalized surveillance. Using large-scale multinational data, we developed a machine learning model to predict the risk of liver-related outcomes (i.e., hepatocellular carcinoma, decompensation, and liver-related death) after the functional cure of chronic hepatitis B (CHB). The new model named PLAN-B-CURE was constructed using seven variables (age, sex, alcohol consumption, diabetes, cirrhosis, serum albumin, and platelet count) and a gradient boosting machine algorithm, and it demonstrated significantly better predictive accuracy than previous models in both the training and validation cohorts. The inclusion of diabetes and significant alcohol intake as model inputs suggests the importance of metabolic risk factor management after the functional cure of CHB. Using seven readily available clinical factors, PLAN-B-CURE, the first machine learning-based model for risk prediction after the functional cure of CHB, may serve as a basis for individualized risk stratification.

High Prevalence of Artifacts in Optical Coherence Tomography With Adequate Signal Strength.

This study aims to investigate the prevalence of artifacts in optical coherence tomography (OCT) images with acceptable signal strength and evaluate the performance of supervised deep learning models in improving OCT image quality assessment. We conducted a retrospective study on 4555 OCT images from 546 patients, with each image having an acceptable signal strength (≥6). A comprehensive analysis of prevalent OCT artifacts was performed, and five pretrained convolutional neural network models were trained and tested to infer images based on quality. Our results showed a high prevalence of artifacts in OCT images with acceptable signal strength. Approximately 21% of images were labeled as nonacceptable quality. The EfficientNetV2 model demonstrated superior performance in classifying OCT image quality, achieving an area under the receiver operating characteristic curve of 0.950 ± 0.007 and an area under the precision recall curve of 0.985 ± 0.002. The findings highlight the limitations of relying solely on signal strength for OCT image quality assessment and the potential of deep learning models in accurately classifying image quality. Application of the deep learning-based OCT image quality assessment models may improve the OCT image data quality for both clinical applications and research.

Adaptive fine-tuning based transfer learning for the identification of MGMT promoter methylation status.

Background.Glioblastoma Multiforme (GBM) is an aggressive form of malignant brain tumor with a generally poor prognosis.O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation has been shown to be a predictive bio-marker for resistance to treatment of GBM, but it is invasive and time-consuming to determine methylation status. There has been effort to predict the MGMT methylation status through analyzing MRI scans using machine learning, which only requires pre-operative scans that are already part of standard-of-care for GBM patients.Purpose.To improve the performance of conventional transfer learning in the identification of MGMT promoter methylation status, we developed a 3D SpotTune network with adaptive fine-tuning capability. Using the pretrained weights of MedicalNet with the SpotTune network, we compared its performance with a randomly initialized network for different combinations of MR modalities.Methods.Using a ResNet50 as the base network, three categories of networks are created: (1) A 3D SpotTune network to process volumetric MR images, (2) a network with randomly initialized weights, and (3) a network pre-trained on MedicalNet. These three networks are trained and evaluated using a public GBM dataset provided by the University of Pennsylvania. The MRI scans from 240 patients are used, with 11 different modalities corresponding to a set of perfusion, diffusion, and structural scans. The performance is evaluated using 5-fold cross validation with a hold-out testing dataset.Results.The SpotTune network showed better performance than the randomly initialized network. The best performing SpotTune model achieved an area under the Receiver Operating Characteristic curve (AUC), average precision of the precision-recall curve (AP), sensitivity, and specificity values of 0.6604, 0.6179, 0.6667, and 0.6061 respectively.Conclusions.SpotTune enables transfer learning to be adaptive to individual patients, resulting in improved performance in predicting MGMT promoter methylation status in GBM using equivalent MRI modalities as compared to a randomly initialized network.