Deep Code-Comment Understanding and Assessment

To develop a method for identifying high-risk patients among septic populations requiring mechanical ventilation, and to conduct phenotypic analysis based on this method. Data from four sources were utilized: the Medical Information Mart for Intensive Care (MIMIC-IV 2.0, MIMIC-III 1.4), the Philips eICU-Collaborative Research Database 2.0 (eICU-CRD 2.0), and the Anhui Medical University Second Affiliated Hospital dataset. The adult patients in intensive care unit (ICU) who met Sepsis-3 and received invasive mechanical ventilation (IMV) on the first day of first admission were enrolled. The MIMIC-IV dataset with the highest data integrity was divided into a training set and a test set at a 6:1 ratio, while the remaining datasets were served as validation sets. The demographic information, comorbidities, laboratory indicators, commonly used ICU scores, and treatment measures of patients were extracted. Clinical data collected within first day of ICU admission were used to calculate the sequential organ failure assessment (SOFA) score. K-means clustering was applied to cluster SOFA score components, and the sum of squared errors (SSE) and Davies-Bouldin index (DBI) were used to determine the optimal number of disease subtypes. For clustering results, normalized methods were employed to compare baseline characteristics by visualization, and Kaplan-Meier curves were used to analyze clinical outcomes across phenotypes. This study enrolled patients from MIMIC-IV dataset (n = 11 166), MIMIC-III dataset (n = 4 821), eICU-CRD dataset (n = 6 624), and a local dataset (n = 110), with the four datasets showing similar median ages and male proportions exceeding 50%; using 85% of the MIMIC-IV dataset as the training set, 15% as the test set, and the rest dataset as the validation set. K-means clustering based on the six-item SOFA score was performed to determine the optimal number of clusters as 3, and patients were finally classified into three phenotypes. In the training set, compared with the patients with phenotype II and phenotype III, those with phenotype I had the more severe circulatory and respiratory dysfunction, a higher proportion of vasoactive drug usage, more obvious metabolic acidosis and hypoxia, and a higher incidence of congestive heart failure. The patients with phenotype II was dominated by respiratory dysfunction with higher visceral injury. The patients with phenotype III had relatively stable organ function. The above characteristics were consistent in both the test and validation sets. Analysis of infection-related indicators showed that the patients with phenotype I had the highest SOFA score within 7 days after ICU admission, initial decreases and later increases in platelet count (PLT), and higher counts of neutrophils, lymphocytes, and monocytes as compared with those with phenotype II and phenotype III, their blood cultures had a higher positivity rates for Gram-positive bacteria, Gram-negative bacteria and fungi as compared with those with phenotype II and phenotype III. The Kaplan-Meier curve indicated that in the training, test, and validation sets, the 28-day cumulative mortality of patients with phenotype I was significantly higher than that of patients with phenotypes II and phenotype III. Three distinct phenotypes in septic patients receiving IMV based on unsupervised machine learning is derived, among which phenotype I, characterized by cardiorespiratory failure, can be used for the early identification of high-risk patients in this population. Moreover, this population is more prone to bloodstream infections, posing a high risk and having a poor prognosis.

Currently, developers are often obligated to enhance code quality. High-quality code is often accompanied with comprehensive summaries, including code documentation and function explanations, which are invaluable for maintenance and further development. Regrettably, few software projects provide sufficient code comments owing to the high costs associated with human labeling. Contemporary researchers in software engineering concentrate on the methods for automated comment generating. Initial algorithms depended on handwritten templates or information retrieval methods. With the advancement of machine learning, researchers construct automated models for machine translation instead. Nonetheless, the produced code comments remain inadequate owing to the significant disparity between code structure and normal language. This study introduces a unique deep learning model, At-ComGen, which utilizes hybrid attention for the automated creation of source code comments. Utilizing two separate LSTM encoders, our approach integrates essential tokens from source code functions with the code structure, represented by a corresponding Abstract Syntax Tree. In contrast to earlier data-driven models, our methodology utilizes code syntax and semantics in the generation of comments. The hybrid attention method, used for comment creation for the first time to our knowledge, enhances the quality of code comments. The tests demonstrate that At-ComGen is efficacious and surpasses other prevalent methodologies. Machine comments from Seq2Seq and CODE-NN disregard code structure underlying DeepCom and At-ComGen. At-ComGen has 59.3%, 36.4%, 43.3%, and 13.1% higher comment BLEU values than baseline models for a 5-line function. Even though model performance reduces with comment length, At-ComGen's comments often outperform others. 5–10-word machine comments work best. For reference length 10, At-ComGen has 38.2%, 23.7%, 9.3%, and 4.4% greater BLEU values than the other baseline models.

Prior to using a quantitative structure activity relationship (QSAR) model for external predictions, its predictive power should be established and validated. In the absence of a true external data set, the best way to validate the predictive ability of a model is to perform its statistical external validation. In statistical external validation, the overall data set is divided into training and test sets. Commonly, this splitting is performed using random division. Rational splitting methods can divide data sets into training and test sets in an intelligent fashion. The purpose of this study was to determine whether rational division methods lead to more predictive models compared to random division. A special data splitting procedure was used to facilitate the comparison between random and rational division methods. For each toxicity end point, the overall data set was divided into a modeling set (80% of the overall set) and an external evaluation set (20% of the overall set) using random division. The modeling set was then subdivided into a training set (80% of the modeling set) and a test set (20% of the modeling set) using rational division methods and by using random division. The Kennard-Stone, minimal test set dissimilarity, and sphere exclusion algorithms were used as the rational division methods. The hierarchical clustering, random forest, and k-nearest neighbor (kNN) methods were used to develop QSAR models based on the training sets. For kNN QSAR, multiple training and test sets were generated, and multiple QSAR models were built. The results of this study indicate that models based on rational division methods generate better statistical results for the test sets than models based on random division, but the predictive power of both types of models are comparable.

Existing studies report very few factors influencing the mortality of infant and toddler heart failure patients during hospitalization. Due to its high mortality rate, it is an important health issue. Therefore, this study aims to explore the factors influencing infant and toddler heart failure patients during hospitalization, establish predictive models, and a mortality scoring table. The study ultimately included 544 cases of infant heart failure patients. They were randomly divided into a training set (380 cases) and a validation set (164 cases) in a ratio of 7:3. The training set was then further divided into the death group and the survival group for further analysis of indicators during hospitalization. Using the Lasso regression method, this study selected the best 14 variables from 88 independent variables of infants and toddlers with heart failure. Multivariate Logistic regression results show that TP<65g/L (OR=2.34), pH<7.35 (OR=2.79), and Respiratory rate (1-12 months: <30times/min and 13-36 months: <25times/min, OR=2.34) are independent risk factors. The model evaluation results for the train and test sets of infant and toddler heart failure patients are as follows: C-index values for discrimination in the train and test sets are 0.721 and 0.728, respectively. Fit test calibration evaluations show P values of 0.9958 and 0.9998, both greater than 0.05, indicating good calibration. The AUC values for the train and test sets are 0.75 and 0.64, respectively, showing a good predictive effect of the model. The mortality scoring table divides patients in the train and test sets into low risk, moderate-risk, and high risk categories. Compared to the low risk group, the OR values for the occurrence of mortality in the Medium risk group and high risk group in the train set are 3.78 and 11.67, respectively; in the test set, the OR values for the moderate-risk group and high-risk group are 1.73 and 6.33, respectively. The predictive models and scoring tables established in this study have a good predictive role in assessing the risk of death in infant and toddler heart failure patients aged 1-36 months during hospitalization, providing clinical guidance and reference value.

To investigative whether radiomics features in bilateral hippocampi from MRI can identify temporal lobe epilepsy (TLE). A total of 131 subjects with MRI (66 TLE patients [35 right and 31 left TLE] and 65 healthy controls [HC]) were allocated to training (n = 90) and test (n = 41) sets. Radiomics features (n = 186) from the bilateral hippocampi were extracted from T1-weighted images. After feature selection, machine learning models were trained. The performance of the classifier was validated in the test set to differentiate TLE from HC and ipsilateral TLE from HC. Identical processes were performed to differentiate right TLE from HC (training set, n = 69; test set; n = 31) and left TLE from HC (training set, n = 66; test set, n = 30). The best-performing model for identifying TLE showed an AUC, accuracy, sensitivity, and specificity of 0.848, 84.8%, 76.2%, and 75.0% in the test set, respectively. The best-performing radiomics models for identifying right TLE and left TLE subgroups showed AUCs of 0.845 and 0.840 in the test set, respectively. In addition, multiple radiomics features significantly correlated with neuropsychological test scores (false discovery rate-corrected p-values < 0.05). The radiomics model from hippocampus can be a potential biomarker for identifying TLE.

This focus issue of Physiological Measurement follows the 38th Annual International Computing in Cardiology (CinC) Conference, hosted in Hangzhou, China in September 2011 by Zhejiang University. Each year, the NIH-sponsored PhysioNet resource (http://physionet.org/) runs an open competition lasting several months, aimed at encouraging the development of solutions to an unsolved or poorly solved problem in biomedicine, in most cases making use of relevant clinical and experimental data provided freely by PhysioNet. Participants in these annual challenges discuss their diverse approaches to the Challenge problems during dedicated scientific sessions at CinC. The topics of these PhysioNet/CinC Challenges range from physiologic signal processing and analysis to forecasting and modelling clinically important events and processes.

Background and AimsThe prognosis of papillary thyroid carcinoma (PTC) is highly variable, even for high-risk cases. The predictive and prognostic role of the lymphocyte-to-monocyte ratio (LMR) has been reported in other cancers. The aim of our present study was to explore the value of LMR prognostic prediction in high-risk PTC patients.Patients and MethodsTwo hundred and twenty-four PTC high-risk cases at West China Hospital were randomized into a training set (112 cases) and testing set (112 cases), while 48 cases in Shang Jin Nan Fu Hospital were included as the external validation set.ResultsA lower preoperative LMR correlated with larger tumor size, advanced N and M stages, and an increased number of multiple PTC cases in the training, testing, and validation sets (all P<0.05 in the three sets). Patients with a high LMR exhibited significantly improved overall and PTC-free survival compared with those of patients with a low LMR in the training, testing, validation, and combined sets (all P<0.05 in the individual and combined sets). Moreover, multivariate analyses identified the LMR as an independent prognostic factor for overall and PTC-free survival. The nomograms for predicting the 5-year mortality and PTC recurrence were developed based on the risk factors in the training set and validated in the independent testing and validation sets.ConclusionThe preoperative LMR was identified as an independent prognostic factor that could be incorporated into the two nomograms with other risk factors to predict overall survival and PTC-free survival for individual patients.

SummaryAccurate prediction of the solubility of chemical substances in solvents remains a challenge. The sparsity of high-quality solubility data is recognized as the biggest hurdle in the development of robust data-driven methods for practical use. Nonetheless, the effects of the quality and quantity of data on aqueous solubility predictions have not yet been scrutinized. In this study, the roles of the size and the quality of data sets on the performances of the solubility prediction models are unraveled, and the concepts of actual and observed performances are introduced. In an effort to curtail the gap between actual and observed performances, a quality-oriented data selection method, which evaluates the quality of data and extracts the most accurate part of it through statistical validation, is designed. Applying this method on the largest publicly available solubility database and using a consensus machine learning approach, a top-performing solubility prediction model is achieved.

Random forest estimation of genomic breeding values for disease susceptibility over different disease incidences and genomic architectures in simulated cow calibration groups

The paper discusses variable selection as used in large metabolomic studies, exemplified by mouse urinary gas chromatography of 441 mice in three experiments to detect the influence of age, diet, and stress on their chemosignal. Partial least squares discriminant analysis (PLS-DA) was applied to obtain class models, using a procedure of 20,000 iterations including the bootstrap for model optimization and random splits into test and training sets for validation. Variables are selected using PLS regression coefficients on the training set using an optimized number of components obtained from the bootstrap. The variables are ranked in order of significance, and the overall optimal variables are selected as those that appear as highly significant over 100 different test and training set splits. Cost/benefit analysis of performing the model on a reduced number of variables is also illustrated. This paper provides a strategy for properly validated methods for determining which variables are most significant for discriminating between two groups in large metabolomic data sets avoiding the common pitfall of overfitting if variables are selected on a combined training and test set and also taking into account that different variables may be selected each time the samples are split into training and test sets using iterative procedures.

We previously identified a 5 methylated DNA marker (MDM) panel for the detection of nonendoscopic Barrett's esophagus (BE). In this study, we aimed to recalibrate the performance of the 5 MDM panel using a simplified assay in a training cohort, validate the panel in an independent test cohort, and explore the accuracy of an MDM panel with only 3 markers. Participants were recruited from 3 medical centers. The sponge on a string device (EsophaCap; CapNostics, Concord, NC, USA) was swallowed and withdrawn, followed by endoscopy, in BE cases and control subjects. A 5 MDM panel was blindly assayed using a simplified assay. Random forest modeling analysis was performed, in silico cross-validated in the training set, and then locked down, before test set analysis. The training set had 199 patients: 110 BE cases and 89 control subjects, and the test set had 89 patients: 60 BE cases and 29 control subjects. Sensitivity of the 5 MDM panel for BE diagnosis was 93% at 90% specificity in the training set and 93% at 93% specificity in the test set. Areas under the receiver operating characteristic curves were .96 and .97 in the training and test sets, respectively. Model accuracy was not influenced by age, sex, or smoking history. Multiple 3 MDM panels achieved similar accuracy. A 5 MDM panel for BE is highly accurate in training and test sets in a blinded multisite case-control analysis using a simplified assay. This panel may be reduced to only 3 MDMs in the future. (Clinical trial registration number: NCT02560623.).

Identification and verification of four candidate biomarkers for early diagnosis of osteoarthritis by machine learning

ObjectiveThe primary objective of this study was to investigate the risk factors for diabetic peripheral neuropathy (DPN) and to establish an early diagnostic prediction model for its onset, based on clinical data and biochemical indices.MethodsRetrospective data were collected from 1,446 diabetic patients at the First Affiliated Hospital of Anhui University of Chinese Medicine and were split into training and internal validation sets in a 7:3 ratio. Additionally, 360 diabetic patients from the Second Affiliated Hospital were used as an external validation cohort. Feature selection was conducted within the training set, where univariate logistic regression identified variables with a p-value < 0.05, followed by backward elimination to construct the logistic regression model. Concurrently, the random forest algorithm was applied to the training set to identify the top 10 most important features, with hyperparameter optimization performed via grid search combined with cross-validation. Model performance was evaluated using ROC curves, decision curve analysis, and calibration curves. Model fit was assessed using the Hosmer-Lemeshow test, followed by Brier Score evaluation for the random forest model. Ten-fold cross-validation was employed for further validation, and SHAP analysis was conducted to enhance model interpretability.ResultsA nomogram model was developed using logistic regression with key features: limb numbness, limb pain, diabetic retinopathy, diabetic kidney disease, urinary protein, diastolic blood pressure, white blood cell count, HbA1c, and high-density lipoprotein cholesterol. The model achieved AUCs of 0.91, 0.88, and 0.88 for the training, validation, and test sets, respectively, with a mean AUC of 0.902 across 10-fold cross-validation. Hosmer-Lemeshow test results showed p-values of 0.595, 0.418, and 0.126 for the training, validation, and test sets, respectively. The random forest model demonstrated AUCs of 0.95, 0.88, and 0.88 for the training, validation, and test sets, respectively, with a mean AUC of 0.886 across 10-fold cross-validation. The Brier score indicates a good calibration level, with values of 0.104, 0.143, and 0.142 for the training, validation, and test sets, respectively.ConclusionThe developed nomogram exhibits promise as an effective tool for the diagnosis of diabetic peripheral neuropathy in clinical settings.

Vessels Encapsulating Tumor Clusters (VETC) are now recognized as independent indicators of recurrence and overall survival in hepatocellular carcinoma (HCC) patients. However, there has been limited investigation into predicting the VETC pattern using hepatobiliary phase (HBP) features from preoperative gadobenate-enhanced MRI. This study involved 252 HCC patients with confirmed VETC status from three different hospitals (Hospital 1: training set with 142 patients; Hospital 2: test set with 64 patients; Hospital 3: validation set with 46 patients). Independent predictive factors for VETC status were determined through univariate and multivariate logistic analyses. Subsequently, these factors were used to construct two distinct VETC prediction models. Model 1 included all independent predictive factors, while Model 2 excluded HBP features. The performance of both models was assessed using the Area Under the Curve (AUC), Decision Curve Analysis, and Calibration Curve. Prediction accuracy between the two models was compared using Net Reclassification Improvement (NRI) and Integrated Discriminant Improvement (IDI). CA199, IBIL, shape, peritumoral hyperintensity on HBP, and arterial peritumoral enhancement were independent predictors of VETC. Model 1 showed robust predictive performance, with AUCs of 0.836 (training), 0.811 (test), and 0.802 (validation). Model 2 exhibited moderate performance, with AUCs of 0.813, 0.773, and 0.783 in the respective sets. Calibration and decision curves for both models indicated consistent predictions between predicted and actual VETC, benefiting HCC patients. NRI showed Model 1 increased by 0.326, 0.389, and 0.478 in the training, test, and validation sets compared to Model 2. IDI indicated Model 1 increased by 0.036, 0.028, and 0.025 in the training, test, and validation sets compared to Model 2. HBP features from preoperative gadobenate-enhanced MRI can enhance the predictive performance of VETC in HCC.

Glioblastoma is a highly aggressive brain tumor with limited survival that poses challenges in predicting patient outcomes. The Karnofsky Performance Status (KPS) score is a valuable tool for assessing patient functionality and contributes to the stratification of patients with poor prognoses. This study aimed to develop a 6-month postoperative KPS prediction model by combining clinical data with deep learning-based image features from pre- and postoperative MRI scans, offering enhanced personalized care for glioblastoma patients. Using 1,476 MRI datasets from the Brain Tumor Segmentation Challenge 2020 public database, we pretrained two variational autoencoders (VAEs). Imaging features from the latent spaces of the VAEs were used for KPS prediction. Neural network-based KPS prediction models were developed to predict scores below 70 at 6 months postoperatively. In this retrospective single-center analysis, we incorporated clinical parameters and pre- and postoperative MRI images from 150 newly diagnosed IDH wild-type glioblastoma, divided into training (100 patients) and test (50 patients) sets. In training set, the performance of these models was evaluated using the area under the curve (AUC), calculated through fivefold cross-validation repeated 10 times. The final evaluation of the developed models assessed in the test set. Among the 150 patients, 61 had 6-month postoperative KPS scores below 70 and 89 scored 70 or higher. We developed three models: a clinical-based model, an MRI-based model, and a multimodal model that incorporated both clinical parameters and MRI features. In the training set, the mean AUC was 0.785±0.051 for the multimodal model, which was significantly higher than the AUCs of the clinical-based model (0.716±0.059, P = 0.038) using only clinical parameters and the MRI-based model (0.651±0.028, P<0.001) using only MRI features. In the test set, the multimodal model achieved an AUC of 0.810, outperforming the clinical-based (0.670) and MRI-based (0.650) models. The integration of MRI features extracted from VAEs with clinical parameters in the multimodal model substantially enhanced KPS prediction performance. This approach has the potential to improve prognostic prediction, paving the way for more personalized and effective treatments for patients with glioblastoma.