LASSO Regression Modeling on Prediction of Medical Terms among Seafarers’ Health Documents Using Tidy Text Mining

Using automated text classification to explore uncertainty in NICE appraisals for drugs for rare diseases

OBJECTIVE

This study examined the application, feasibility, and validity of supervised learning models for text classification in appraisals for rare disease treatments (RDTs) in relation to uncertainty, and analyzed differences between appraisals based on the classification results.

METHODS

We analyzed appraisals for RDTs (n = 94) published by the National Institute for Health and Care Excellence (NICE) between January 2011 and May 2023. We used Naïve Bayes, Lasso, and Support Vector Machine models in a binary text classification task (classifying paragraphs as either referencing uncertainty in the evidence base or not). To illustrate the results, we tested hypotheses in relation to the appraisal guidance, advanced therapy medicinal product (ATMP) status, disease area, and age group.

RESULTS

The best performing (Lasso) model achieved 83.6 percent classification accuracy (sensitivity = 74.4 percent, specificity = 92.6 percent). Paragraphs classified as referencing uncertainty were significantly more likely to arise in highly specialized technology (HST) appraisals compared to appraisals from the technology appraisal (TA) guidance (adjusted odds ratio = 1.44, 95 percent CI 1.09, 1.90, p = 0.004). There was no significant association between paragraphs classified as referencing uncertainty and appraisals for ATMPs, non-oncology RDTs, and RDTs indicated for children only or adults and children. These results were robust to the threshold value used for classifying paragraphs but were sensitive to the choice of classification model.

CONCLUSION

Using supervised learning models for text classification in NICE appraisals for RDTs is feasible, but the results of downstream analyses may be sensitive to the choice of classification model.

Electrical energy recovery from wastewater: prediction with machine learning algorithms

Wind, solar, biomass, tidal, etc. are renewable energy sources obtained from natural sources. Among these resources, biomass can be characterized as a significant energy source. Today, the process of producing biogas from waste and turning it into electrical energy has become more popular. So, clean, sustainable, and eco-friendly energy is generated as the waste is managed and converted into electrical energy. The estimation of the electrical energy that will be produced by wastewater recovery using machine learning (ML) algorithms is vital and has not yet been investigated. Thus, this study fills this gap. In this study, it is aimed to predict the electrical energy recovery potential of the sewage sludge of Kahramanmaraş Advanced Biological Wastewater Treatment Plant (KABWWTP) (Turkey), through incineration and anaerobic digestion. For this aim, 6 distinct ML algorithms including linear regression (LR), extreme gradient boosting (XGB), Gaussian process regression (GPR), ridge regression (RR), Lasso regression (LASReg), and Bayesian ridge regression (BR) have been used. Another novelty in this study is the restricted number of input parameters. That is, the electrical energy (output parameter) is predicted using only 3 distinct input parameters (gas flow, conductivity, and TSS). With a MAPE value of 1.032, the XGB method has been determined as the most successful model. Heat mapping and correlation analyses are used to evaluate the relationship between these parameters. Performance results are presented in tables and graphs.

Integrated bioinformatics analysis of noncoding RNAs with tumor immune microenvironment in gastric cancer

In recent years, molecular and genetic research hotspots of gastric cancer have been investigated, including microRNAs, long noncoding RNAs (lncRNAs) and messenger RNA (mRNAs). The study on the role of lncRNAs may help to develop personalized treatment and identify potential prognostic biomarkers in gastric cancer. The RNA-seq and miRNA-seq data of gastric cancer were downloaded from the TCGA database. Differential analysis of RNA expression between gastric cancer samples and normal samples was performed using the edgeR package. The ceRNA regulatory network was visualized using Cytoscape. KEGG pathway analysis of mRNAs in the ceRNA network was performed using the clusterProfiler package. CIBERSORT was used to distinguish 22 immune cell types and the prognosis-related genes and immune cells were determined using Kaplan-Meier and Cox proportional hazard analyses. To estimate these nomograms, we used receiver operating characteristic and calibration curve studies. The ceRNA regulation network of gastric cancer was built in this study, and the genes in the network were analyzed for prognosis. A total of 980 lncRNAs were differentially expressed, of which 774 were upregulated and 206 were downregulated. A survival study identified 15 genes associated with gastric cancer prognosis, including VCAN-AS1, SERPINE1, AL139002.1, LINC00326, AC018781.1, C15orf54, hsa-miR-145. Monocytes and Neutrophils were associated with the survival rate of gastric cancer. Our research uncovers new ceRNA network for the detection, treatment, and monitoring of gastric cancer.

Comparison of MRI Sequences to Predict ATRX Status Using Radiomics-Based Machine Learning

ATRX is an important molecular marker according to the 2021 WHO classification of adult-type diffuse glioma. We aim to predict the ATRX mutation status non-invasively using radiomics-based machine learning models on MRI and to determine which MRI sequence is best suited for this purpose. In this retrospective study, we used MRI images of patients with histologically confirmed glioma, including the sequences T1w without and with the administration of contrast agent, T2w, and the FLAIR. Radiomics features were extracted from the corresponding MRI images by hand-delineated regions of interest. Data partitioning into training data and independent test data was repeated 100 times to avoid random effects. Feature preselection and subsequent model development were performed using Lasso regression. The T2w sequence was found to be the most suitable and the FLAIR sequence the least suitable for predicting ATRX mutations using radiomics-based machine learning models. For the T2w sequence, our seven-feature model developed with Lasso regression achieved a mean AUC of 0.831, a mean accuracy of 0.746, a mean sensitivity of 0.772, and a mean specificity of 0.697. In conclusion, for the prediction of ATRX mutation using radiomics-based machine learning models, the T2w sequence is the most suitable among the commonly used MRI sequences.

Clinical features and risk factors of adverse clinical outcomes in central pulmonary embolism using machine learning analysis

BACKGROUND

In prior studies, central pulmonary embolism (PE) was associated with high clot burden and was considered an independent predictor for thrombolysis. Further information about predictors of adverse outcomes in these patients is needed for better risk stratification. The objective is to describe independent predictors of adverse clinical outcomes in patients with central PE.

METHODS

Large retrospective, observational, and single-center study of hospitalized patients with central PE. Data were gathered on demographics, comorbidities, clinical features on admission, imaging, treatments, and outcomes. Multivariable standard and Least Absolute Shrinkage and Selection Operator (LASSO) machine learning logistic regressions and sensitivity analyses were used to analyze factors associated with a composite of adverse clinical outcomes, including vasopressor use, mechanical ventilation, and inpatient mortality.

RESULTS

A total of 654 patients had central PE. The mean age was 63.1 years, 59% were women, and 82% were African American. The composite adverse outcome was observed in 18% (n = 115) of patients. Serum creatinine elevation (odds ratio [OR] = 1.37, 95% CI = 1.20-1.57; p = 0.0001), white blood cell (WBC) count elevation (OR = 1.10, 95% CI = 1.05-1.15; p < 0.001), higher simplified pulmonary embolism severity index (sPESI) score (OR = 1.47, 95% CI = 1.18-1.84; p = 0.001), serum troponin elevation (OR = 1.26, 95% CI 1.02-1.56; p = 0.03), and respiratory rate increase (OR = 1.03, 95% CI = 1.0-1.05; p = 0.02) were independent predictors of adverse clinical outcomes.

CONCLUSION

Among patients with central PE, higher sPESI score, WBC count elevation, serum creatinine elevation, serum troponin elevation, and respiratory rate increase were independent predictors of adverse clinical outcomes. Right ventricular dysfunction on imaging and saddle PE location did not predict adverse outcomes.

Analyzing Community Care Research Trends Using Text Mining

Purpose

This study utilized text mining to analyze research trends around community care, which focuses on improving patients' quality of life by lessening the financial burden on caregivers and relieving patient discomfort.

Methods

To examine research trends by community care stage, Section 1 is set from 2017 to 2019, when the community care was implemented, and Section 2 from 2020 to 2021, after the end of the community care. Papers used for the analysis were extracted using the Korea Citation Index (KCI); a total of 132 articles were selected and subjected to text mining analysis.

Results

First, the main community care research areas included work, housing, economy, disability, and mind. Second, from 2017 to 2019, there was considerable interest in community care centered on households, and main keywords, such as nursing, family, and experience, appeared. Third, from 2020 to the present, there was high interest in community care centered on disabilities, and keywords, such as space, business, and Seoul City, appeared.

Conclusion

The results reveal the changing issues, with the implementation of community care. Overall, research has tended to focus on social and welfare systems, rather than health and medical systems. In the future, local, community-integrated health and medical care systems should be restructured and regional delivery systems established to make them more accessible.

A Fair and Safe Usage Drug Recommendation System in Medical Emergencies by a Stacked ANN

The importance of online recommender systems for drugs, medical professionals, and hospitals is growing. Today, the majority of people use online consultations for drug recommendations for all types of health issues. Emergencies such as pandemics, floods, or cyclones can be helped by the medical recommender system. In the era of machine learning (ML), recommender systems produce more accurate, quick, and reliable clinical predictions with minimal costs. As a result, these systems maintain better performance, integrity, and privacy of patient data in the decision-making process and provide precise information at any time. Therefore, we present drug recommender systems with a stacked artificial neural network (ANN) model to improve the fairness and safety of treatment for infectious diseases. To reduce side effects, drugs are recommended based on a patient’s previous health profile, lifestyle, and habits. The proposed system produced results with 97.5% accuracy. A system such as this could be useful in recommending safe medicines to patients, especially during health emergencies.