Comprehensive clinical application analysis of artificial intelligence-enabled electrocardiograms for screening multiple valvular heart diseases.

BACKGROUND: Valvular heart disease (VHD) is becoming increasingly important to manage the risk of future complications. Electrocardiographic (ECG) changes may be related to multiple VHDs, and (AI)-enabled ECG has been able to detect some VHDs. We aimed to develop five deep learning models (DLMs) to identify aortic stenosis, aortic regurgitation, pulmonary regurgitation, tricuspid regurgitation, and mitral regurgitation. METHODS: Between 2010 and 2021, 77,047 patients with echocardiography and 12-lead ECG performed within 7 days were identified from an academic medical center to provide DLM development (122,728 ECGs), and internal validation (7,637 ECGs). Additional 11,800 patients from a community hospital were identified to external validation. The ECGs were classified as with or without moderate-to-severe VHDs according to transthoracic echocardiography (TTE) records, and we also collected the other echocardiographic data and follow-up TTE records to identify new-onset valvular heart diseases. RESULTS: AI-ECG adjusted for age and sex achieved areas under the curves (AUCs) of >0.84, >0.80, >0.77, >0.83, and >0.81 for detecting aortic stenosis, aortic regurgitation, pulmonary regurgitation, tricuspid regurgitation, and mitral regurgitation, respectively. Since predictions of each DLM shared similar components of ECG rhythms, the positive findings of each DLM were highly correlated with other valvular heart diseases. Of note, a total of 37.5-51.7% of false-positive predictions had at least one significant echocardiographic finding, which may lead to a significantly higher risk of future moderate-to-severe VHDs in patients with initially minimal-to-mild VHDs. CONCLUSION: AI-ECG may be used as a large-scale screening tool for detecting VHDs and a basis to undergo an echocardiography.

AI-enabled electrocardiography alert intervention and all-cause mortality: a pragmatic randomized clinical trial.

The early identification of vulnerable patients has the potential to improve outcomes but poses a substantial challenge in clinical practice. This study evaluated the ability of an artificial intelligence (AI)-enabled electrocardiogram (ECG) to identify hospitalized patients with a high risk of mortality in a multisite randomized controlled trial involving 39 physicians and 15,965 patients. The AI-ECG alert intervention included an AI report and warning messages delivered to the physicians, flagging patients predicted to be at high risk of mortality. The trial met its primary outcome, finding that implementation of the AI-ECG alert was associated with a significant reduction in all-cause mortality within 90 days: 3.6% patients in the intervention group died within 90 days, compared to 4.3% in the control group (4.3%) (hazard ratio (HR) = 0.83, 95% confidence interval (CI) = 0.70-0.99). A prespecified analysis showed that reduction in all-cause mortality associated with the AI-ECG alert was observed primarily in patients with high-risk ECGs (HR = 0.69, 95% CI = 0.53-0.90). In analyses of secondary outcomes, patients in the intervention group with high-risk ECGs received increased levels of intensive care compared to the control group; for the high-risk ECG group of patients, implementation of the AI-ECG alert was associated with a significant reduction in the risk of cardiac death (0.2% in the intervention arm versus 2.4% in the control arm, HR = 0.07, 95% CI = 0.01-0.56). While the precise means by which implementation of the AI-ECG alert led to decreased mortality are to be fully elucidated, these results indicate that such implementation assists in the detection of high-risk patients, prompting timely clinical care and reducing mortality. ClinicalTrials.gov registration: NCT05118035 .

Artificial Intelligence-enabled Chest X-ray Classifies Osteoporosis and Identifies Mortality Risk.

A deep learning model was developed to identify osteoporosis from chest X-ray (CXR) features with high accuracy in internal and external validation. It has significant prognostic implications, identifying individuals at higher risk of all-cause mortality. This Artificial Intelligence (AI)-enabled CXR strategy may function as an early detection screening tool for osteoporosis. The aim of this study was to develop a deep learning model (DLM) to identify osteoporosis via CXR features and investigate the performance and clinical implications. This study collected 48,353 CXRs with the corresponding T score according to Dual energy X-ray Absorptiometry (DXA) from the academic medical center. Among these, 35,633 CXRs were used to identify CXR- Osteoporosis (CXR-OP). Another 12,720 CXRs were used to validate the performance, which was evaluated by the area under the receiver operating characteristic curve (AUC). Furthermore, CXR-OP was tested to assess the long-term risks of mortality, which were evaluated by Kaplan‒Meier survival analysis and the Cox proportional hazards model. The DLM utilizing CXR achieved AUCs of 0.930 and 0.892 during internal and external validation, respectively. The group that underwent DXA with CXR-OP had a higher risk of all-cause mortality (hazard ratio [HR] 2.59, 95% CI: 1.83-3.67), and those classified as CXR-OP in the group without DXA also had higher all-cause mortality (HR: 1.67, 95% CI: 1.61-1.72) in the internal validation set. The external validation set produced similar results. Our DLM uses CXRs for early detection of osteoporosis, aiding physicians to identify those at risk. It has significant prognostic implications, improving life quality and reducing mortality. AI-enabled CXR strategy may serve as a screening tool.

Opportunistic Screening for Asymptomatic Left Ventricular Dysfunction With the Use of Electrocardiographic Artificial Intelligence: A Cost-Effectiveness Approach.

BACKGROUND: The burden of asymptomatic left ventricular dysfunction (LVD) is greater than that of heart failure; however, a cost-effective tool for asymptomatic LVD screening has not been well validated. We aimed to prospectively validate an artificial intelligence (AI)-enabled electrocardiography (ECG) algorithm for asymptomatic LVD detection and evaluate its cost-effectiveness for opportunistic screening. METHODS: In this prospective observational study, patients undergoing ECG at outpatient clinics or health check-ups were enrolled in 2 hospitals in Taiwan. Patients were stratified into LVD (left ventricular ejection fraction ≤ 40%) risk groups according to a previously developed ECG algorithm. The performance of AI-ECG was used to conduct a cost-effectiveness analysis of LVD screening compared with no screening. Incremental cost-effectiveness ratio (ICER) and sensitivity analyses were used to examine the cost-effectiveness and robustness of the results. RESULTS: Among the 29,137 patients, the algorithm demonstrated areas under the receiver operating characteristic curves of 0.984 and 0.945 for detecting LVD within 28 days in the 2 hospital cohorts. For patients not initially scheduled for ECG, the algorithm predicted future echocardiograms (high-risk, 46.2%; medium-risk, 31.4%; low-risk, 14.6%) and LVD (high-risk, 26.2%; medium-risk, 3.4%; low-risk, 0.1%) at 12 months. Opportunistic screening with AI-ECG could result in a negative ICER of -$7,439 for patients aged 65 years, with consistent cost-savings across age groups and particularly in men. Approximately 91.5% of the cases were found to be cost-effective at the willingness-to-pay threshold of $30,000 in the probabilistic analysis. CONCLUSIONS: The use of AI-ECG for asymptomatic LVD risk stratification is promising, and opportunistic screening in outpatient clinics has the potential to reduce costs.

Artificial Intelligence-Enabled Electrocardiography Detects B-Type Natriuretic Peptide and N-Terminal Pro-Brain Natriuretic Peptide.

BACKGROUND: The B-type natriuretic peptide (BNP) and N-terminal pro-brain natriuretic peptide (pBNP) are predictors of cardiovascular morbidity and mortality. Since the artificial intelligence (AI)-enabled electrocardiogram (ECG) system is widely used in the management of many cardiovascular diseases (CVDs), patients requiring intensive monitoring may benefit from an AI-ECG with BNP/pBNP predictions. This study aimed to develop an AI-ECG to predict BNP/pBNP and compare their values for future mortality. METHODS: The development, tuning, internal validation, and external validation sets included 47,709, 16,249, 4001, and 6042 ECGs, respectively. Deep learning models (DLMs) were trained using a development set for estimating ECG-based BNP/pBNP (ECG-BNP/ECG-pBNP), and the tuning set was used to guide the training process. The ECGs in internal and external validation sets belonging to nonrepeating patients were used to validate the DLMs. We also followed-up all-cause mortality to explore the prognostic value. RESULTS: The DLMs accurately distinguished mild (≥500 pg/mL) and severe (≥1000 pg/mL) an abnormal BNP/pBNP with AUCs of ≥0.85 in the internal and external validation sets, which provided sensitivities of 68.0-85.0% and specificities of 77.9-86.2%. In continuous predictions, the Pearson correlation coefficient between ECG-BNP and ECG-pBNP was 0.93, and they were both associated with similar ECG features, such as the T wave axis and correct QT interval. ECG-pBNP provided a higher all-cause mortality predictive value than ECG-BNP. CONCLUSIONS: The AI-ECG can accurately estimate BNP/pBNP and may be useful for monitoring the risk of CVDs. Moreover, ECG-pBNP may be a better indicator to manage the risk of future mortality.

Mortality risk prediction of the electrocardiogram as an informative indicator of cardiovascular diseases.

Background: The electrocardiogram (ECG) may be the most popular test in the management of cardiovascular disease (CVD). Although wide applications of artificial intelligence (AI)-enabled ECG have been developed, an integrating indicator for CVD risk stratification was not investigated. Since mortality may be the most important global outcome, this study aimed to develop a survival deep learning model (DLM) to establish a critical ECG value and explore the associations with various CVD events. Methods: We trained a DLM with 451,950 12-lead resting ECGs obtained from 210,552 patients, for whom 23,592 events occurred. The internal validation set included 27,808 patients with one ECG for each patient. The external validations were performed in a community hospital with 33,047 patients and two transnational data sets with 233,647 and 1631 ECGs. We distinguished the cause of mortality and additionally investigated CVD-related outcomes, including new-onset acute myocardial infarction (AMI), stroke (STK), and heart failure (HF). Results: The DLM achieved C-indices of 0.858/0.836 in internal/external validation sets by using ECG over a 10-year period. The high-mortality-risk group identified by the proposed DLM presented a hazard ratio (HR) of 14.16 (95% confidence interval (CI): 11.33-17.70) compared to the low-risk group in the internal validation and presented a higher risk of cardiovascular (CV) mortality (HR: 18.50, 95% CI: 9.82-34.84), non-CV mortality (HR: 13.68, 95% CI: 10.76-17.38), AMI (HR: 4.01, 95% CI: 2.24-7.17), STK (HR: 2.15, 95% CI: 1.70-2.72), and HF (HR: 6.66, 95% CI: 4.54-9.77), which was consistent in an independent community hospital. The transnational validation also revealed HRs of 4.91 (95% CI: 2.63-9.16) and 2.29 (95% CI: 2.15-2.44) for all-cause mortality in the SaMi-Trop and Clinical Outcomes in Digital Electrocardiography 15% (CODE15) cohorts. Conclusions: The mortality risk by AI-enabled ECG may be applied in passive electronic-health-record-based CVD risk screening, which may identify more asymptomatic and unaware high-risk patients.

An AI-Enabled Dynamic Risk Stratification for Emergency Department Patients with ECG and CXR Integration.

Emergency department (ED) triage scale determines the priority of patient care and foretells the prognosis. However, the information retrieved from the initial assessment is limited, hindering the risk identification accuracy of triage. Therefore, we sought to develop a 'dynamic' triage system as secondary screening, using artificial intelligence (AI) techniques to integrate information from initial assessment data and subsequent examinations. This retrospective cohort study included 134,112 ED visits with at least one electrocardiography (ECG) and chest X-ray (CXR) in a medical center from 2012 to 2022. Additionally, an independent community hospital provided 45,614 ED visits as an external validation set. We trained an eXtreme gradient boosting (XGB) model using initial assessment data to predict all-cause mortality in 7 days. Two deep learning models (DLMs) using ECG and CXR were trained to stratify mortality risks. The dynamic triage levels were based on output from the XGB-triage and DLMs from ECG and CXR. During the internal and external validation, the area under the receiver operating characteristic curve (AUC) of the XGB-triage model was >0.866; furthermore, the AUCs of DLMs using ECG and CXR were >0.862 and >0.886, respectively. The dynamic triage scale provided a higher C-index (0.914-0.920 vs. 0.827-0.843) than the original one and demonstrated better predictive ability for 5-year mortality, 30-day ED revisit, and 30-day discharge. The AI-based risk scale provides a more accurate and dynamic stratification of mortality risk in ED patients, particularly in identifying patients who tend to be overlooked due to atypical symptoms.

Artificial Intelligence-Enabled Electrocardiography Detects Hypoalbuminemia and Identifies the Mechanism of Hepatorenal and Cardiovascular Events.

Background: Albumin, an important component of fluid balance, is associated with kidney, liver, nutritional, and cardiovascular diseases (CVD) and is measured by blood tests. Since fluid balance is associated with electrocardiography (ECG) changes, we established a deep learning model (DLM) to estimate albumin via ECG. Objective: This study aimed to develop a DLM to estimate albumin via ECG and explored its contribution to future complications. Materials and Methods: A DLM was trained for estimating ECG-based albumin (ECG-Alb) using 155,078 ECGs corresponding to albumin from 79,111 patients, and another independent 13,335 patients from an academic medical center and 11,370 patients from a community hospital were used for internal and external validation. The primary analysis focused on distinguishing patients with mild to severe hypoalbuminemia, and the secondary analysis aimed to provide additional prognostic value from ECG-Alb for future complications, which included mortality, new-onset hypoalbuminemia, chronic kidney disease (CKD), new onset hepatitis, CVD mortality, new-onset acute myocardial infarction (AMI), new-onset stroke (STK), new-onset coronary artery disease (CAD), new-onset heart failure (HF), and new-onset atrial fibrillation (Afib). Results: The AUC to identify hypoalbuminemia was 0.8771 with a sensitivity of 56.0% and a specificity of 90.7% in the internal validation set, and the Pearson correlation coefficient was 0.69 in the continuous analysis. The most important ECG features contributing to ECG-Alb were ordered in terms of heart rate, corrected QT interval, T wave axis, sinus rhythm, P wave axis, etc. The group with severely low ECG-Alb had a higher risk of all-cause mortality [hazard ratio (HR): 2.45, 95% CI: 1.81-3.33] and the other hepatorenal and cardiovascular events in the internal validation set. The external validation set yielded similar results. Conclusion: Hypoalbuminemia and its complications can be predicted using ECG-Alb as a novel biomarker, which may be a non-invasive tool to warn asymptomatic patients.

Development and Validation of an Artificial Intelligence Electrocardiogram Recommendation System in the Emergency Department.

The machine learning-assisted electrocardiogram (ECG) is increasingly recognized for its unprecedented capabilities in diagnosing and predicting cardiovascular diseases. Identifying the need for ECG examination early in emergency department (ED) triage is key to timely artificial intelligence-assisted analysis. We used machine learning to develop and validate a clinical decision support tool to predict ED triage patients' need for ECG. Data from 301,658 ED visits from August 2017 to November 2020 in a tertiary hospital were divided into a development cohort, validation cohort, and two test cohorts that included admissions before and during the COVID-19 pandemic. Models were developed using logistic regression, decision tree, random forest, and XGBoost methods. Their areas under the receiver operating characteristic curves (AUCs), positive predictive values (PPVs), and negative predictive values (NPVs) were compared and validated. In the validation cohort, the AUCs were 0.887 for the XGBoost model, 0.885 for the logistic regression model, 0.878 for the random forest model, and 0.845 for the decision tree model. The XGBoost model was selected for subsequent application. In test cohort 1, the AUC was 0.891, with sensitivity of 0.812, specificity of 0.814, PPV of 0.708 and NPV of 0.886. In test cohort 2, the AUC was 0.885, with sensitivity of 0.816, specificity of 0.812, PPV of 0.659, and NPV of 0.908. In the cumulative incidence analysis, patients not receiving an ECG yet positively predicted by the model had significantly higher probability of receiving the examination within 48 h compared with those negatively predicted by the model. A machine learning model based on triage datasets was developed to predict ECG acquisition with high accuracy. The ECG recommendation can effectively predict whether patients presenting at ED triage will require an ECG, prompting subsequent analysis and decision-making in the ED.

Clinical Characteristics of COVID-19 Patients and Application to an Artificial Intelligence System for Disease Surveillance.

During the coronavirus disease (COVID-19) pandemic, we admitted suspected or confirmed COVID-19 patients to our isolation wards between 2 March 2020 and 4 May 2020, following a well-designed and efficient assessment protocol. We included 217 patients suspected of COVID-19, of which 27 had confirmed COVID-19. The clinical characteristics of these patients were used to train artificial intelligence (AI) models such as support vector machine (SVM), decision tree, random forest, and artificial neural network for diagnosing COVID-19. When analyzing the performance of the models, SVM showed the highest sensitivity (SVM vs. decision tree vs. random forest vs. artificial neural network: 100% vs. 42.86% vs. 28.57% vs. 71.43%), while decision tree and random forest had the highest specificity (SVM vs. decision tree vs. random forest vs. artificial neural network: 88.37% vs. 100% vs. 100% vs. 94.74%) in the diagnosis of COVID-19. With the aid of AI models, physicians may identify COVID-19 patients earlier, even with few baseline data available, and segregate infected patients earlier to avoid hospital cluster infections and to ensure the safety of medical professionals and ordinary patients in the hospital.

The Effect of PPARγ rs1801282 Variant on Mortality Risk Among Asians With Chronic Kidney Disease: A Cohort Study and Meta-Analysis.

Background: Chronic kidney disease (CKD) is a public health issue, and an independent risk factor for cardiovascular disease. The peroxisome proliferator-activated receptor gamma (PPARG) plays an important role in the cardiovascular system. Previous studies have examined one important exon polymorphism, Pro12Ala, in PPARG with respect to mortality of CKD patients, but the results were inconsistent and current evidence is insufficient to support a strong conclusion. This study aimed to examine the correlation between Pro12Ala gene polymorphism and mortality among Asians with CKD by trial sequential analysis (TSA). Methods: The research was divided into observational research and meta-analysis. For the cohort study, 767 subjects from dialysis centers in Taipei were selected as samples, and tracked from December 2015 to February 2017. For the meta-analysis, relevant literature from "PubMed" and "Embase" databases (until December 2016), was searched and TSA was used to verify the results. In order to achieve the best evidence hierarchies, our retrospective cohort study was added to the meta-analysis and the TSA. Results: The combined sample size for Asian was 1,685 after adding our cohort study, and there was no significant correlation between PPARG Pro12Ala and mortality by the allele model (RR: 0.85, 95% CI: 0.39-1.83, I2 = 79.3%). Under the parameter setting with the RR value of 1.5, TSA estimation presented that the cumulative sample size entered into the futility area, and it confirmed the conclusion in this study. Conclusion: We found that PPARG Pro12Ala gene polymorphism was not related to mortality in CKD Asians patients, and validated our conclusion using TSA after adding our sample.

The Polymorphism at PLCB4 Promoter (rs6086746) Changes the Binding Affinity of RUNX2 and Affects Osteoporosis Susceptibility: An Analysis of Bioinformatics-Based Case-Control Study and Functional Validation.

Purpose: Genome-wide association studies have identified numerous genetic variants that are associated with osteoporosis risk; however, most of them are present in the non-coding regions of the genome and the functional mechanisms are unknown. In this study, we aimed to investigate the potential variation in runt domain transcription factor 2 (RUNX2), which is an osteoblast-specific transcription factor that normally stimulates bone formation and osteoblast differentiation, regarding variants within RUNX2 binding sites and risk of osteoporosis in postmenopausal osteoporosis (PMOP). Methods: We performed bioinformatics-based prediction by combining whole genome sequencing and chromatin immunoprecipitation sequencing to screen functional SNPs in the RUNX2 binding site using data from the database of Taiwan Biobank; Case-control studies with 651 postmenopausal women comprising 107 osteoporosis patients, 290 osteopenia patients, and 254 controls at Tri-Service General Hospital between 2015 and 2019 were included. The subjects were examined for bone mass density and classified into normal and those with osteopenia or osteoporosis by T-scoring with dual-energy X-ray absorptiometry. Furthermore, mRNA expression and luciferase reporter assay were used to provide additional evidence regarding the associations identified in the association analyses. Chi-square tests and logistic regression were mainly used for statistical assessment. Results: Through candidate gene approaches, 3 SNPs in the RUNX2 binding site were selected. A novel SNP rs6086746 in the PLCB4 promoter was identified to be associated with osteoporosis in Chinese populations. Patients with AA allele had higher risk of osteoporosis than those with GG+AG (adjusted OR = 6.89; 95% confidence intervals: 2.23-21.31, p = 0.001). Moreover, the AA genotype exhibited lower bone mass density (p < 0.05). Regarding mRNA expression, there were large differences in the correlation between PLCB4 and different RUNX2 alleles (Cohen's q = 0.91). Functionally, the rs6086746 A allele reduces the RUNX2 binding affinity, thus enhancing the suppression of PLCB4 expression (p < 0.05). Conclusions: Our results provide further evidence to support the important role of the SNP rs6086746 in the etiology of osteopenia/osteoporosis, thereby enhancing the current understanding of the susceptibility to osteoporosis. We further studied the mechanism underlying osteoporosis regulation by PLCB4.

Genetic association between TNF-α G-308A and osteoarthritis in Asians: A case-control study and meta-analysis.

BACKGROUND: Osteoarthritis (OA) is an important health issue in elderly people. Many studies have suggested that genetic factors are important risk factors for OA, of which tumor necrosis factor-α (TNF-α) is one of the most examined genes. Moreover, several studies have investigated the relationship between TNF-α G-308A polymorphisms and OA risk, but consistent results have not been obtained. OBJECTIVE: This study examines the association between TNF-α G-308A polymorphisms and knee OA. Moreover, meta-analysis and trial sequential analysis (TSA) was used to determine whether this is a susceptibility gene for knee OA. METHODS: Between 2015 and 2019, 591 knee OA cases and 536 healthy controls were recruited. The Kellgren-Lawrence grading system was used to identify the knee OA cases. A meta-analysis was conducted including related studies published until 2020 from PubMed, Embase, and previous meta-analysis to improve the evidence level of the current study. The results were expressed as odds ratios (ORs) with corresponding 95% confidence intervals (CI) to evaluate the effect of this polymorphism on knee OA risk. The TSA was used to estimate the sample sizes required in this issue. RESULTS: A nonsignificant association was found between the AA genotype and knee OA [adjusted OR, 0.84; 95% CI, 0.62-1.15) in the recessive model] in the present case-control study, and analysis of other genetic models showed a similar trend. After adding the critical case-control samples for Asians, the TNF-α G-308A, AA genotype exhibited 2.57 times more risk of developing arthritis when compared with the GG + GA genotype (95% CI, 1.56-4.23), and the cumulative samples for TSA (n = 2182) were sufficient to obtain a definite conclusion. CONCLUSIONS: The results of this meta-analysis revealed that the TNF-α G-308A, AA genotype is a susceptible genotype for OA in the Asian population. This study integrated all current evidence to arrive at this conclusion, suggesting that future studies on Asians are not required.

Association between nitric oxide synthase T-786C genetic polymorphism and chronic kidney disease: Meta-analysis incorporating trial sequential analysis.

BACKGROUND: Several meta-analyses of the relationship between endothelial nitric oxide synthase (eNOS) T-786C gene polymorphism and chronic kidney disease (CKD) have been published. However, the results of these studies were inconsistent, and it is undetermined whether sample sizes are sufficient to reach a definite conclusion. OBJECTIVE: To elucidate the relationship between T-786C and CKD by combining previous studies with our case-control sample and incorporate trial sequential analysis (TSA) to verify whether the sample size is adequate to draw a definite conclusion. METHODS: PubMed and Embase databases were searched for relevant articles on eNOS T-786C and CKD before February 28, 2021. TSA was also incorporated to ascertain a conclusion. A total of 558 hemodialysis cases in the case-control study was recruited from nine dialysis centers in the northern area of Taiwan in 2020. Additionally, 640 healthy subjects of the control group, with estimated glomerular filtration rate (eGFR) ≥ 60 mL/min/1.73 m2, were selected from participants of the annual elderly health examination program at the Tri-Service General Hospital. The functional analysis was based on eQTL data from GTExPortal. RESULTS: After screening with eligibility criteria, 15 papers were included and eventually combined in a meta-analysis. The result of the TSA showed that the sample size for Caucasians was adequate to ascertain the correlation between eNOS T-786C and CKD but was insufficient for Asians. Therefore, we added our case-control samples (n = 1198), though not associated with CKD (odds ratio [OR] = 1.01, 95% confidence interval [CI] = 0.69-1.46), into a meta-analysis, which supported that eNOS T-786C was significantly associated with CKD in Asians (OR = 1.39, 95% CI = 1.04-1.85) by using an adequate cumulative sample size (n = 4572) analyzed by TSA. Data of eQTL from GTEx showed that T-786C with the C minor allele exhibited relatively lower eNOS mRNA expression in whole blood, indicating the hazardous role of eNOS T-786C in CKD. CONCLUSIONS: eNOS T-786C genetic polymorphism was of conclusive significance in the association with CKD among Asians in our meta-analysis. Our case-control samples play a decisive role in changing conclusions from indefinite to definite.

The Application of Projection Word Embeddings on Medical Records Scoring System.

Medical records scoring is important in a health care system. Artificial intelligence (AI) with projection word embeddings has been validated in its performance disease coding tasks, which maintain the vocabulary diversity of open internet databases and the medical terminology understanding of electronic health records (EHRs). We considered that an AI-enhanced system might be also applied to automatically score medical records. This study aimed to develop a series of deep learning models (DLMs) and validated their performance in medical records scoring task. We also analyzed the practical value of the best model. We used the admission medical records from the Tri-Services General Hospital during January 2016 to May 2020, which were scored by our visiting staffs with different levels from different departments. The medical records were scored ranged 0 to 10. All samples were divided into a training set (n = 74,959) and testing set (n = 152,730) based on time, which were used to train and validate the DLMs, respectively. The mean absolute error (MAE) was used to evaluate each DLM performance. In original AI medical record scoring, the predicted score by BERT architecture is closer to the actual reviewer score than the projection word embedding and LSTM architecture. The original MAE is 0.84 ± 0.27 using the BERT model, and the MAE is 1.00 ± 0.32 using the LSTM model. Linear mixed model can be used to improve the model performance, and the adjusted predicted score was closer compared to the original score. However, the project word embedding with the LSTM model (0.66 ± 0.39) provided better performance compared to BERT (0.70 ± 0.33) after linear mixed model enhancement (p < 0.001). In addition to comparing different architectures to score the medical records, this study further uses a mixed linear model to successfully adjust the AI medical record score to make it closer to the actual physician's score.

The Decisive Case-Control Study Elaborates the Null Association between ESR1 XbaI and Osteoarthritis in Asians: A Case-Control Study and Meta-Analysis.

(1) Background: The prevalence of knee osteoarthritis (OA) in women is significantly higher than in men. The estrogen receptor α (ERα) has been considered to play a key role due to a large gender difference in its expression. ERα is encoded by the gene estrogen receptor 1 (ESR1), which is widely studied to explore the gender difference in knee OA. Several polymorphisms in ESR1 [PvuII (rs2234693) and BtgI (rs2228480)] were confirmed as the risk factors of OA. However, the evidence of the last widely investigated polymorphism, ESR1 Xbal (rs9340799), is still insufficient for concluding its effect on knee OA. (2) Objective: This study proposed a case-control study to investigate the association between ESR1 Xbal and knee OA. Moreover, a meta-analysis and trial sequential analysis (TSA) were conducted to enlarge the sample size for obtaining a conclusive evidence. (3) Methods: In total, 497 knee OA cases and 473 healthy controls were recruited between March 2015 and July 2018. The Kellgren-Lawrence grading system was used to identify the knee OA cases. To improve the evidence level of our study, we conducted a meta-analysis including the related studies published up until December 2018 from PubMed, Embase, and previous meta-analysis. The results are expressed as odds ratios (ORs) with corresponding 95% confidence intervals (CI) for evaluating the effect of this polymorphism on knee OA risk. TSA was used to estimate the sample sizes required in this issue. (4) Results: We found non-significant association between the G allele and knee OA [Crude-OR: 0.97 (95% CI: 0.78-1.20) and adjusted-OR: 0.90 (95% CI: 0.71-1.15) in allele model] in the present case-control study, and the analysis of other genetic models showed a similar trend. After including six published studies and our case-control studies, the current evidence with 3174 Asians showed the conclusively null association between ESR1 XbaI and knee OA [OR: 0.78 (95% CI: 0.59-1.04)] with a high heterogeneity (I2: 78%). The result of Caucasians also concluded the null association [OR: 1.05 (95% CI: 0.56-1.95), I2: 87%]. (5) Conclusions: The association between ESR1 XbaI and knee OA was not similar with other polymorphisms in ESR1, which is not a causal relationship. This study integrated all current evidence to elaborate this conclusion for suggesting no necessity of future studies.

Polymorphism of rs6426749 at 1p36.12 is associated with the risk of osteoarthritis in Taiwanese female population.

BACKGROUND: Osteoarthritis (OA) is a multifactorial disease that is associated with several genetic factors. TFAP2A with a motif of C allele at rs6426749 demonstrates a higher binding ability, thereby increasing CDC42 expression, potentially affecting OA occurrence. In this study, we evaluated the role of rs6426749 polymorphisms on knee OA in a female Taiwanese population. METHODS: We performed a case-control study of 368 OA cases and 379 controls between March 2017 and October 2018. Knee OA was defined using the Kellgren-Lawrence grading system, and genotypes were determined using the Sequenom MassArray iPLEX Gold assay. Stratified sex and body mass index (BMI) analyses were performed using logistic regression to explore interactions between genes and the environment. We also used expression quantitative trait loci data from the genotype-tissue expression project to conduct functional analyses. RESULTS: The C allele of rs6426749 was associated with the risk of knee OA (odds ratio [OR] = 1.31, 95% confidence interval [CI], 1.01-1.71; p = 0.042), after adjusting for gender, age, and BMI. In addition, subgroup analyses indicated that females expressing C alleles showed an increased risk for knee OA (OR = 1.56; 95% CI, 1.12-2.18; p = 0.009). Females with a normal BMI and the C allele had the highest OA risk (OR = 1.73; 95% CI, 1.08-2.76; p = 0.022). CONCLUSION: Our findings indicated that rs6426749 may be related to OA susceptibility in the Taiwanese population. This was particularly true for women with normal BMI.

Decisive evidence corroborates a null relationship between MTHFR C677T and chronic kidney disease: A case-control study and a meta-analysis.

BACKGROUND: Previous meta-analyses have explored the association between the C677T polymorphism of methyltetrahydrofolate reductase (MTHFR) and chronic kidney disease (CKD) but there were no studies with a decisive conclusion. Furthermore, the high heterogeneity among different populations is not yet interpreted. OBJECTIVES: This study used trial sequential analysis (TSA) to evaluate whether the nowadays conclusion supported by current cumulative samples. We also applied case-weighted meta-regression to explore the potential gene-environment interactions. METHODS: For the first stage of this study we conducted a case-control study involving 847 dialysis patients from 7 hemodialysis centers in Taipei during 2015 to 2018 and 755 normal controls from a health center in the Tri-Service General Hospital. The second stage combined the results from the first stage with previous studies. The previous studies were collected from PubMed, EMBASE, and Web of Science databases before January 2018. RESULTS: From the case-control study, the T allele of MTHFR C677T appeared to have a protective effect on end-stage renal disease compared with the C allele [odds ratio (OR): 0.80, 95% CI (confidence interval) = 0.69-0.93]. However, the meta-analysis contradicted the results in Asian (OR = 1.12, 95% CI = 0.96-1.30). The same analysis was also applied in Caucasian and presented similar results from Asian (OR = 1.18, 95% CI = 0.98-1.42). The TSA showed our case-control study to be the decisive sample leading to a null association among Asian population. The high heterogeneity (I = 75%) could explain the contradictory results between the case-control study and the meta-analysis. However, further case-weighted meta-regression did not find any significant interaction between measured factors and MTHFR C677T on CKD. CONCLUSIONS: High heterogeneities were found in both Caucasian and Asian, which caused the null relationship in meta-analysis while there were significant effects in individual studies. Future studies should further explore the high heterogeneity that might be hidden in unmeasured gene-environment interactions, to explain the diverse findings among different populations.

PPARG Pro12Ala Polymorphism with CKD in Asians: A Meta-Analysis Combined with a Case-Control Study-A Key for Reaching Null Association.

BACKGROUND: So far, numerous meta-analyses have been published regarding the correlation between peroxisome proliferator-activated receptor gamma (PPARG) proline 12 alanine (Pro12Ala) gene polymorphism and chronic kidney disease (CKD); however, the results appear to be contradictory. Hence, this study is formulated with the objective of using existing meta-analysis data together with our research population to study the correlation between PPARG Pro12Ala gene polymorphism and CKD and evaluate whether an accurate result can be obtained. METHODS: First, literature related to CKD and PPARG Pro12Ala available on the PubMed and EMBASE databases up to December 2016 was gathered from 20 publications. Then, the gathered results were combined with our case-control study of 1693 enrolled subjects and a trial sequential analysis (TSA) was performed to verify existing evidence and determine whether a firm conclusion can be drawn. RESULTS: The TSA results showed that the cumulative sample size for the Asian sample was 6078 and was sufficient to support a definite result. The results of this study confirmed that there is no obvious correlation between PPARG Pro12Ala and CKD for Asians (OR = 0.82 (95% CI = 0.66-1.02), I2 = 63.1%), but this was not confirmed for Caucasians. Furthermore, the case-control sample in our study was shown to be the key for reaching this conclusion. CONCLUSIONS: The meta-analysis results of this study suggest no significant correlation between PPARG Pro12Ala gene polymorphism and CKD for Asians after adding our samples, but not for Caucasian.

Accuracy augmentation of body composition measurement by bioelectrical impedance analyzer in elderly population.

Bioelectrical impedance analysis (BIA) is currently the most commonly used method in clinical practice to measure body composition. However, the bioelectrical impedance analyzer is not designed according to different countries, races, and elderly populations. Because different races may have different body compositions, a prediction model for the elderly population in Taiwan should be developed to avoid population bias, thereby improving the accuracy of community evaluation surveys.Dual energy X-ray absorptiometry (DXA) was used as a standard method for comparison, and impedance analysis was used for the development of a highly accurate predictive model that is suitable for assessing the body composition of elderly people.This study employed a cross-sectional design and recruited 438 elderly people who were undergoing health examinations at the health management center in the Tri-Service General Hospital as study subjects. Basic demographic variables and impedance analysis values were used in four predictive models, namely, linear regression, random forest, support vector machine (SVM), and eXtreme Gradient Boosting (XGBoost) models, to predict DXA body composition. The data from 354 study subjects were used to develop the predictive model, while the data from 84 study subjects were used to validate the accuracy of the predictive model.The body composition of elderly people as estimated by InBody 720 was highly correlated with that estimated by DXA. The correlation coefficient between InBody 720 and DXA for muscle mass was 0.969, and that for fat mass was 0.935. Consistency analysis results showed that InBody 720 tends to underestimate muscle mass and fat mass. A comparison of the accuracy of the linear regression, random forest, SVM, and XGBoost models showed that the linear regression has the highest accuracy. The correlation coefficient between the new model and DXA for muscle mass and fat mass were 0.977 and 0.978, respectively.The new predictive model can be used to monitor the nutrition status of elderly people and identify people with sarcopenia in the community.