Ways to Generate Synthetic Data for AI Training without Leaking Information

Body fat assessment is needed in individuals with HIV. The objective was to identify the discriminatory capacity of the abdominal skinfold (ASF) tricipital skinfold (TSF), subscapular fold (SSF), calf skinfold (CSF), body adiposity index (BAI), body mass index, conicity index (IC), mid-upper arm circumference (MUAC), waist circumference (WC), perimeter of neck (PN) and waist-to-height ratio (WHtR) for high body fat in children and adolescents with HIV, compared Dual energy X-ray absorptiometry (DXA) and air displacement plethysmography (ADP). Descriptive study, cross - sectional study, with 65 children and adolescents with HIV by vertical transmission. Body fat was measured by DXA and ADP. Measures were measured by international standardization. The diagnostic properties for high body fat were assessed by area under the ROC curve (AUC). For boys, having DXA as a reference for fat, ASF (AUC: 0.920), TSF (AUC: 0.792), SSF (AUC: 0.766), CSF (AUC: 0.866), BAI satisfactory discriminatory capacity. With ADP as the reference method, ASF (AUC: 0.920), TSF (AUC: 0.921), SSF (AUC: 0.766), CSF (AUC: 0.901), BAI (AUC: 0.756) and BMI (AUC: 0.699) presented satisfactory results. For girls, having DXA as a reference for fat, ASF (AUC: 0.838), TSF (AUC: 0.842), SSF (AUC: 0.840), CSF (AUC: 0.887), BAI (AUC: 0.846), and BMI (AUC: 0.859) presented satisfactory discriminatory capacity. Assuming ADP as a reference for fat, ASF (AUC [AUC: 0.799], TSF [AUC: 0.825], SSF [AUC: 0.767], CSF [AUC: 0.897], BAI 0.788), were satisfactory. The ASF, TSF, SSF, CSF, BAI and BMI anthropometric indicators may be suggested as the most suitable for the detection of high body fat in children and adolescents with HIV.

The objective was to study the dynamics of B-type natriuretic peptide (BNP) and its relationship with hemodynamic parameters during on-pump coronary artery bypass grafting (CABG), and to evaluate the informativeness of the biomarker as a predictor of myocardial dysfunction.Materials and methods. The study involved 127 patients aged 59 [54–66.75] years with ischemic heart disease who underwent CABG. The BNP blood level was determined in the operating room at stages: I – before surgery (BNP1 ), II – at the end of surgery (BNP2 ). Hemodynamic parameters were analyzed at the same stages. Correlation analysis, logistic regression with the calculation of the odds ratio (OR) and 95% confidence interval (95% CI) and ROC analysis with the calculation of the area under the ROC curve (AUC) were used.Results. BNP1 blood level was 49 [25.6–91.6], BNP2 – 90 [47.8–140.2] pg/ml (p < 0.0001). BNP1 correlated with central venous pressure (CVP) at stage I (rho = 0.212; p = 0.017) and with pulmonary artery wedge pressure (PAWP) at stage II (rho = 0.204; p = 0.045). BNP2 correlated with PAWP at stage II (rho = 0.204; p = 0.045). BNP1 > 52.1 pg/ml was the predictor of ICU length of stay > 24 hours (OR 1.0290, 95% CI 1.0154– 1.0427, p < 0.0001, AUC 0.775), BNP1 > 71 pg/ml was the predictor of inotropic index > 5 c. u. (OR 1.0076, 95% CI 1.0015–1.0138, p = 0.014, AUC 0.705) and BNP1 > 90.8 pg/ml was the predictor of vasoactive inotropic index > 10 c. u. (OR 1.0070, 95% CI 1.0014–1.0126, p = 0.013, AUC 0.727). BNP2 > 67.5 pg/ml was the predictor of ICU length of stay > 24 hours (OR 1.0179, 95% CI 1.0073–1.0287, p < 0.0009, AUC 0.763), BNP2 > 94.3 pg/ml was the predictor of inotropic index > 5 c. u. (OR 1.0063, 95% CI 1.0010–1.0117, p = 0.020, AUC 0.713), BNP2 > 144 pg/ml was the predictor of intra-aortic balloon pumping (OR 1.0037, 95% CI 1 .0000–1.0074, p = 0.048, AUC 0.854), BNP2 > 159 pg/ml was the predictor of vasoactive inotropic index > 10 c. u. (OR 1.0072, 95% CI 1.0006–1.0139, p = 0.033, AUC 0.729) and BNP2 > 161 pg/ml was the predictor of early mortality in the ICU (OR 1.0040, 95% CI 1, 0000-1.0080, p = 0.049, AUC 0.845). Conclusion. In 78.7% of patients undergoing on-pump CABG, BNP blood level does not exceed the upper limit of normal; by the end of surgery, the biomarker level increases by 32.9 [17.7–62.0] pg/ml. Before and at the end of surgery, BNP values are weakly correlated with CVP and PAWP and do not correlate with other hemodynamic parameters. Before surgery, BNP blood level in the range of 52.1–90.8 pg/ml are predictors of ICU stay > 24 hours (AUC 0.775), inotropic scale > 5 (AUC 0.705) and vasoactive-inotropic scale > 10 c. u. (AUC 0.727). At the end of surgery, BNP > 67.5 pg/ml is associated with an ICU stay > 24 hours (AUC 0.763), and BNP > 90.4 pg/ml is associated with inotropic scale > 5 c. u. (AUC 0.713). The BNP, increased to 144.0–161.0 pg/ml, indicates severe myocardial dysfunction, including hemodynamic support with intra-aortic balloon pumping (AUC 0.854), vasoactive-inotropic scale > 10 c. u. (AUC 0.729) and the risk of early mortality in the ICU (AUC 0.845).

Introduction Atrial fibrillation (AF) is the most common arrhythmia, linked to increased risks of stroke, heart failure, and cardiovascular mortality. Advances in machine learning and artificial intelligence have led to the development of ECG-based neural networks (NN) as promising tools for AF risk stratification. Purpose This analysis aims to evaluate and compare the predictive accuracy of two ECG-based NN models in forecasting incident AF across distinct clinical subgroups within two independent population-based cohorts. Methods The predictive performance of two NN models, Model 1 (a single-lead ECG NN) and Model 2 (a 12-channel ECG NN), was evaluated on resting ECGs from the combined cohorts (N = 4,943; incident AF = 67 at follow-up after 5-7 years), excluding individuals with AF at baseline. Predictive accuracy for a 5 to 7-year risk of incident AF was determined using the area under the ROC curve (AUC). For the combined model, a scaled average from Model 1 and Model 2 predictive scores was used to analyse the joint performance. Subgroup analyses examined model performances across clinical factors, including age, body mass index, lipid profiles (total cholesterol, LDL, HDL), levels of N-terminal prohormone of brain natriuretic peptide (NT-proBNP), cardiovascular comorbidities (e.g., hypertension, diabetes, stroke, and myocardial infarction), and medication use. Results The combined model achieved the highest overall accuracy (AUC = 0.84), while Model 1 (AUC = 0.80) and Model 2 (AUC = 0.81) also showed strong predictive accuracy. Subgroup analysis revealed varying levels of predictive accuracy across specific groups. Higher predictive accuracy was observed in individuals with myocardial infarction (AUC = 0.90 for Model 1 vs. AUC = 0.84 for Model 2) and in those with well-managed LDL levels (LDL < 70 mg/dL; AUC = 0.90 vs. AUC = 0.93). In contrast, predictive accuracy was lower in subgroups with hypertension (AUC = 0.75 vs. AUC = 0.78) and elevated NT-proBNP (≥ 300 pg/mL; AUC = 0.68 vs. AUC = 0.82), revealing distinct performance variations between the two NNs across different subgroups. Conclusions This analysis highlights the effectiveness of NN models in predicting incident AF across a range of population subgroups. The results underscore the importance of subgroup-specific validation to capture variability within the population. While NN models can perform well in high-risk groups, such as those with myocardial infarction, accuracy may vary across other groups, suggesting the need for refinement. With improvements, NN-based tools hold promise for enhancing population-level AF risk assessment, supporting earlier detection, and possibly improving outcomes across diverse communities.ROC Curves of Neural Network ModelsScatter and boxplots of NN performance

BackgroundAn accurate assessment of nailfold capillaroscopy (NFC) images has great importance in the diagnosis and prognosis of systemic sclerosis (SSc). To overcome some of the inherent problems with NFC image...

Single-cell co-expression analysis reveals that transcriptional modules are shared across cell types in the brain.

Study question Is the efficacy of embryo selection models altered by the conditions employed for embryo culture? Summary answer The predictive capacity of morphology and morphokinetics-based selection models might be dependent on the incubator and culture conditions employed. What is known already Morphological assessment remains the gold-standard for embryo selection. However, since the introduction of time-lapse technology to the incubators used for embryo culture, many selection algorithms based on morphokinetic annotations have been developed, adding objectivity to the embryo selection process. The efficacy and reproducibility of these algorithms have been questioned by other authors, being potentially influenced by characteristics of the patients and differing culture conditions. Nowadays, there are several incubators with time-lapse technology in the market, some of them providing novel features such as culture in high humidity conditions, which has been related to differences in the morphokinetics of embryo development. Study design, size, duration Retrospective external validation of an embryo selection algorithm based on morphokinetic annotations, in a set of 555 transferred blastocysts cultured in a time-lapse system in dry (DC, n = 281) or humid conditions (HC, n = 274), and comparison with selection by morphological criteria. A novel selection model was developed considering morphokinetic annotations of our embryo dataset, including blastocysts cultured in DC and HC. Embryos belong to autologous and oocyte-donation ICSI cycles performed in a clinic over 3 years. Participants/materials, setting, methods Embryos were cultured in a Geri incubator (Genea Biomedx) and automatically annotated (Connect&Assess2.0). The efficacy of the algorithm published by Motato et al, 2016, and ASEBIR morphological grading was assessed by Generalized Estimating Equations (GEE), considering possible confounders. Efficacy was quantified by the Area Under the ROC Curve (AUC), its 95% confidence interval (CI) and statistical significance was assessed by the Mann-Whitney test. A novel algorithm was developed aided by the visual tool FertAI (Merck). Main results and the role of chance Transferred blastocysts with known implantation data were classified A-D in base of the algorithm published by Motato et al, 2016, according to the optimal range of 2 morphokinetic parameters, tEB and s3, empirically-obtained in embryos cultured in an Embryoscope incubator. The algorithm had an AUC = 0.591, 95%CI(0.542−0.64), resulting significantly predictive of implantation (P<0.001), but lower than the efficacy reported by the authors (AUC = 0.602, 95%CI(0.559−0.645)). The efficacy was different in the two culture conditions: AUC(DC) = 0.608, 95%CI(0.54−0.676), P = 0.002; AUC(HC) = 0.588, 95%CI(0.518−0.657), P = 0.103. The morphological evaluation (3 categories: A (best) to C (worse)) resulted statistically predictive of implantation: AUC of 0.596, 95%CI(0.547−0.646), P<0.001. Again, its efficacy was different in DC (AUC = 0.626, 95%CI(0.559−0.693), P<0.001) and HC (AUC = 0.589, 95%CI(0.52−0.657), P = 0.013). The lower efficacy shown by these algorithms might be associated with different morphokinetic development of embryos cultured in a different incubator and different culture conditions. Hence, a novel scoring model was developed with empirically-determinated optimal ranges of three morphokinetic parameters, considering, for the first time, embryos cultured in DC and HC: tEB<113.874; (t5-t3)/(t5-t2) = [0.521, 0.554] and cc2 = [10.34, 11.58], yielding a score from 0 to 3. The selection model resulted in an AUC = 0.637, 95%CI(0.59−0.684), and was equally efficient in DC and HC: AUC(DC) = 0.645, 95%CI(0.579−0.712); AUC(HC) = 0.645, 95%CI(0.579−0.711); P<0.001). Limitations, reasons for caution This is a primary approach to a development of a selection algorithm using morphokinetic data of embryos cultured until blastocyst stage in a Geri incubator. The efficacy and reproducibility of the model must be validated in a different dataset. Wider implications of the findings The lower efficacy shown by a selection algorithm developed in a different incubator supports the necessity of adjusting selection tools for the specific culture conditions employed by each IVF laboratory. This is the first scoring model developed for selection of blastocysts using morphokinetic parameters recorded in a Geri time-lapse incubator. Trial registration number Not applicable

Introduction: Annually, over 500,000 Americans are hospitalized due to heart failure (HF), marking it as a major contributor to morbidity and mortality. It also poses a significant financial burden and leads to considerable losses in productivity. Objective: This study investigates the predictive accuracy of different socioeconomic metrics on the risk and outcomes of HF in Maryland. Methodology: A retrospective analysis of the Maryland State Inpatient Database (2016-2020) was conducted to assess the predictive accuracy of race/ethnicity, insurance status, household median income, and neighborhood poverty level (measured by the Distressed Communities Index) on the risk of heart failure-related hospital admissions and outcomes. Multivariate logistic regression models were also used to adjust for confounders. Result: During the study period, a total of 389,220 cases of HF were reported in the Maryland SID. The majority of these patients were white (56.8%) and female (51.1%), with a median age of 73 years (interquartile range [IQR] 62-82 years). The in-hospital mortality rate was 5.1%, while rates of atrial fibrillation, cardiac arrest and prolonged hospital stay were 34.4%, 0.3%, and 48.4%, respectively. Multivariate analysis revealed a substantial area under the ROC curve (AUC) indicating good model performance: 0.88 for predicting HF, 0.64 for atrial fibrillation, 0.64 for cardiac arrest 0.57 for prolonged hospital stays, 0.63 for mortality. Subgroup analyses showed variable predictiveness by race (AUC = 0.4378), payment method (AUC = 0.5754), income quartile (AUC = 0.5202), and deprivation composite score (AUC = 0.4751). Patients with private insurance had the highest risk of stress cardiomyopathy (odds ratio [OR] = 1.98; 95% confidence interval [CI] 1.70-2.29). Socioeconomic metrics, including neighborhood distress, showed varying predictive accuracy for the HF-related admissions and selected short-term outcomes, with the highest predictive accuracy for neighborhood distress on the risk of HF (AUC = 0.50, std: 0.006), atrial fibrillation (AUC = 0.48, std: 0.0007), cardiac arrest (AUC = 0.51, std: 0.007), and prolonged hospital stays (AUC = 0.53, std: 0.0005) and mortality (AUC = 0.50, std: 0.0015). Conclusion: Neighborhood poverty level demonstrates significant predictive power for assessing the risk of HF-related hospital admissions and the short-term outcomes among Maryland residents, exceeding factors like insurance and race/ethnicity.

PurposePost-hepatectomy liver failure (PHLF) is a severe complication in hepatocellular carcinoma (HCC) patients who have undergone hepatectomy. This research aimed to investigate the combination of albumin–bilirubin (ALBI) score and aspartate aminotransferase-platelet ratio index (APRI) as a novel approach in predicting PHLF risk in hepatitis B virus (HBV)-related HCC patients.Patients and methodsHBV-related HCC patients who underwent hepatectomy from January 2006 to October 2013 were enrolled in this study. A novel model was constructed using a combination of ALBI and APRI scores to predict PHLF risk, and the prognostic value of the model was evaluated and compared with Child-Pugh (C-P) grade, ALBI score and APRI score.ResultsA total of 1,055 HCC patients were retrospectively studied, which included 151 experienced PHLF. Univariable and multivariate analyses showed that the ALBI and APRI scores were independent predictors of PHLF. The area under the ROC curve (AUC) of the ALBI score, APRI score, and C-P grade was 0.717, 0.720, and 0.602, respectively, with AUC (ALBI) > AUC (C-P) (P <0.001) and AUC (APRI) > AUC (C-P) (P <0.001). After ALBI was associated with APRI, the AUC (ALBI-APRI) was 0.766, and AUC (ALBI-APRI) > AUC (ALBI) (P <0.001), AUC (ALBI-APRI) > AUC (APRI) (P =0.047). Our results indicated that ALBI and APRI scores had higher discriminatory abilities than C-P grade in predicting the risk of PHLF, and the ALBI-APRI model could enhance the capability of predicting PHLF compared to ALBI or APRI alone.ConclusionALBI-APRI score is a novel and effective predictive model of PHLF for HBV-related HCC patients, and its accuracy in predicting the risk of PHLF is better than that of C-P, ALBI and APRI scores.

Body fat assessment is needed in individuals with HIV. The objective was to identify the discriminatory capacity of the abdominal skinfold (ASF) tricipital skinfold (TSF), subscapular fold (SSF), calf skinfold (CSF), body adiposity index (BAI), body mass index, conicity index (IC), mid-upper arm circumference (MUAC), waist circumference (WC), perimeter of neck (PN) and waist-to-height ratio (WHtR) for high body fat in children and adolescents with HIV, compared Dual energy x-ray Absorciometry (DXA) and air displacement plethysmography (ADP). Descriptive study, cross - sectional study, with 65 children and adolescents with HIV by vertical transmission. Body fat was measured by DXA and ADP. Measures were measured by international standardization. The diagnostic properties for high body fat were assessed by area under the ROC curve (AUC). For boys, having DXA as a reference for fat, ASF (AUC-0.920), TSF (AUC-0.792), SSF (AUC 0.766), CSF (AUC-0.866), BAI satisfactory discriminatory capacity. With ADP as the reference method, ASF (AUC-0.920), TSF (AUC-0.921), SSF (AUC-0.766), CSF (AUC-0.901), BAI (AUC- 0.756) and BMI (AUC- 0.699) presented satisfactory results. For girls, having DXA as a reference for fat, ASF (AUC- 0.838), TSF (AUC-0.842), SSF (AUC-0.840), CSF (AUC-0.887), BAI (AUC-0.846), and BMI (AUC-0.859) presented satisfactory discriminatory capacity. Assuming ADP as a reference for fat, ASF (AUC- (AUC-0.799), TSF (AUC-0.825), SSF (AUC-0.767), CSF (AUC-0.897), BAI 0.788), were satisfactory. The ASF, TSF, SSF, CSF, BAI and BMI anthropometric indicators may be suggested as the most suitable for the detection of high body fat in children and adolescents with HIV.

To explore the basic characteristics of various types of intensive care unit (ICU) patients and the predictive value of six common disease severity scores in critically ill patients on the first day on the 28-day death risk. The general information, disease severity scores [acute physiology score III (APS III), Oxford acute disease severity (OASIS) score, Logistic organ dysfunction score (LODS), simplified acute physiology score II (SAPS II), systemic inflammatory response syndrome (SIRS) score and sequential organ failure assessment (SOFA) score], prognosis and other indicators of critically ill patients admitted from 2008 to 2019 were extracted from Medical Information Mart for Intensive Care-IV 2.0 (MIMIC-IV 2.0). The receiver operator characteristic curve (ROC curve) of six critical illness scores for 28-day death risk of patients in various ICU, and the area under the ROC curve (AUC) was calculated, the optimal Youden index was used to determine the cut-off value, and the AUC of various ICU was verified by Delong method. A total of 53 150 critically ill patients were enrolled, with medical ICU (MICU) accounted for the most (19.25%, n = 10 233), followed by cardiac vascular ICU (CVICU) with 17.78%(n = 9 450), and neurological ICU (NICU) accounted for the least (6.25%, n = 3 320). The patients in coronary care unit (CCU) were the oldest [years old: 71.79 (60.27, 82.33)]. The length of ICU stay in NICU was the longest [days: 2.84 (1.51, 5.49)] and accounted for the highest proportion of total length of hospital stay [63.51% (34.61%, 97.07%)]. The patients in comprehensive ICU had the shortest length of ICU stay [days: 1.75 (0.99, 3.05)]. The patients in CVICU had the lowest proportion of length of ICU stay to total length of hospital stay [27.69% (18.68%, 45.18%)]. The six scores within the first day of ICU admission in NICU patients were lower than those in the other ICU, while APS III, LODS, OASIS, and SOFA scores in MICU patients were higher than those in the other ICU. SAP II and SIRS scores were both the highest in CVICU, respectively. In terms of prognosis, MICU patients had the highest 28-day mortality (14.14%, 1 447/10 233), while CVICU patients had the lowest (2.88%, 272/9 450). ROC curve analysis of the predictive value of each score on the 28-day death risk of various ICU patients showed that, the predictive value of APS III, LODS, and SAPS II in comprehensive ICU were higher [AUC and 95% confidence interval (95%CI) were 0.84 (0.83-0.85), 0.82 (0.81-0.84), and 0.83 (0.82-0.84), respectively]. The predictive value of OASIS, LODS, and SAPS II in surgical ICU (SICU) were higher [AUC and 95%CI were 0.80 (0.79-0.82), 0.79 (0.78-0.81), and 0.79 (0.77-0.80), respectively]. The predictive value of APS III and SAPS II in MICU were higher [AUC and 95%CI were 0.84 (0.82-0.85) and 0.82 (0.81-0.83), respectively]. The predictive value of APS III and SAPS II in CCU were higher [AUC and 95%CI were 0.86 (0.85-0.88) and 0.85 (0.83-0.86), respectively]. The predictive value of LODS and SAPS II in trauma ICU (TICU) were higher [AUC and 95%CI were 0.83 (0.82-0.83) and 0.83 (0.82-0.84), respectively]. The predictive value of OASIS and SAPS II in NICU were higher [AUC and 95%CI were 0.83 (0.80-0.85) and 0.81 (0.78-0.83), respectively]. The predictive value of APS III, LODS, and SAPS II in CVICU were higher [AUC and 95%CI were 0.84 (0.83-0.85), 0.81 (0.80-0.82), and 0.78 (0.77-0.78), respectively]. For the patients in comprehensive ICU, MICU, CCU, and CVICU, APS III or SAPS II can be applied for predicting 28-day death risk. For the patients in SICU and NICU, OASIS or SAPS II can be applied to predict 28-day death risk. For the patients in TICU, SAPS II or LODS can be applied for predicting 28-day death risk. For CVICU patients, APS III or LODS can be applied to predict 28-day death risk.

A modified area under the ROC curve and its application to marker selection and classification

Research Type: Level 3 - Retrospective cohort study, Case-control study, Meta-analysis of Level 3 studies Introduction/Purpose: The risk profiles associated with complications following total ankle arthroplasty (TAA) are not fully understood. Previous efforts to assess complication risk have identified statistically significant risk factors, but small effect sizes limit these models’ clinical utility. Given its ability to process extensive and complex data, machine learning may be a clinically relevant predictive tool. We sought to evaluate the accuracy and effectiveness of four different models for predicting short term complications, extended length of stay, and mechanical failures. Methods: The National Readmissions Database (NRD) was queried for adult patients (≥ 18) who underwent TAA from 2015-2020. Primary outcomes were complications within 180 days, extended length of stay (LOS), and mechanical failure of hardware. For each outcome, four models were created (weighted logistic regression (LR), random forest classifier (RF), gradient boosting classifier (GBC), and an artificial neural network (ANN)) using Python v3.9. Model performance was assessed using accuracy and the area under the ROC curve (AUC). AUC was categorized into poor (AUC &lt;.70), acceptable (0.70&lt; AUC &lt; 0.80), excellent (0.80&lt; AUC &lt; 0.90), and outstanding (0.90&lt; AUC &lt; 1.0), regarding the predictive capability of each model. Results: A total of 8,362 patients underwent TAA from 2015-2020. For predicting short term complications, random forest classification was marginally superior (RF- AUC: 0.58), though all models offered poor predictive capability. For predicting extended length of stay, weighted logistic regression proved most effective (LR-AUC: 0.66). All four models were relatively ineffective at predicting mechanical complications (AUC: 0.51-0.52). Conclusion: While the models created in this study offer relatively poor predictive capability, machine learning has the potential to accurately predict rare outcomes with heterogenous data; however, increasingly complex data requires much larger datasets. As national databases continue to expand, machine learning will become more accurate. In the meantime, simpler modeling techniques provide more accurate and interpretable predictions.

BackgroundDecreased tongue strength, pressure, and endurance are key indicators in determining oropharyngeal dysphagia (OD). This study aimed to examine the accuracy of the Iowa Oral Performance Instrument (IOPI) in assessing tongue strength, endurance, and pressure, and to identify predictors of OD.MethodsIn this study, we analyzed data of community-dwelling older adults (age ≥ 65 years) collected between March to December 2022. The accuracy for IOPI was examined with Receiver operating characteristic curve using area under the ROC curve (AUC), sensitivity (Se) and specificity (Sp) and optimal cutoff with Youden index (J). Bivariate and multivariate logistic regression analysis for predictors of OD were performed presenting odds ratio (OR) with 95% confidence interval (CI).ResultsThe cohort consisted of 85 older adults with mean age of 83.25 years (SD 6.76), of which 64 (75.3%) were female. The prevalence of OD using EAT-10 was 8.3%. Tongue strength demonstrated better diagnostic accuracy using anterior tongue strength (ATS): cut-off: 37.5 kPa (AUC: 0.79, Se: 0.86, and Sp: 0.65) and posterior tongue strength: cut-off: 31.5 kPa (AUC: 0.73, Se: 0.71, and Sp: 0.79). Tongue endurance demonstrated better diagnostic accuracy using anterior endurance target second (ATE-Target Sec): cut-off: 2.4 (AUC: 0.96, Se: 0.86, and Sp: 0.90), PTE-Target Sec: cut-off: 1.7 (AUC: 0.93, Se: 0.86, and Sp: 0.83), ATE-Target Max with cut-off: 34.4 kPa (AUC 0.81, Se = 0.86, and Sp = 0.64), and PTE-Target Max with cut-off: 29.5 kPa (AUC: 0.77, Se = 0.86, and Sp = 0.69). Tongue pressure revealed limited diagnostic accuracy using saliva swallowing pressure with cut-off: 23.3 kPa (AUC: 0.60) and effortful swallowing pressure with cut-off: 28.5 kPa (AUC: 0.62). Significant predictors for OD were frailty (3.02, 95%CI: 1.56–5.88), age (1.17, 95%CI: 1.01–1.35), nutritional status (0.72, 95%CI: 0.57–0.92), ATS (0.86, 95%CI: 0.77–0.97), ATE-Target Max (0.90, 95%CI: 0.84–0.97), PTE-Target Max (0.92, 95%CI: 0.86–0.99), ADL (0.91), IADL (0.67), and depression (1.32).ConclusionsThe findings suggest that tongue strength and endurance, measured by IOPI, are more effective parameters than tongue pressure, with frailty, age, nutritional status, ATS, ATE-Target Max, PTE-Target Max, ADL, IADL, and depression being essential for early screening of OD in community-dwelling older adults.Clinical trial numberNot applicable.

Introduction: HDL cholesterol (HDL-C) may not always represent the HDL function. This study was conducted to help predict the risk of coronary artery disease (CAD) in patients with type 2 diabetes (T2D) based on substituting paraoxonase 1 (PON1) as an important enzyme in HDL function for that of HDL-C in the atherogenic ratios LDL-C/HDL-C and log (TG/HDL-C). Methods: A total of 274 subjects undergoing diagnostic coronary angiography in this study; 92 without significant CAD and 182 with severe CAD. The hydrolysis rate of phenylacetate as a PON1 substrate was spectrophotometrically measured in kinetic mode. The area under the ROC curve (AUC) was calculated to quantify the overall discriminative capacity of the variables. Results: ROC analyses showed that LDL-C/PON1 (AUC=0.66, p=0.02) and log (TG/PON1) (AUC=0.65, p=0.019) had higher predictive powers compared with LDL-C/HDL-C (AUC=0.55, p=0.221) and log (TG/HDL-C) (AUC=0.57, p=0.121) for CAD characterization in T2D patients. Also, AUCs of LDL-C/PON1 (AUC=0.74, p=0.014) and log (TG/PON1) (AUC=0.67, p=0.057) were higher than those of LDL-C/HDL-C (AUC=0.51, p=0.469) and log (TG/HDL-C) (AUC=0.53, p=0.364) for predicting CAD in T2D patients with HbA1c ≥7%. The multiple regression analysis with adjustments for gender, age, BMI, duration of T2D, metformin and statin therapy indicated that LDL-C/PON1 [adjusted OR=1.1 (0.99,1.23), p=0.085] and log (TG/PON1) [adjusted OR=3.76 (0.94,14.89), p=0.058] were independently predictors of CAD in T2D patients. Conclusions: New cardiometabolic biomarkers using PON1 provide an incremental predictive value compared with those using HDL-C in identifying subjects at increased risk for CAD among patients with T2D and poor glycemic control. This study sought to exploit the lipoprotein-related biomarkers of CAD from a more effective perspective in T2D. Our findings may help develop clinically cost-effective and functional laboratory measurements to evaluate HDL function.

The area under the ROC curve (AUC) is a widely used measure for evaluating classification performance on heavily imbalanced data. The kernelized AUC maximization machines have established a superior generalization ability compared to linear AUC machines because of their capability in modeling the complex nonlinear structures underlying most real-world data. However, the high training complexity renders the kernelized AUC machines infeasible for large-scale data. In this paper, we present two nonlinear AUC maximization algorithms that optimize linear classifiers over a finite-dimensional feature space constructed via the k-means Nystrom approximation. Our first algorithm maximizes the AUC metric by optimizing a pairwise squared hinge loss function using the truncated Newton method. However, the second-order batch AUC maximization method becomes expensive to optimize for extremely massive datasets. This motivates us to develop a first-order stochastic AUC maximization algorithm that incorporates a scheduled regularization update and scheduled averaging to accelerate the convergence of the classifier. Experiments on several benchmark datasets demonstrate that the proposed AUC classifiers are more efficient than kernelized AUC machines while they are able to surpass or at least match the AUC performance of the kernelized AUC machines. We also show experimentally that the proposed stochastic AUC classifier is able to reach the optimal solution, while the other state-of-the-art online and stochastic AUC maximization methods are prone to suboptimal convergence. Code related to this paper is available at: https://sites.google.com/view/majdikhalid/.