A Study on Survival Analysis Methods Using Neural Network to Prevent Cancers.

Background: Cancer is one of the main global health threats. Early personalized prediction of cancer incidence is crucial for the population at risk. This study introduces a novel cancer prediction model based on modern recurrent survival deep learning algorithms. Methods: The study includes 160,407 participants from the blood-based cohort of the Korea Cancer Prevention Research-II Biobank, which has been ongoing since 2004. Data linkages were designed to ensure anonymity, and data collection was carried out through nationwide medical examinations. Predictive performance on ten cancer sites, evaluated using the concordance index (c-index), was compared among nDeep and its multitask variation, Cox proportional hazard (PH) regression, DeepSurv, and DeepHit. Results: Our models consistently achieved a c-index of over 0.8 for all ten cancers, with a peak of 0.8922 for lung cancer. They outperformed Cox PH regression and other survival deep neural networks. Conclusion: This study presents a survival deep learning model that demonstrates the highest predictive performance on censored health dataset, to the best of our knowledge. In the future, we plan to investigate the causal relationship between explanatory variables and cancer to reduce cancer incidence and mortality.

10-year follow-up study on medical expenses and medical care use according to biological age: National Health Insurance Service Health Screening Cohort (NHIS-HealS 2002~2019).

OBJECTIVES: The world is witnessing a sharp increase in its elderly population, accelerated by longer life expectancy and lower birth rates, which in turn imposes enormous medical burden on society. Although numerous studies have predicted medical expenses based on region, gender, and chronological age (CA), any attempt has rarely been made to utilize biological age (BA)-an indicator of health and aging-to ascertain and predict factors related to medical expenses and medical care use. Thus, this study employs BA to predict factors that affect medical expenses and medical care use. MATERIALS AND METHODS: Referring to the health screening cohort database of the National Health Insurance Service (NHIS), this study targeted 276,723 adults who underwent health check-ups in 2009-2010 and kept track of the data on their medical expenses and medical care use up to 2019. The average follow-up period is 9.12 years. Twelve clinical indicators were used to measure BA, while the total annual medical expenses, total annual number of outpatient days, total annual number of days in hospital, and average annual increases in medical expenses were used as the variables for medical expenses and medical care use. For statistical analysis, this study employed Pearson correlation analysis and multiple regression analysis. RESULTS: Regression analysis of the differences between corrected biological age (cBA) and CA exhibited statistically significant increases (p<0.05) in all the variables of the total annual medical expenses, total annual number of outpatient days, total annual number of days in hospital, and average annual increases in medical expenses. CONCLUSIONS: This study quantified decreases in the variables for medical expenses and medical care use based on improved BA, thereby motivating people to become more health-conscious. In particular, this study is significant in that it is the first of its kind to predict medical expenses and medical care use through BA.

Effects of Customized Digital Health Care Service on Metabolic Syndrome Status and Lifestyle Using a Health Care App: Clinical Trial.

BACKGROUND: Untact cultures have rapidly spread around the world as a result of the prolongation of the COVID-19 pandemic, leading to various types of research and technological developments in the fields of medicine and health care, where digital health care refers to health care services provided in a digital environment. Previous studies relating to digital health care demonstrated its effectiveness in managing chronic diseases such as hypertension and diabetes. While many studies have applied digital health care to various diseases, daily health care is needed for healthy individuals before they are diagnosed with a disease. Accordingly, research on individuals who have not been diagnosed with a disease is also necessary. OBJECTIVE: This study aimed to identify the effects of using a customized digital health care service (CDHCS) on risk factors for metabolic syndrome (MS) and lifestyle improvement. METHODS: The population consisted of 63 adults who underwent a health checkup at the National Health Insurance Service Ilsan (NHIS) Hospital in 2020. Measured variables include basic clinical indicators, MS-related variables, and lifestyle variables. All items were measured at NHIS Ilsan Hospital before the use of the CDHCS and 3 months thereafter. The CDHCS used in this study is a mobile app that analyzes the health condition of the user by identifying their risk factors and provides appropriate health care content. For comparison between before and after CDHCS use (pre-post comparison), paired t test was used for continuous variables, and a chi-square test was used for nominal variables. RESULTS: The study population included 30 (47.6%) male and 33 (52.4%) female participants, and the mean age was 47.61 (SD 13.93) years. The changes in clinical indicators before and after intervention results showed a decrease in weight, waist circumference, triglyceride, and high-density lipoprotein cholesterol and increases in systolic blood pressure and diastolic blood pressure. The distribution of the risk group increased from 32 (50.8%) to 34 (54%) and that of the MS group decreased from 18 (28.6%) to 16 (25.4%). The mean metabolic syndrome age-chronological age before the CDHCS was 2.20 years, which decreased to 1.72 years after CDHCS, showing a decrease of 0.48 years in the mean metabolic syndrome age-chronological age after the intervention. While all lifestyle variables, except alcohol consumption, showed a tendency toward improvement, the differences were not statistically significant. CONCLUSIONS: Although there was no statistical significance in the variables under study, this pilot study will provide a foundation for more accurate verification of CDHCS in future research.

Predicting the incidence of age-related diseases based on biological age: The 11-year national health examination data follow-up.

PURPOSE: As the population ages rapidly, the incidence of age-related diseases (ARDs) is also increasing fast. Predicting the incidence of ARDs is a challenge since the rates of individual aging vary, and objective assessments of the stages of aging based on chronological age (CA) may be inaccurate. Thus, in this study, we developed a biological age (BA) model based on the National Health Examination (NHE) data and analyzed the model prediction results for the incidence of 16 ARDs. METHODS: This study was based on the 2002-2019 National Health Information Databases of the National Health Insurance Service (NHIS-NHID). The data from a total of 10,002,494 subjects were selected between 2009 and 2010, and the principal component analysis (PCA) was performed to develop the BA model. The Cox-proportional hazard model was used to perform predictive analysis of the ARD incidence. RESULTS: For the unit increase in the difference between corrected biological age (cBA) and chronological age (CA), the hazard ratios (HRs) of ARDs increased significantly for both sexes (p < 0.001). In descending order, the corresponding ARDs' HRs were obesity (1.655), chronic renal failure (1.362), hypertension (1.301), hyperlipidemia (1.264), diabetes mellitus (1.261), fracture (1.119), dementia (1.163), cataract (1.116), myocardial infarction (1.097), stroke (1.169), macular degeneration (1.075), osteoarthritis (1.059), osteoporosis (1.124), Parkinson's disease (1.048), and chronic obstructive pulmonary disease (1.026). CONCLUSIONS: In this study, the incidence of 16 ARDs were analyzed based on BA. Therefore, conducting the NHIS health examination can facilitate the prevention of ARDs by estimating HRs for at least 16 diseases.

Biological age and lifestyle in the diagnosis of metabolic syndrome: the NHIS health screening data, 2014-2015.

Metabolic syndrome (MS) is diagnosed using absolute criteria that do not consider age and sex, but most studies have shown that the prevalence of MS increases with age in both sexes. Thus, the evaluation of MS should consider sex and age. We aimed to develop a new index that considers the age and sex for evaluating an individual's relative overall MS status. Data of 16,518,532 subjects (8,671,838 males and 7,846,694 females) who completed a validated health survey of the National Health Insurance Service of the Republic of Korea (2014‒2015) were analyzed to develop an MS-biological age model. Principal component score analysis using waist circumference, pulse pressure, fasting blood sugar, triglyceride levels, and high-density lipoprotein level, but not age, as independent variables were performed to derive an index of health status and biological age. In both sexes, the age according to the MS-biological age model increased with rising smoking and alcohol consumption habits and decreased with rising physical activity. Particularly, smoking and drinking affected females, whereas physical activity affected males. The MS-biological age model can be a supplementary tool for evaluating and managing MS, quantitatively measuring the effect of lifestyle changes on MS, and motivating patients to maintain a healthy lifestyle.

Biological age as a health index for mortality and major age-related disease incidence in Koreans: National Health Insurance Service - Health screening 11-year follow-up study.

PURPOSE: A comprehensive health index is needed to measure an individual's overall health and aging status and predict the risk of death and age-related disease incidence, and evaluate the effect of a health management program. The purpose of this study is to demonstrate the validity of estimated biological age (BA) in relation to all-cause mortality and age-related disease incidence based on National Sample Cohort database. PATIENTS AND METHODS: This study was based on National Sample Cohort database of the National Health Insurance Service - Eligibility database and the National Health Insurance Service - Medical and Health Examination database of the year 2002 through 2013. BA model was developed based on the National Health Insurance Service - National Sample Cohort (NHIS - NSC) database and Cox proportional hazard analysis was done for mortality and major age-related disease incidence. RESULTS: For every 1 year increase of the calculated BA and chronological age difference, the hazard ratio for mortality significantly increased by 1.6% (1.5% in men and 2.0% in women) and also for hypertension, diabetes mellitus, heart disease, stroke, and cancer incidence by 2.5%, 4.2%, 1.3%, 1.6%, and 0.4%, respectively (p<0.001). CONCLUSION: Estimated BA by the developed BA model based on NHIS - NSC database is expected to be used not only as an index for assessing health and aging status and predicting mortality and major age-related disease incidence, but can also be applied to various health care fields.

Models for estimating the metabolic syndrome biological age as the new index for evaluation and management of metabolic syndrome.

PURPOSE: This study aims to propose a metabolic syndrome (MS) biological age model, through which overall evaluation and management of the health status and aging state in MS can be done easily. Through this model, we hope to provide a novel evaluation and management health index that can be utilized in various health care fields. PATIENT AND METHODS: MS parameters from American Heart Association/National Heart, Lung, and Blood Institute guidelines in 2005 were used as biomarkers for the estimation of MS biological age. MS biological age model development was done by analyzing data of 263,828 participants and clinical application of the developed MS biological age was assessed by analyzing the data of 188,886 subjects. RESULTS: The principal component accounted for 36.1% in male and 38.9% in female of the total variance in the battery of five variables. The correlation coefficient between corrected biological age and chronological age in males and females were 0.711 and 0.737, respectively. Significant difference for mean MS biological age and chronological age between the three groups, normal, at risk and MS, was seen (P<0.001). CONCLUSION: For the comprehensive approach in MS management, MS biological age is expected to be additionally utilized as a novel evaluation and management index along with the traditional MS diagnosis.

Models for estimating the biological age of five organs using clinical biomarkers that are commonly measured in clinical practice settings.

OBJECTIVES: To date, no worldwide studies have been conducted to estimate the biological age of five organs using clinical biomarkers that are associated with the aging status. Therefore, we conducted this study to develop the models for estimating the biological age of five organs (heart, lung, liver, pancreas, and kidney) using clinical biomarkers which are commonly measured in clinical practice. DESIGN: A cross sectional study. METHODS: Subjects were recruited from the routine health check-up centers in Korea from 2004 through 2010. Data obtained from 121,189 subjects (66,168 men and 55,021 women) were used for clinical evaluation and statistical analysis. We examined the relations between clinical biomarkers associated with five organs and the chronological age and proposed a model for estimating the biological age of five organs. RESULTS: In the models for predicting the biological ages of the heart, lung, liver, pancreas and kidney in men, 12, 2, 8, 3, and 5 parameters were respectively included (R(2)=0.652, 0.427, 0.107, 0.245, and 0.651). In contrast to men, 10, 2, 8, 3, and 5 parameters in women were respectively included (R(2)=0.780, 0.435, 0.140, 0.384, and 0.501). CONCLUSION: We first proposed the models for predicting the biological age of five organs in the current study. We developed those using clinical parameters that can be easily obtained in clinical practice settings. Our biological age prediction models may be used as supplementary tools to assess the aging status of five organs in clinical practice settings.

A model for estimating body shape biological age based on clinical parameters associated with body composition.

BACKGROUND: To date, no studies have attempted to estimate body shape biological age using clinical parameters associated with body composition for the purposes of examining a person's body shape based on their age. OBJECTIVE: We examined the relations between clinical parameters associated with body composition and chronological age, and proposed a model for estimating the body shape biological age. METHODS: The study was conducted in 243,778 subjects aged between 20 and 90 years who received a general medical checkup at health promotion centers at university and community hospitals in Korea from 2004 to 2011. RESULTS: In men, the clinical parameters with the highest correlation to age included the waist- to-hip ratio (r = 0.786, P < 0.001), hip circumference (r = -0.448, P < 0.001), and height (r = -0.377, P < 0.001). In women, the clinical parameters with the highest correlation to age include the waist-to-hip ratio (r = 0.859, P < 0.001), waist circumference (r = 0.580, P < 0.001), and hip circumference (r = 0.520, P < 0.001). To estimate the optimal body shape biological age based on clinical parameters associated with body composition, we performed a multiple regression analysis. In a model estimating the body shape biological age, the coefficient of determination (R(2)) was 0.71 in men and 0.76 in women. CONCLUSION: Our model for estimating body shape biological age might be a novel approach to variation in body shape that is due to aging. We assume that our estimation model would be used as an adjunctive measure in easily predicting differences in body shape with the use of clinical parameters that are commonly used to assess the status of obesity in a clinical setting.

Development of models for predicting biological age (BA) with physical, biochemical, and hormonal parameters.

Individual differences are the hallmark of aging. Chronological age (CHA) is known that fails to provide an accurate indicator of the aging but biological age (BA) estimates the functional status of an individual in reference to his or her chronological peers on the basis of how well he or she functions in comparison with others of the same CHA. Therefore, we developed models for predicting BA that can be applicable in clinical practice settings. This was a community-based cross-sectional study. Subjects were recruited from the health promotion center in Korea from 2001 to 2005. Among these, data obtained from the 3575 participants (1302 men and 2273 women) was used for clinical evaluation and statistical analysis. For our test battery we selected 25 parameters among the routine tests. For males, the best models were developed using 15, 7, 5, and 4 of the 25 chosen parameters for total, physical, biochemical and hormonal characteristics, respectively (R(2)=0.62, 0.38, 0.33, and 0.36, respectively). Similar to males, for the females, 14, 6, 8, and 3 parameters were developed as the models (R(2)=0.66, 0.40, 0.42, and 0.37, respectively). Our BA prediction models may be used as supplementary tools adding knowledge in the evaluation of aging status.

Hypocholesterolemic property of Yucca schidigera and Quillaja saponaria extracts in human body.

This study was undertaken to observe the effects of the blend of partially purified Yucca schidigera and Quillaja saponaria extracts on cholesterol levels in the human's blood and gastrointestinal functions, and to determine if a new cholesterol-lowering drug can be developed by the further purification of the extracts. Ultrafiltration and sequential diafiltration increased the amounts of steroidal saponin in aqueous yucca extract and terpenoid saponin in aqueous quillaja extract from 9.3% and 21.4% to 17.2% and 61.8%, respectively. Taking 0.9 mg of the blend (6:4, v:v) of the resulting filtrates a day for 4 weeks resulted in the decreases in total and LDL cholesterol levels in blood plasma of hyper-cholesterolemic patients with enhancement in gastrointestinal symptoms of patients.