Comparison of Biological Age Prediction Models Using Clinical Biomarkers Commonly Measured in Clinical Practice Settings: AI Techniques Vs. Traditional Statistical Methods

Bae, Chul Young; Im, Yoori; Lee, Jonghoon; Park, Choong-Shik; Kim, Miyoung; Kwon, Hee Uk; Kim, Boseon; Park, Hye ri; Lee, Chun-Koo; Kim, Inhee; Kim, JeongHoon

doi:10.3389/frans.2021.709589

Cited by 12 publications

(6 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work considered machine learning methods (Multiple Linear Regression (MLR), Generalized Additive Models (GAM), Support Vector Machine (SVM), Adaboost, Gradient Boosting Decision Tree (GBDT), Light Gradient Boosting Machine (LGBM), Catboost, Xgboost, Extra Trees) and neural network methods (Deep Neural Networks (DNN), Convolutional Neural Network (CNN)) that can be used for regression analysis. The Pearson correlations, MAE, and RMSE between BA and CA are the indicators used to compare different BA estimation algorithms, which are done in the test set [ 25 , 26 ].…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

et al. 2022

View full text Add to dashboard Cite

Background Biological age (BA) has been recognized as a more accurate indicator of aging than chronological age (CA). However, the current limitations include: insufficient attention to the incompleteness of medical data for constructing BA; Lack of machine learning-based BA (ML-BA) on the Chinese population; Neglect of the influence of model overfitting degree on the stability of the association results. Methods and results Based on the medical examination data of the Chinese population (45–90 years), we first evaluated the most suitable missing interpolation method, then constructed 14 ML-BAs based on biomarkers, and finally explored the associations between ML-BAs and health statuses (healthy risk indicators and disease). We found that round-robin linear regression interpolation performed best, while AutoEncoder showed the highest interpolation stability. We further illustrated the potential overfitting problem in ML-BAs, which affected the stability of ML-Bas’ associations with health statuses. We then proposed a composite ML-BA based on the Stacking method with a simple meta-model (STK-BA), which overcame the overfitting problem, and associated more strongly with CA (r = 0.66, P < 0.001), healthy risk indicators, disease counts, and six types of disease. Conclusion We provided an improved aging measurement method for middle-aged and elderly groups in China, which can more stably capture aging characteristics other than CA, supporting the emerging application potential of machine learning in aging research.

show abstract

Section: Methodsmentioning

confidence: 99%

“…The Pearson correlations, MAE, and RMSE between BA and CA are the preferred and most commonly used indicators to compare different BA estimation algorithms [ 25 , 26 ]. Exploring the associations of ML-BA with epidemiological variables (e.g.…”

Section: Introductionmentioning

confidence: 99%

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Negative values of Δage (i.e., where BA is less than CA) are associated with the deceleration of aging and a lower risk of morbidity, hospitalization, and mortality [ 15 , 16 ]. One of the most innovative ways to estimate biological aging is by applying deep neural networks to circulating biomarkers [ 16 , 17 , 18 , 19 ]. Indeed, although this represents only a generic marker of biological aging and other markers or scales such as frailty and cognitive performance may better tag organ-specific aging [ 20 ] or the intrinsic aging capacity [ 21 , 22 ], blood-based estimates of BA can provide information on several aging domains within the human body because it can be based on a range of different circulating biomarkers.…”

Section: Introductionmentioning

confidence: 99%

“…Indeed, although this represents only a generic marker of biological aging and other markers or scales such as frailty and cognitive performance may better tag organ-specific aging [ 20 ] or the intrinsic aging capacity [ 21 , 22 ], blood-based estimates of BA can provide information on several aging domains within the human body because it can be based on a range of different circulating biomarkers. Indeed, previous studies identified prominent roles of glucose homeostasis, liver and kidney functionality, and inflammation, among other biomarkers [ 16 , 17 , 18 , 19 ]. Moreover, the wide availability of routine blood tests resulting from common clinical practice makes this a cost-effective estimator of biological aging, which could be used as a public health and healthy aging screening tool in the general population [ 16 ].…”

Section: Introductionmentioning

confidence: 99%

Association between the Inflammatory Potential of the Diet and Biological Aging: A Cross-Sectional Analysis of 4510 Adults from the Moli-Sani Study Cohort

et al. 2023

View full text Add to dashboard Cite

Chronological age (CA) may not accurately reflect the health status of an individual. Rather, biological age (BA) or hypothetical underlying “functional” age has been proposed as a relevant indicator of healthy aging. Observational studies have found that decelerated biological aging or Δage (BA-CA) is associated with a lower risk of disease and mortality. In general, CA is associated with low-grade inflammation, a condition linked to the risk of the incidence of disease and overall cause-specific mortality, and is modulated by diet. To address the hypothesis that diet-related inflammation is associated with Δage, a cross-sectional analysis of data from a sub-cohort from the Moli-sani Study (2005–2010, Italy) was performed. The inflammatory potential of the diet was measured using the Energy-adjusted Dietary Inflammatory Index (E-DIITM) and a novel literature-based dietary inflammation score (DIS). A deep neural network approach based on circulating biomarkers was used to compute BA, and the resulting Δage was fit as the dependent variable. In 4510 participants (men 52.0%), the mean of CA (SD) was 55.6 y (±11.6), BA 54.8 y (±8.6), and Δage −0.77 (±7.7). In a multivariable-adjusted analysis, an increase in E-DIITM and DIS scores led to an increase in Δage (β = 0.22; 95%CI 0.05, 0.38; β = 0.27; 95%CI 0.10, 0.44, respectively). We found interaction for DIS by sex and for E-DIITM by BMI. In conclusion, a pro-inflammatory diet is associated with accelerated biological aging, which likely leads to an increased long-term risk of inflammation-related diseases and mortality.

show abstract

“…The Pearson correlations, MAE, and RMSE between BA and CA are the preferred and most commonly used indicators to compare different BA estimation algorithms, which are usually done in the test set [27,28]. In addition, exploring the associations of ML-BA with epidemiological variables (e.g.…”

Section: Introductionmentioning

confidence: 99%

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

Yang

Gao

Lin

et al. 2022

Preprint

View full text Add to dashboard Cite

Background: Biological age (BA) has been recognized as a more accurate indicator of aging than chronological age (CA). However, current limitations include insufficient attention to the incompleteness of medical data used to construct BA; lack of machine learning-based BA (ML-BA) on the Chinese population; neglect of the influence of the degree of model overfitting on the stability of the correlation results.Methods and Results: Based on the medical examination data of the Chinese population (45-90 years), we first evaluated the most suitable missing interpolation method, then constructed 14 ML-BAs based on biomarkers, and finally explored the associations between ML-BAs and health statuses (healthy risk indicators and disease). We found that round-robin linear regression interpolation performed best, while AE showed the highest stability at high missing rates. We further illustrated the potential overfitting problem in ML-BAs, which affected the stability of ML-BAs' associations with health statuses. We then proposed a composite ML-BA based on the Stacking method with a simple meta-model (STK-BA), which overcame the overfitting problem, and associated more strongly with CA (r = 0.66, P < 0.001), healthy risk indicators, disease counts, and six types of disease.Conclusion: We provided an improved aging measurement method for middle-aged and elderly groups in China, which can more stably capture aging characteristics other than CA, supporting the emerging application potential of machine learning in aging research.

show abstract

Comparison of Biological Age Prediction Models Using Clinical Biomarkers Commonly Measured in Clinical Practice Settings: AI Techniques Vs. Traditional Statistical Methods

Cited by 12 publications

References 41 publications

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

Association between the Inflammatory Potential of the Diet and Biological Aging: A Cross-Sectional Analysis of 4510 Adults from the Moli-Sani Study Cohort

A machine learning-based data mining in medical examination data: a biological features-based biological age prediction model

Contact Info

Product

Resources

About