Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program
Introduction Although various lipid and non-lipid analytes measured by nuclear magnetic resonance (NMR) spectroscopy have been associated with type 2 diabetes, a structured comparison of the ability of NMR-derived biomarkers and standard lipids to predict individual diabetes risk has not been undert...
Main Authors: | , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMJ Publishing Group
2021-08-01
|
Series: | BMJ Open Diabetes Research & Care |
Online Access: | https://drc.bmj.com/content/9/1/e001953.full |
id |
doaj-6e1d050c124d4a529a5724c96d670a0f |
---|---|
record_format |
Article |
spelling |
doaj-6e1d050c124d4a529a5724c96d670a0f2021-08-10T10:31:05ZengBMJ Publishing GroupBMJ Open Diabetes Research & Care2052-48972021-08-019110.1136/bmjdrc-2020-001953Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention ProgramSamuel Dagogo-Jack0Carlos Lorenzo1Kieren J Mather2Marinella Temprosa3Tibor V Varga4Jinxi Liu5Ronald B Goldberg6Guannan Chen7Xavier Pi-Sunyer8University of Tennessee Health Science Center, Memphis, Tennessee, USAThe University of Texas Health Science Center at San Antonio, San Antonio, Texas, USAIndiana University School of Medicine, Indianapolis, Indiana, USABiostatistics Center and Department of Biostatistics and Bioinformatics, Milken Institute School of Public Health, George Washington University, Rockville, Maryland, USASection of Epidemiology, Department of Public Health, University of Copenhagen, Copenhagen, DenmarkBiostatistics Center and Department of Biostatistics and Bioinformatics, Milken Institute School of Public Health, George Washington University, Rockville, Maryland, USADepartment of Medicine, University of Miami, Miami, Florida, USABiostatistics Center and Department of Biostatistics and Bioinformatics, Milken Institute School of Public Health, George Washington University, Rockville, Maryland, USAColumbia University Medical Center, New York City, New York, USAIntroduction Although various lipid and non-lipid analytes measured by nuclear magnetic resonance (NMR) spectroscopy have been associated with type 2 diabetes, a structured comparison of the ability of NMR-derived biomarkers and standard lipids to predict individual diabetes risk has not been undertaken in larger studies nor among individuals at high risk of diabetes.Research design and methods Cumulative discriminative utilities of various groups of biomarkers including NMR lipoproteins, related non-lipid biomarkers, standard lipids, and demographic and glycemic traits were compared for short-term (3.2 years) and long-term (15 years) diabetes development in the Diabetes Prevention Program, a multiethnic, placebo-controlled, randomized controlled trial of individuals with pre-diabetes in the USA (N=2590). Logistic regression, Cox proportional hazards model and six different hyperparameter-tuned machine learning algorithms were compared. The Matthews Correlation Coefficient (MCC) was used as the primary measure of discriminative utility.Results Models with baseline NMR analytes and their changes did not improve the discriminative utility of simpler models including standard lipids or demographic and glycemic traits. Across all algorithms, models with baseline 2-hour glucose performed the best (max MCC=0.36). Sophisticated machine learning algorithms performed similarly to logistic regression in this study.Conclusions NMR lipoproteins and related non-lipid biomarkers were associated but did not augment discrimination of diabetes risk beyond traditional diabetes risk factors except for 2-hour glucose. Machine learning algorithms provided no meaningful improvement for discrimination compared with logistic regression, which suggests a lack of influential latent interactions among the analytes assessed in this study.Trial registration number Diabetes Prevention Program: NCT00004992; Diabetes Prevention Program Outcomes Study: NCT00038727.https://drc.bmj.com/content/9/1/e001953.full |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Samuel Dagogo-Jack Carlos Lorenzo Kieren J Mather Marinella Temprosa Tibor V Varga Jinxi Liu Ronald B Goldberg Guannan Chen Xavier Pi-Sunyer |
spellingShingle |
Samuel Dagogo-Jack Carlos Lorenzo Kieren J Mather Marinella Temprosa Tibor V Varga Jinxi Liu Ronald B Goldberg Guannan Chen Xavier Pi-Sunyer Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program BMJ Open Diabetes Research & Care |
author_facet |
Samuel Dagogo-Jack Carlos Lorenzo Kieren J Mather Marinella Temprosa Tibor V Varga Jinxi Liu Ronald B Goldberg Guannan Chen Xavier Pi-Sunyer |
author_sort |
Samuel Dagogo-Jack |
title |
Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program |
title_short |
Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program |
title_full |
Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program |
title_fullStr |
Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program |
title_full_unstemmed |
Predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the Diabetes Prevention Program |
title_sort |
predictive utilities of lipid traits, lipoprotein subfractions and other risk factors for incident diabetes: a machine learning approach in the diabetes prevention program |
publisher |
BMJ Publishing Group |
series |
BMJ Open Diabetes Research & Care |
issn |
2052-4897 |
publishDate |
2021-08-01 |
description |
Introduction Although various lipid and non-lipid analytes measured by nuclear magnetic resonance (NMR) spectroscopy have been associated with type 2 diabetes, a structured comparison of the ability of NMR-derived biomarkers and standard lipids to predict individual diabetes risk has not been undertaken in larger studies nor among individuals at high risk of diabetes.Research design and methods Cumulative discriminative utilities of various groups of biomarkers including NMR lipoproteins, related non-lipid biomarkers, standard lipids, and demographic and glycemic traits were compared for short-term (3.2 years) and long-term (15 years) diabetes development in the Diabetes Prevention Program, a multiethnic, placebo-controlled, randomized controlled trial of individuals with pre-diabetes in the USA (N=2590). Logistic regression, Cox proportional hazards model and six different hyperparameter-tuned machine learning algorithms were compared. The Matthews Correlation Coefficient (MCC) was used as the primary measure of discriminative utility.Results Models with baseline NMR analytes and their changes did not improve the discriminative utility of simpler models including standard lipids or demographic and glycemic traits. Across all algorithms, models with baseline 2-hour glucose performed the best (max MCC=0.36). Sophisticated machine learning algorithms performed similarly to logistic regression in this study.Conclusions NMR lipoproteins and related non-lipid biomarkers were associated but did not augment discrimination of diabetes risk beyond traditional diabetes risk factors except for 2-hour glucose. Machine learning algorithms provided no meaningful improvement for discrimination compared with logistic regression, which suggests a lack of influential latent interactions among the analytes assessed in this study.Trial registration number Diabetes Prevention Program: NCT00004992; Diabetes Prevention Program Outcomes Study: NCT00038727. |
url |
https://drc.bmj.com/content/9/1/e001953.full |
work_keys_str_mv |
AT samueldagogojack predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT carloslorenzo predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT kierenjmather predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT marinellatemprosa predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT tiborvvarga predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT jinxiliu predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT ronaldbgoldberg predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT guannanchen predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram AT xavierpisunyer predictiveutilitiesoflipidtraitslipoproteinsubfractionsandotherriskfactorsforincidentdiabetesamachinelearningapproachinthediabetespreventionprogram |
_version_ |
1721212263168212992 |