Osteoporosis risk prediction for bone mineral density assessment of postmenopausal women using machine learning

Tae Keun Yoo, Sung Kean Kim, Deok Won Kim, Joon Yul Choi, Wan Hyung Lee, Ein Oh, Euncheol Park

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

Purpose: A number of clinical decision tools for osteoporosis risk assessment have been developed to select postmenopausal women for the measurement of bone mineral density. We developed and validated machine learning models with the aim of more accurately identifying the risk of osteoporosis in postmenopausal women compared to the ability of conventional clinical decision tools. Materials and Methods: We collected medical records from Korean postmenopausal women based on the Korea National Health and Nutrition Examination Surveys. The training data set was used to construct models based on popular machine learning algorithms such as support vector machines (SVM), random forests, artificial neural networks (ANN), and logistic regression (LR) based on simple surveys. The machine learning models were compared to four conventional clinical decision tools: osteoporosis self-assessment tool (OST), osteoporosis risk assessment instrument (ORAI), simple calculated osteoporosis risk estimation (SCORE), and osteoporosis index of risk (OSIRIS). Results: SVM had significantly better area under the curve (AUC) of the receiver operating characteristic than ANN, LR, OST, ORAI, SCORE, and OSIRIS for the training set. SVM predicted osteoporosis risk with an AUC of 0.827, accuracy of 76.7%, sensitivity of 77.8%, and specificity of 76.0% at total hip, femoral neck, or lumbar spine for the testing set. The significant factors selected by SVM were age, height, weight, body mass index, duration of menopause, duration of breast feeding, estrogen therapy, hyperlipidemia, hypertension, osteoarthritis, and diabetes mellitus. Conclusion: Considering various predictors associated with low bone density, the machine learning methods may be effective tools for identifying postmenopausal women at high risk for osteoporosis.

Original languageEnglish
Pages (from-to)1321-1330
Number of pages10
JournalYonsei medical journal
Volume54
Issue number6
DOIs
Publication statusPublished - 2013 Oct 31

Fingerprint

Bone Density
Osteoporosis
Area Under Curve
Machine Learning
Logistic Models
Postmenopausal Osteoporosis
Aptitude
Nutrition Surveys
Femur Neck
Korea
Menopause
Hyperlipidemias
Breast Feeding
ROC Curve
Osteoarthritis
Medical Records
Hip
Diabetes Mellitus
Estrogens
Spine

All Science Journal Classification (ASJC) codes

  • Medicine(all)

Cite this

Yoo, Tae Keun ; Kim, Sung Kean ; Kim, Deok Won ; Choi, Joon Yul ; Lee, Wan Hyung ; Oh, Ein ; Park, Euncheol. / Osteoporosis risk prediction for bone mineral density assessment of postmenopausal women using machine learning. In: Yonsei medical journal. 2013 ; Vol. 54, No. 6. pp. 1321-1330.
@article{064b0904720642be9a73d48ad70cabbd,
title = "Osteoporosis risk prediction for bone mineral density assessment of postmenopausal women using machine learning",
abstract = "Purpose: A number of clinical decision tools for osteoporosis risk assessment have been developed to select postmenopausal women for the measurement of bone mineral density. We developed and validated machine learning models with the aim of more accurately identifying the risk of osteoporosis in postmenopausal women compared to the ability of conventional clinical decision tools. Materials and Methods: We collected medical records from Korean postmenopausal women based on the Korea National Health and Nutrition Examination Surveys. The training data set was used to construct models based on popular machine learning algorithms such as support vector machines (SVM), random forests, artificial neural networks (ANN), and logistic regression (LR) based on simple surveys. The machine learning models were compared to four conventional clinical decision tools: osteoporosis self-assessment tool (OST), osteoporosis risk assessment instrument (ORAI), simple calculated osteoporosis risk estimation (SCORE), and osteoporosis index of risk (OSIRIS). Results: SVM had significantly better area under the curve (AUC) of the receiver operating characteristic than ANN, LR, OST, ORAI, SCORE, and OSIRIS for the training set. SVM predicted osteoporosis risk with an AUC of 0.827, accuracy of 76.7{\%}, sensitivity of 77.8{\%}, and specificity of 76.0{\%} at total hip, femoral neck, or lumbar spine for the testing set. The significant factors selected by SVM were age, height, weight, body mass index, duration of menopause, duration of breast feeding, estrogen therapy, hyperlipidemia, hypertension, osteoarthritis, and diabetes mellitus. Conclusion: Considering various predictors associated with low bone density, the machine learning methods may be effective tools for identifying postmenopausal women at high risk for osteoporosis.",
author = "Yoo, {Tae Keun} and Kim, {Sung Kean} and Kim, {Deok Won} and Choi, {Joon Yul} and Lee, {Wan Hyung} and Ein Oh and Euncheol Park",
year = "2013",
month = "10",
day = "31",
doi = "10.3349/ymj.2013.54.6.1321",
language = "English",
volume = "54",
pages = "1321--1330",
journal = "Yonsei Medical Journal",
issn = "0513-5796",
publisher = "Yonsei University College of Medicine",
number = "6",

}

Osteoporosis risk prediction for bone mineral density assessment of postmenopausal women using machine learning. / Yoo, Tae Keun; Kim, Sung Kean; Kim, Deok Won; Choi, Joon Yul; Lee, Wan Hyung; Oh, Ein; Park, Euncheol.

In: Yonsei medical journal, Vol. 54, No. 6, 31.10.2013, p. 1321-1330.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Osteoporosis risk prediction for bone mineral density assessment of postmenopausal women using machine learning

AU - Yoo, Tae Keun

AU - Kim, Sung Kean

AU - Kim, Deok Won

AU - Choi, Joon Yul

AU - Lee, Wan Hyung

AU - Oh, Ein

AU - Park, Euncheol

PY - 2013/10/31

Y1 - 2013/10/31

N2 - Purpose: A number of clinical decision tools for osteoporosis risk assessment have been developed to select postmenopausal women for the measurement of bone mineral density. We developed and validated machine learning models with the aim of more accurately identifying the risk of osteoporosis in postmenopausal women compared to the ability of conventional clinical decision tools. Materials and Methods: We collected medical records from Korean postmenopausal women based on the Korea National Health and Nutrition Examination Surveys. The training data set was used to construct models based on popular machine learning algorithms such as support vector machines (SVM), random forests, artificial neural networks (ANN), and logistic regression (LR) based on simple surveys. The machine learning models were compared to four conventional clinical decision tools: osteoporosis self-assessment tool (OST), osteoporosis risk assessment instrument (ORAI), simple calculated osteoporosis risk estimation (SCORE), and osteoporosis index of risk (OSIRIS). Results: SVM had significantly better area under the curve (AUC) of the receiver operating characteristic than ANN, LR, OST, ORAI, SCORE, and OSIRIS for the training set. SVM predicted osteoporosis risk with an AUC of 0.827, accuracy of 76.7%, sensitivity of 77.8%, and specificity of 76.0% at total hip, femoral neck, or lumbar spine for the testing set. The significant factors selected by SVM were age, height, weight, body mass index, duration of menopause, duration of breast feeding, estrogen therapy, hyperlipidemia, hypertension, osteoarthritis, and diabetes mellitus. Conclusion: Considering various predictors associated with low bone density, the machine learning methods may be effective tools for identifying postmenopausal women at high risk for osteoporosis.

AB - Purpose: A number of clinical decision tools for osteoporosis risk assessment have been developed to select postmenopausal women for the measurement of bone mineral density. We developed and validated machine learning models with the aim of more accurately identifying the risk of osteoporosis in postmenopausal women compared to the ability of conventional clinical decision tools. Materials and Methods: We collected medical records from Korean postmenopausal women based on the Korea National Health and Nutrition Examination Surveys. The training data set was used to construct models based on popular machine learning algorithms such as support vector machines (SVM), random forests, artificial neural networks (ANN), and logistic regression (LR) based on simple surveys. The machine learning models were compared to four conventional clinical decision tools: osteoporosis self-assessment tool (OST), osteoporosis risk assessment instrument (ORAI), simple calculated osteoporosis risk estimation (SCORE), and osteoporosis index of risk (OSIRIS). Results: SVM had significantly better area under the curve (AUC) of the receiver operating characteristic than ANN, LR, OST, ORAI, SCORE, and OSIRIS for the training set. SVM predicted osteoporosis risk with an AUC of 0.827, accuracy of 76.7%, sensitivity of 77.8%, and specificity of 76.0% at total hip, femoral neck, or lumbar spine for the testing set. The significant factors selected by SVM were age, height, weight, body mass index, duration of menopause, duration of breast feeding, estrogen therapy, hyperlipidemia, hypertension, osteoarthritis, and diabetes mellitus. Conclusion: Considering various predictors associated with low bone density, the machine learning methods may be effective tools for identifying postmenopausal women at high risk for osteoporosis.

UR - http://www.scopus.com/inward/record.url?scp=84886503138&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84886503138&partnerID=8YFLogxK

U2 - 10.3349/ymj.2013.54.6.1321

DO - 10.3349/ymj.2013.54.6.1321

M3 - Article

VL - 54

SP - 1321

EP - 1330

JO - Yonsei Medical Journal

JF - Yonsei Medical Journal

SN - 0513-5796

IS - 6

ER -