Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Age assessment of youth and young adults using magnetic resonance imaging of the knee: A deep learning approach
Blekinge Institute of Technology, Faculty of Engineering, Department of Health.ORCID iD: 0000-0002-6752-017X
Blekinge Institute of Technology, Faculty of Engineering, Department of Health.ORCID iD: 0000-0003-4312-2246
Optriva AB, SWE.
Karolinska, SWE.
Show others and affiliations
2019 (English)In: JMIR Medical Informatics, E-ISSN 2291-9694, Vol. 7, no 4, p. 419-436, article id e16291Article in journal (Refereed) Published
Abstract [en]

Background: Bone age assessment (BAA) is an important tool for diagnosis and in determining the time of treatment in a number of pediatric clinical scenarios, as well as in legal settings where it is used to estimate the chronological age of an individual where valid documents are lacking. Traditional methods for BAA suffer from drawbacks, such as exposing juveniles to radiation, intra- and interrater variability, and the time spent on the assessment. The employment of automated methods such as deep learning and the use of magnetic resonance imaging (MRI) can address these drawbacks and improve the assessment of age. Objective: The aim of this paper is to propose an automated approach for age assessment of youth and young adults in the age range when the length growth ceases and growth zones are closed (14-21 years of age) by employing deep learning using MRI of the knee. Methods: This study carried out MRI examinations of the knee of 402 volunteer subjects-221 males (55.0%) and 181 (45.0%) females-aged 14-21 years. The method comprised two convolutional neural network (CNN) models: the first one selected the most informative images of an MRI sequence, concerning age-assessment purposes; these were then used in the second module, which was responsible for the age estimation. Different CNN architectures were tested, both training from scratch and employing transfer learning. Results: The CNN architecture that provided the best results was GoogLeNet pretrained on the ImageNet database. The proposed method was able to assess the age of male subjects in the range of 14-20.5 years, with a mean absolute error (MAE) of 0.793 years, and of female subjects in the range of 14-19.5 years, with an MAE of 0.988 years. Regarding the classification of minors-with the threshold of 18 years of age-an accuracy of 98.1% for male subjects and 95.0% for female subjects was achieved. Conclusions: The proposed method was able to assess the age of youth and young adults from 14 to 20.5 years of age for male subjects and 14 to 19.5 years of age for female subjects in a fully automated manner, without the use of ionizing radiation, addressing the drawbacks of traditional methods. © 2019 Journal of Medical Internet Research. All rights reserved.

Place, publisher, year, edition, pages
JMIR PUBLICATIONS , 2019. Vol. 7, no 4, p. 419-436, article id e16291
Keywords [en]
Age assessment, Bone age, Convolutional neural networks, Deep learning, Knee, Machine learning, Magnetic resonance imaging, Medical imaging, Skeletal maturity, Transfer learning, adolescent, adult, article, bone age determination, child, convolutional neural network, diagnostic imaging, employment, female, human, human experiment, juvenile, major clinical study, male, maturity, nuclear magnetic resonance imaging, transfer of learning, young adult
National Category
Radiology, Nuclear Medicine and Medical Imaging Orthopaedics
Identifiers
URN: urn:nbn:se:bth-19085DOI: 10.2196/16291ISI: 000510198100031Scopus ID: 2-s2.0-85077013897OAI: oai:DiVA.org:bth-19085DiVA, id: diva2:1383941
Note

open access

Available from: 2020-01-09 Created: 2020-01-09 Last updated: 2022-05-25Bibliographically approved
In thesis
1. Machine learning applications in healthcare
Open this publication in new window or tab >>Machine learning applications in healthcare
2020 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Healthcare is an important and high cost sector that involves many decision-making tasks based on the analysis of data, from its primary activities up till management itself. A technology that can be useful in an environment as data-intensive as healthcare is machine learning. This thesis investigates the application of machine learning in healthcare contexts as an applied health technology (AHT). AHT refers to application of scientific methods for the development of interventions targeting practical problems related to health and healthcare.

The two research contexts in this thesis regard two pivotal activities in the healthcare systems: diagnosis and prognosis. The diagnosis research context regards the age assessment of the young individuals, which aims to address the drawbacks in the bone age assessment research, investigating new age assessment methods. The prognosis research context regards the prognosis of dementia, which aims to investigate prognostic estimates for older individuals who came to develop the dementia disorder, in a time frame of 10 years. Machine learning applications were shown to be useful in both research contexts.

In the diagnosis research context, study I summarized the state of the art evidence in the area of bone age assessment with the use of machine learning, identifying both automated and non-automated approaches for age assessment. Study II investigated a non-automated approach based on the radiologists' assessment and study III investigated an automated approach based on deep learning. Both studies used magnetic resonance imaging. The results showed that the radiologists' assessment as input was not precise enough for the estimation of age. However, the deep learning method was able to extract more useful features from the images and provided better diagnostic performance for the age assessment.

In the research context of prognosis, study IV conducted a review on the relevant evidence in on the prognosis of dementia with machine learning techniques, identifying a focus on the research on neuroimaging studies dedicated to validating biomarkers for pharmaceutical research. Study V proposed a multifactorial decision tree approach for the prognosis of dementia in older individuals as to their development or not of dementia in 10 years. Achieving consistent performance results, it provided an interpretable prognostic model identifying possible modifiable and non-modifiable risk factors and possible patient subgroups of importance for the dementia research.

Place, publisher, year, edition, pages
Karlskrona: Blekinge Tekniska Högskola, 2020. p. 202
Series
Blekinge Institute of Technology Doctoral Dissertation Series, ISSN 1653-2090 ; 6
Keywords
Machine learning, Healthcare, Diagnosis, Prognosis, Age assessment, Bone age assessment, Dementia, Applied health technology
National Category
Other Medical Sciences
Research subject
Applied Health Technology
Identifiers
urn:nbn:se:bth-19513 (URN)978-91-7295-405-2 (ISBN)
Public defence
2020-09-16, 14:00
Opponent
Supervisors
Available from: 2020-05-29 Created: 2020-05-28 Last updated: 2021-01-27Bibliographically approved

Open Access in DiVA

Age assessment of youth and young adults using magnetic resonance imaging of the knee(960 kB)278 downloads
File information
File name FULLTEXT01.pdfFile size 960 kBChecksum SHA-512
daa5d5069a3c4c4b95a03e81c92f50b4580ce645c254885e3c42ee2dda7cfb65691c81edef6b1034ad91cd7f0b600de313019e903a586e0dec7d5665efc5e41b
Type fulltextMimetype application/pdf

Other links

Publisher's full textScopus

Authority records

Dallora Moraes, Ana LuizaSanmartin Berglund, JohanAnderberg, Peter

Search in DiVA

By author/editor
Dallora Moraes, Ana LuizaSanmartin Berglund, JohanAnderberg, Peter
By organisation
Department of Health
In the same journal
JMIR Medical Informatics
Radiology, Nuclear Medicine and Medical ImagingOrthopaedics

Search outside of DiVA

GoogleGoogle Scholar
Total: 278 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

doi
urn-nbn

Altmetric score

doi
urn-nbn
Total: 305 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf