Loading…
Detecting pathological features and predicting fracture risk from dual-energy X-ray absorptiometry images using deep learning
Dual-energy X-ray absorptiometry (DXA) is the gold standard imaging method for diagnosing osteoporosis in clinical practice. The DXA images are commonly used to estimate a numerical value for bone mineral density (BMD), which decreases in osteoporosis. Low BMD is a known risk factor for osteoporotic...
Saved in:
Published in: | Bone Reports 2021-06, Vol.14, p.101070-101070, Article 101070 |
---|---|
Main Authors: | , , , , , , , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3 |
---|---|
cites | cdi_FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3 |
container_end_page | 101070 |
container_issue | |
container_start_page | 101070 |
container_title | Bone Reports |
container_volume | 14 |
creator | Nissinen, Tomi Suoranta, Sanna Saavalainen, Taavi Sund, Reijo Hurskainen, Ossi Rikkonen, Toni Kröger, Heikki Lähivaara, Timo Väänänen, Sami P. |
description | Dual-energy X-ray absorptiometry (DXA) is the gold standard imaging method for diagnosing osteoporosis in clinical practice. The DXA images are commonly used to estimate a numerical value for bone mineral density (BMD), which decreases in osteoporosis. Low BMD is a known risk factor for osteoporotic fractures. In this study, we used deep learning to identify lumbar scoliosis and structural abnormalities that potentially affect BMD but are often neglected in lumbar spine DXA analysis. In addition, we tested the approach's ability to predict fractures using only DXA images. A dataset of 2949 images gathered by Kuopio Osteoporosis Risk Factor and Prevention Study was used to train a convolutional neural network (CNN) for classification. The model was able to classify scoliosis with an AUC of 0.96 and structural abnormalities causing unreliable BMD measurement with an AUC of 0.91. It predicted fractures occurring within 5 years from the lumbar spine DXA scan with an AUC of 0.63, meeting the predictive performance of combined BMD measurements from the lumbar spine and hip. In an independent test set of 574 clinical patients, AUC for lumbar scoliosis was 0.93 and AUC for unreliable BMD measurements was 0.94. In each classification task, neural network visualizations indicated the model's predictive strategy. We conclude that deep learning could complement the well established DXA method for osteoporosis diagnostics by analyzing incidental findings and image reliability, and improve its predictive ability in the future. |
doi_str_mv | 10.1016/j.bonr.2021.101070 |
format | article |
fullrecord | <record><control><sourceid>proquest_doaj_</sourceid><recordid>TN_cdi_doaj_primary_oai_doaj_org_article_269a660dff4141c79db473c488914663</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><els_id>S2352187221003259</els_id><doaj_id>oai_doaj_org_article_269a660dff4141c79db473c488914663</doaj_id><sourcerecordid>2528432247</sourcerecordid><originalsourceid>FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3</originalsourceid><addsrcrecordid>eNp9kktv1DAQxy0EotXSL8AB-cgli1-JYwkhofKqVIkLSNwsxx6nXrJxsJNKe-C745BStRdO9rx-M5r5I_SSkj0ltHlz2HdxTHtGGF0dRJIn6JzxmlW0lezpg_8Zusj5QAihtRJSqefojHOlJBXyHP3-ADPYOYw9nsx8E4fYB2sG7MHMS4KMzejwlMCFLcknY9cATiH_LFY8YreYoYIRUn_CP6pkTth0OaZpDvEIczrhcDR9IS15BTiACQ9g0lisF-iZN0OGi7t3h75_-vjt8kt1_fXz1eX768rWjM6VIlbaWgJ1zErPvLWMkZbUjMtOeel423Bbq7o4Oka6xinWUuUb4kTDuQe-Q1cb10Vz0FMqE6WTjibov46Yem3SHOwAmjXKNKXSe0EFtVK5TkhuRdsqKpqC26F3G2tauiM4C-OczPAI-jgyhhvdx1vdUsIEWQGv7wAp_logz_oYsoVhMCPEJWtWs1ZwxkrbHWJbqk0x5wT-vg0lepWBPuhVBnqVgd5kUIpePRzwvuTf0UvC2y0ByspvAySdbYDRliOnooWyk_A__h959MVq</addsrcrecordid><sourcetype>Open Website</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2528432247</pqid></control><display><type>article</type><title>Detecting pathological features and predicting fracture risk from dual-energy X-ray absorptiometry images using deep learning</title><source>ScienceDirect</source><source>PubMed Central</source><creator>Nissinen, Tomi ; Suoranta, Sanna ; Saavalainen, Taavi ; Sund, Reijo ; Hurskainen, Ossi ; Rikkonen, Toni ; Kröger, Heikki ; Lähivaara, Timo ; Väänänen, Sami P.</creator><creatorcontrib>Nissinen, Tomi ; Suoranta, Sanna ; Saavalainen, Taavi ; Sund, Reijo ; Hurskainen, Ossi ; Rikkonen, Toni ; Kröger, Heikki ; Lähivaara, Timo ; Väänänen, Sami P.</creatorcontrib><description>Dual-energy X-ray absorptiometry (DXA) is the gold standard imaging method for diagnosing osteoporosis in clinical practice. The DXA images are commonly used to estimate a numerical value for bone mineral density (BMD), which decreases in osteoporosis. Low BMD is a known risk factor for osteoporotic fractures. In this study, we used deep learning to identify lumbar scoliosis and structural abnormalities that potentially affect BMD but are often neglected in lumbar spine DXA analysis. In addition, we tested the approach's ability to predict fractures using only DXA images. A dataset of 2949 images gathered by Kuopio Osteoporosis Risk Factor and Prevention Study was used to train a convolutional neural network (CNN) for classification. The model was able to classify scoliosis with an AUC of 0.96 and structural abnormalities causing unreliable BMD measurement with an AUC of 0.91. It predicted fractures occurring within 5 years from the lumbar spine DXA scan with an AUC of 0.63, meeting the predictive performance of combined BMD measurements from the lumbar spine and hip. In an independent test set of 574 clinical patients, AUC for lumbar scoliosis was 0.93 and AUC for unreliable BMD measurements was 0.94. In each classification task, neural network visualizations indicated the model's predictive strategy. We conclude that deep learning could complement the well established DXA method for osteoporosis diagnostics by analyzing incidental findings and image reliability, and improve its predictive ability in the future.</description><identifier>ISSN: 2352-1872</identifier><identifier>EISSN: 2352-1872</identifier><identifier>DOI: 10.1016/j.bonr.2021.101070</identifier><identifier>PMID: 33997147</identifier><language>eng</language><publisher>United States: Elsevier Inc</publisher><subject>Deep learning ; Degeneration ; Dual-energy X-ray absorptiometry ; Fracture risk ; Lumbar spine ; Scoliosis ; Trabecular bone score</subject><ispartof>Bone Reports, 2021-06, Vol.14, p.101070-101070, Article 101070</ispartof><rights>2021 The Author(s)</rights><rights>2021 The Author(s).</rights><rights>2021 The Author(s) 2021</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3</citedby><cites>FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://www.ncbi.nlm.nih.gov/pmc/articles/PMC8102403/pdf/$$EPDF$$P50$$Gpubmedcentral$$Hfree_for_read</linktopdf><linktohtml>$$Uhttps://www.sciencedirect.com/science/article/pii/S2352187221003259$$EHTML$$P50$$Gelsevier$$Hfree_for_read</linktohtml><link.rule.ids>230,314,727,780,784,885,3549,27924,27925,45780,53791,53793</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/33997147$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Nissinen, Tomi</creatorcontrib><creatorcontrib>Suoranta, Sanna</creatorcontrib><creatorcontrib>Saavalainen, Taavi</creatorcontrib><creatorcontrib>Sund, Reijo</creatorcontrib><creatorcontrib>Hurskainen, Ossi</creatorcontrib><creatorcontrib>Rikkonen, Toni</creatorcontrib><creatorcontrib>Kröger, Heikki</creatorcontrib><creatorcontrib>Lähivaara, Timo</creatorcontrib><creatorcontrib>Väänänen, Sami P.</creatorcontrib><title>Detecting pathological features and predicting fracture risk from dual-energy X-ray absorptiometry images using deep learning</title><title>Bone Reports</title><addtitle>Bone Rep</addtitle><description>Dual-energy X-ray absorptiometry (DXA) is the gold standard imaging method for diagnosing osteoporosis in clinical practice. The DXA images are commonly used to estimate a numerical value for bone mineral density (BMD), which decreases in osteoporosis. Low BMD is a known risk factor for osteoporotic fractures. In this study, we used deep learning to identify lumbar scoliosis and structural abnormalities that potentially affect BMD but are often neglected in lumbar spine DXA analysis. In addition, we tested the approach's ability to predict fractures using only DXA images. A dataset of 2949 images gathered by Kuopio Osteoporosis Risk Factor and Prevention Study was used to train a convolutional neural network (CNN) for classification. The model was able to classify scoliosis with an AUC of 0.96 and structural abnormalities causing unreliable BMD measurement with an AUC of 0.91. It predicted fractures occurring within 5 years from the lumbar spine DXA scan with an AUC of 0.63, meeting the predictive performance of combined BMD measurements from the lumbar spine and hip. In an independent test set of 574 clinical patients, AUC for lumbar scoliosis was 0.93 and AUC for unreliable BMD measurements was 0.94. In each classification task, neural network visualizations indicated the model's predictive strategy. We conclude that deep learning could complement the well established DXA method for osteoporosis diagnostics by analyzing incidental findings and image reliability, and improve its predictive ability in the future.</description><subject>Deep learning</subject><subject>Degeneration</subject><subject>Dual-energy X-ray absorptiometry</subject><subject>Fracture risk</subject><subject>Lumbar spine</subject><subject>Scoliosis</subject><subject>Trabecular bone score</subject><issn>2352-1872</issn><issn>2352-1872</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2021</creationdate><recordtype>article</recordtype><sourceid>DOA</sourceid><recordid>eNp9kktv1DAQxy0EotXSL8AB-cgli1-JYwkhofKqVIkLSNwsxx6nXrJxsJNKe-C745BStRdO9rx-M5r5I_SSkj0ltHlz2HdxTHtGGF0dRJIn6JzxmlW0lezpg_8Zusj5QAihtRJSqefojHOlJBXyHP3-ADPYOYw9nsx8E4fYB2sG7MHMS4KMzejwlMCFLcknY9cATiH_LFY8YreYoYIRUn_CP6pkTth0OaZpDvEIczrhcDR9IS15BTiACQ9g0lisF-iZN0OGi7t3h75_-vjt8kt1_fXz1eX768rWjM6VIlbaWgJ1zErPvLWMkZbUjMtOeel423Bbq7o4Oka6xinWUuUb4kTDuQe-Q1cb10Vz0FMqE6WTjibov46Yem3SHOwAmjXKNKXSe0EFtVK5TkhuRdsqKpqC26F3G2tauiM4C-OczPAI-jgyhhvdx1vdUsIEWQGv7wAp_logz_oYsoVhMCPEJWtWs1ZwxkrbHWJbqk0x5wT-vg0lepWBPuhVBnqVgd5kUIpePRzwvuTf0UvC2y0ByspvAySdbYDRliOnooWyk_A__h959MVq</recordid><startdate>20210601</startdate><enddate>20210601</enddate><creator>Nissinen, Tomi</creator><creator>Suoranta, Sanna</creator><creator>Saavalainen, Taavi</creator><creator>Sund, Reijo</creator><creator>Hurskainen, Ossi</creator><creator>Rikkonen, Toni</creator><creator>Kröger, Heikki</creator><creator>Lähivaara, Timo</creator><creator>Väänänen, Sami P.</creator><general>Elsevier Inc</general><general>Elsevier</general><scope>6I.</scope><scope>AAFTH</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7X8</scope><scope>5PM</scope><scope>DOA</scope></search><sort><creationdate>20210601</creationdate><title>Detecting pathological features and predicting fracture risk from dual-energy X-ray absorptiometry images using deep learning</title><author>Nissinen, Tomi ; Suoranta, Sanna ; Saavalainen, Taavi ; Sund, Reijo ; Hurskainen, Ossi ; Rikkonen, Toni ; Kröger, Heikki ; Lähivaara, Timo ; Väänänen, Sami P.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2021</creationdate><topic>Deep learning</topic><topic>Degeneration</topic><topic>Dual-energy X-ray absorptiometry</topic><topic>Fracture risk</topic><topic>Lumbar spine</topic><topic>Scoliosis</topic><topic>Trabecular bone score</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Nissinen, Tomi</creatorcontrib><creatorcontrib>Suoranta, Sanna</creatorcontrib><creatorcontrib>Saavalainen, Taavi</creatorcontrib><creatorcontrib>Sund, Reijo</creatorcontrib><creatorcontrib>Hurskainen, Ossi</creatorcontrib><creatorcontrib>Rikkonen, Toni</creatorcontrib><creatorcontrib>Kröger, Heikki</creatorcontrib><creatorcontrib>Lähivaara, Timo</creatorcontrib><creatorcontrib>Väänänen, Sami P.</creatorcontrib><collection>ScienceDirect Open Access Titles</collection><collection>Elsevier:ScienceDirect:Open Access</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>MEDLINE - Academic</collection><collection>PubMed Central (Full Participant titles)</collection><collection>Directory of Open Access Journals</collection><jtitle>Bone Reports</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Nissinen, Tomi</au><au>Suoranta, Sanna</au><au>Saavalainen, Taavi</au><au>Sund, Reijo</au><au>Hurskainen, Ossi</au><au>Rikkonen, Toni</au><au>Kröger, Heikki</au><au>Lähivaara, Timo</au><au>Väänänen, Sami P.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Detecting pathological features and predicting fracture risk from dual-energy X-ray absorptiometry images using deep learning</atitle><jtitle>Bone Reports</jtitle><addtitle>Bone Rep</addtitle><date>2021-06-01</date><risdate>2021</risdate><volume>14</volume><spage>101070</spage><epage>101070</epage><pages>101070-101070</pages><artnum>101070</artnum><issn>2352-1872</issn><eissn>2352-1872</eissn><abstract>Dual-energy X-ray absorptiometry (DXA) is the gold standard imaging method for diagnosing osteoporosis in clinical practice. The DXA images are commonly used to estimate a numerical value for bone mineral density (BMD), which decreases in osteoporosis. Low BMD is a known risk factor for osteoporotic fractures. In this study, we used deep learning to identify lumbar scoliosis and structural abnormalities that potentially affect BMD but are often neglected in lumbar spine DXA analysis. In addition, we tested the approach's ability to predict fractures using only DXA images. A dataset of 2949 images gathered by Kuopio Osteoporosis Risk Factor and Prevention Study was used to train a convolutional neural network (CNN) for classification. The model was able to classify scoliosis with an AUC of 0.96 and structural abnormalities causing unreliable BMD measurement with an AUC of 0.91. It predicted fractures occurring within 5 years from the lumbar spine DXA scan with an AUC of 0.63, meeting the predictive performance of combined BMD measurements from the lumbar spine and hip. In an independent test set of 574 clinical patients, AUC for lumbar scoliosis was 0.93 and AUC for unreliable BMD measurements was 0.94. In each classification task, neural network visualizations indicated the model's predictive strategy. We conclude that deep learning could complement the well established DXA method for osteoporosis diagnostics by analyzing incidental findings and image reliability, and improve its predictive ability in the future.</abstract><cop>United States</cop><pub>Elsevier Inc</pub><pmid>33997147</pmid><doi>10.1016/j.bonr.2021.101070</doi><tpages>1</tpages><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | ISSN: 2352-1872 |
ispartof | Bone Reports, 2021-06, Vol.14, p.101070-101070, Article 101070 |
issn | 2352-1872 2352-1872 |
language | eng |
recordid | cdi_doaj_primary_oai_doaj_org_article_269a660dff4141c79db473c488914663 |
source | ScienceDirect; PubMed Central |
subjects | Deep learning Degeneration Dual-energy X-ray absorptiometry Fracture risk Lumbar spine Scoliosis Trabecular bone score |
title | Detecting pathological features and predicting fracture risk from dual-energy X-ray absorptiometry images using deep learning |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-23T14%3A34%3A50IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_doaj_&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Detecting%20pathological%20features%20and%20predicting%20fracture%20risk%20from%20dual-energy%20X-ray%20absorptiometry%20images%20using%20deep%20learning&rft.jtitle=Bone%20Reports&rft.au=Nissinen,%20Tomi&rft.date=2021-06-01&rft.volume=14&rft.spage=101070&rft.epage=101070&rft.pages=101070-101070&rft.artnum=101070&rft.issn=2352-1872&rft.eissn=2352-1872&rft_id=info:doi/10.1016/j.bonr.2021.101070&rft_dat=%3Cproquest_doaj_%3E2528432247%3C/proquest_doaj_%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c521t-90c7c57e1d2c7f2fcc220805237b9f7d3863c595523b20b6d92819f60d4633fe3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2528432247&rft_id=info:pmid/33997147&rfr_iscdi=true |