Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis
Identifieur interne : 000831 ( Pmc/Checkpoint ); précédent : 000830; suivant : 000832Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis
Auteurs : Toru Hirano ; Masayuki Nishide ; Naoki Nonaka ; Jun Seita ; Kosuke Ebina [Japon] ; Kazuhiro Sakurada ; Atsushi KumanogohSource :
- Rheumatology Advances in Practice [ 2514-1775 ] ; 2019.
Abstract
The purpose of this research was to develop a deep-learning model to assess radiographic finger joint destruction in RA.
The model comprises two steps: a joint-detection step and a joint-evaluation step. Among 216 radiographs of 108 patients with RA, 186 radiographs were assigned to the training/validation dataset and 30 to the test dataset. In the training/validation dataset, images of PIP joints, the IP joint of the thumb or MCP joints were manually clipped and scored for joint space narrowing (JSN) and bone erosion by clinicians, and then these images were augmented. As a result, 11 160 images were used to train and validate a deep convolutional neural network for joint evaluation. Three thousand seven hundred and twenty selected images were used to train machine learning for joint detection. These steps were combined as the assessment model for radiographic finger joint destruction. Performance of the model was examined using the test dataset, which was not included in the training/validation process, by comparing the scores assigned by the model and clinicians.
The model detected PIP joints, the IP joint of the thumb and MCP joints with a sensitivity of 95.3% and assigned scores for JSN and erosion. Accuracy (percentage of exact agreement) reached 49.3–65.4% for JSN and 70.6–74.1% for erosion. The correlation coefficient between scores by the model and clinicians per image was 0.72–0.88 for JSN and 0.54–0.75 for erosion.
Image processing with the trained convolutional neural network model is promising to assess radiographs in RA.
Url:
DOI: 10.1093/rap/rkz047
PubMed: 31872173
PubMed Central: 6921374
Affiliations:
Links toward previous steps (curation, corpus...)
Links to Exploration step
PMC:6921374Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis</title><author><name sortKey="Hirano, Toru" sort="Hirano, Toru" uniqKey="Hirano T" first="Toru" last="Hirano">Toru Hirano</name><affiliation><nlm:aff id="rkz047-aff1"><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</nlm:aff><wicri:noCountry code="subfield">Osaka</wicri:noCountry></affiliation></author><author><name sortKey="Nishide, Masayuki" sort="Nishide, Masayuki" uniqKey="Nishide M" first="Masayuki" last="Nishide">Masayuki Nishide</name><affiliation><nlm:aff id="rkz047-aff1"><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</nlm:aff><wicri:noCountry code="subfield">Osaka</wicri:noCountry></affiliation></author><author><name sortKey="Nonaka, Naoki" sort="Nonaka, Naoki" uniqKey="Nonaka N" first="Naoki" last="Nonaka">Naoki Nonaka</name><affiliation><nlm:aff id="rkz047-aff2"><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</nlm:aff><wicri:noCountry code="subfield">Kanagawa</wicri:noCountry></affiliation></author><author><name sortKey="Seita, Jun" sort="Seita, Jun" uniqKey="Seita J" first="Jun" last="Seita">Jun Seita</name><affiliation><nlm:aff id="rkz047-aff2"><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</nlm:aff><wicri:noCountry code="subfield">Kanagawa</wicri:noCountry></affiliation></author><author><name sortKey="Ebina, Kosuke" sort="Ebina, Kosuke" uniqKey="Ebina K" first="Kosuke" last="Ebina">Kosuke Ebina</name><affiliation wicri:level="1"><nlm:aff id="rkz047-aff3"><institution>Department of Musculoskeletal Regenerative Medicine, Graduate School of Medicine, Osaka University</institution>, Suita, Osaka,<country country="JP">Japan</country></nlm:aff><country xml:lang="fr">Japon</country><wicri:regionArea># see nlm:aff country strict</wicri:regionArea></affiliation></author><author><name sortKey="Sakurada, Kazuhiro" sort="Sakurada, Kazuhiro" uniqKey="Sakurada K" first="Kazuhiro" last="Sakurada">Kazuhiro Sakurada</name><affiliation><nlm:aff id="rkz047-aff2"><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</nlm:aff><wicri:noCountry code="subfield">Kanagawa</wicri:noCountry></affiliation></author><author><name sortKey="Kumanogoh, Atsushi" sort="Kumanogoh, Atsushi" uniqKey="Kumanogoh A" first="Atsushi" last="Kumanogoh">Atsushi Kumanogoh</name><affiliation><nlm:aff id="rkz047-aff1"><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</nlm:aff><wicri:noCountry code="subfield">Osaka</wicri:noCountry></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">PMC</idno><idno type="pmid">31872173</idno><idno type="pmc">6921374</idno><idno type="url">http://www.ncbi.nlm.nih.gov/pmc/articles/PMC6921374</idno><idno type="RBID">PMC:6921374</idno><idno type="doi">10.1093/rap/rkz047</idno><date when="2019">2019</date><idno type="wicri:Area/Pmc/Corpus">000803</idno><idno type="wicri:explorRef" wicri:stream="Pmc" wicri:step="Corpus" wicri:corpus="PMC">000803</idno><idno type="wicri:Area/Pmc/Curation">000803</idno><idno type="wicri:explorRef" wicri:stream="Pmc" wicri:step="Curation">000803</idno><idno type="wicri:Area/Pmc/Checkpoint">000831</idno><idno type="wicri:explorRef" wicri:stream="Pmc" wicri:step="Checkpoint">000831</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en" level="a" type="main">Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis</title><author><name sortKey="Hirano, Toru" sort="Hirano, Toru" uniqKey="Hirano T" first="Toru" last="Hirano">Toru Hirano</name><affiliation><nlm:aff id="rkz047-aff1"><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</nlm:aff><wicri:noCountry code="subfield">Osaka</wicri:noCountry></affiliation></author><author><name sortKey="Nishide, Masayuki" sort="Nishide, Masayuki" uniqKey="Nishide M" first="Masayuki" last="Nishide">Masayuki Nishide</name><affiliation><nlm:aff id="rkz047-aff1"><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</nlm:aff><wicri:noCountry code="subfield">Osaka</wicri:noCountry></affiliation></author><author><name sortKey="Nonaka, Naoki" sort="Nonaka, Naoki" uniqKey="Nonaka N" first="Naoki" last="Nonaka">Naoki Nonaka</name><affiliation><nlm:aff id="rkz047-aff2"><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</nlm:aff><wicri:noCountry code="subfield">Kanagawa</wicri:noCountry></affiliation></author><author><name sortKey="Seita, Jun" sort="Seita, Jun" uniqKey="Seita J" first="Jun" last="Seita">Jun Seita</name><affiliation><nlm:aff id="rkz047-aff2"><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</nlm:aff><wicri:noCountry code="subfield">Kanagawa</wicri:noCountry></affiliation></author><author><name sortKey="Ebina, Kosuke" sort="Ebina, Kosuke" uniqKey="Ebina K" first="Kosuke" last="Ebina">Kosuke Ebina</name><affiliation wicri:level="1"><nlm:aff id="rkz047-aff3"><institution>Department of Musculoskeletal Regenerative Medicine, Graduate School of Medicine, Osaka University</institution>, Suita, Osaka,<country country="JP">Japan</country></nlm:aff><country xml:lang="fr">Japon</country><wicri:regionArea># see nlm:aff country strict</wicri:regionArea></affiliation></author><author><name sortKey="Sakurada, Kazuhiro" sort="Sakurada, Kazuhiro" uniqKey="Sakurada K" first="Kazuhiro" last="Sakurada">Kazuhiro Sakurada</name><affiliation><nlm:aff id="rkz047-aff2"><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</nlm:aff><wicri:noCountry code="subfield">Kanagawa</wicri:noCountry></affiliation></author><author><name sortKey="Kumanogoh, Atsushi" sort="Kumanogoh, Atsushi" uniqKey="Kumanogoh A" first="Atsushi" last="Kumanogoh">Atsushi Kumanogoh</name><affiliation><nlm:aff id="rkz047-aff1"><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</nlm:aff><wicri:noCountry code="subfield">Osaka</wicri:noCountry></affiliation></author></analytic><series><title level="j">Rheumatology Advances in Practice</title><idno type="eISSN">2514-1775</idno><imprint><date when="2019">2019</date></imprint></series></biblStruct></sourceDesc></fileDesc><profileDesc><textClass></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en"><title>Abstract</title><sec id="s1"><title>Objective</title><p>The purpose of this research was to develop a deep-learning model to assess radiographic finger joint destruction in RA.</p></sec><sec id="s2"><title>Methods</title><p>The model comprises two steps: a joint-detection step and a joint-evaluation step. Among 216 radiographs of 108 patients with RA, 186 radiographs were assigned to the training/validation dataset and 30 to the test dataset. In the training/validation dataset, images of PIP joints, the IP joint of the thumb or MCP joints were manually clipped and scored for joint space narrowing (JSN) and bone erosion by clinicians, and then these images were augmented. As a result, 11 160 images were used to train and validate a deep convolutional neural network for joint evaluation. Three thousand seven hundred and twenty selected images were used to train machine learning for joint detection. These steps were combined as the assessment model for radiographic finger joint destruction. Performance of the model was examined using the test dataset, which was not included in the training/validation process, by comparing the scores assigned by the model and clinicians.</p></sec><sec id="s3"><title>Results</title><p>The model detected PIP joints, the IP joint of the thumb and MCP joints with a sensitivity of 95.3% and assigned scores for JSN and erosion. Accuracy (percentage of exact agreement) reached 49.3–65.4% for JSN and 70.6–74.1% for erosion. The correlation coefficient between scores by the model and clinicians per image was 0.72–0.88 for JSN and 0.54–0.75 for erosion.</p></sec><sec id="s4"><title>Conclusion</title><p>Image processing with the trained convolutional neural network model is promising to assess radiographs in RA.</p></sec></div></front><back><div1 type="bibliography"><listBibl><biblStruct><analytic><author><name sortKey="Lecun, Y" uniqKey="Lecun Y">Y LeCun</name></author><author><name sortKey="Bengio, Y" uniqKey="Bengio Y">Y Bengio</name></author><author><name sortKey="Hinton, G" uniqKey="Hinton G">G. Hinton</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Webb, S" uniqKey="Webb S">S. Webb</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Chartrand, G" uniqKey="Chartrand G">G Chartrand</name></author><author><name sortKey="Cheng, Pm" uniqKey="Cheng P">PM Cheng</name></author><author><name sortKey="Vorontsov, E" uniqKey="Vorontsov E">E Vorontsov</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Gulshan, V" uniqKey="Gulshan V">V Gulshan</name></author><author><name sortKey="Peng, L" uniqKey="Peng L">L Peng</name></author><author><name sortKey="Coram, M" uniqKey="Coram M">M Coram</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Esteva, A" uniqKey="Esteva A">A Esteva</name></author><author><name sortKey="Kuprel, B" uniqKey="Kuprel B">B Kuprel</name></author><author><name sortKey="Novoa, Ra" uniqKey="Novoa R">RA Novoa</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Fujisawa, Y" uniqKey="Fujisawa Y">Y Fujisawa</name></author><author><name sortKey="Otomo, Y" uniqKey="Otomo Y">Y Otomo</name></author><author><name sortKey="Ogata, Y" uniqKey="Ogata Y">Y Ogata</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Coudray, N" uniqKey="Coudray N">N Coudray</name></author><author><name sortKey="Ocampo, Ps" uniqKey="Ocampo P">PS Ocampo</name></author><author><name sortKey="Sakellaropoulos, T" uniqKey="Sakellaropoulos T">T Sakellaropoulos</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Walsh, Slf" uniqKey="Walsh S">SLF Walsh</name></author><author><name sortKey="Calandriello, L" uniqKey="Calandriello L">L Calandriello</name></author><author><name sortKey="Silva, M" uniqKey="Silva M">M Silva</name></author><author><name sortKey="Sverzellati, N" uniqKey="Sverzellati N">N. Sverzellati</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Smolen, Js" uniqKey="Smolen J">JS Smolen</name></author><author><name sortKey="Landewe, R" uniqKey="Landewe R">R Landewé</name></author><author><name sortKey="Bijlsma, J" uniqKey="Bijlsma J">J Bijlsma</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Steinbrocker, O" uniqKey="Steinbrocker O">O Steinbrocker</name></author><author><name sortKey="Traeger, Ch" uniqKey="Traeger C">CH Traeger</name></author><author><name sortKey="Batterman, Rc" uniqKey="Batterman R">RC. Batterman</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Kellgren, Jh" uniqKey="Kellgren J">JH Kellgren</name></author><author><name sortKey="Bier, F" uniqKey="Bier F">F. Bier</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Sharp, Jt" uniqKey="Sharp J">JT Sharp</name></author><author><name sortKey="Lidsky, Md" uniqKey="Lidsky M">MD Lidsky</name></author><author><name sortKey="Collins, Lc" uniqKey="Collins L">LC Collins</name></author><author><name sortKey="Moreland, J" uniqKey="Moreland J">J. Moreland</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Sharp, Jt" uniqKey="Sharp J">JT Sharp</name></author><author><name sortKey="Bluhm, Gb" uniqKey="Bluhm G">GB Bluhm</name></author><author><name sortKey="Brook, A" uniqKey="Brook A">A Brook</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Larsen, A" uniqKey="Larsen A">A Larsen</name></author><author><name sortKey="Dale, K" uniqKey="Dale K">K Dale</name></author><author><name sortKey="Eek, M" uniqKey="Eek M">M. Eek</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Genant, Hk" uniqKey="Genant H">HK. Genant</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Van Der Heijde, Dm" uniqKey="Van Der Heijde D">DM Van der Heijde</name></author><author><name sortKey="Van Riel, Pl" uniqKey="Van Riel P">PL Van Riel</name></author><author><name sortKey="Nuver Zwart, Ih" uniqKey="Nuver Zwart I">IH Nuver-Zwart</name></author><author><name sortKey="Gribnau, Fw" uniqKey="Gribnau F">FW Gribnau</name></author><author><name sortKey="Van De Putte, Lb" uniqKey="Van De Putte L">LB. Van de Putte</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Arnett, Fc" uniqKey="Arnett F">FC Arnett</name></author><author><name sortKey="Edworthy, Sm" uniqKey="Edworthy S">SM Edworthy</name></author><author><name sortKey="Bloch, Da" uniqKey="Bloch D">DA Bloch</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Aletaha, D" uniqKey="Aletaha D">D Aletaha</name></author><author><name sortKey="Neogi, T" uniqKey="Neogi T">T Neogi</name></author><author><name sortKey="Silman, Aj" uniqKey="Silman A">AJ Silman</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Van Der Heijde, D" uniqKey="Van Der Heijde D">D. van der Heijde</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Viola, P" uniqKey="Viola P">P Viola</name></author><author><name sortKey="Jones, M" uniqKey="Jones M">M. Jones</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Kingma, Dp" uniqKey="Kingma D">DP Kingma</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Loffe, S" uniqKey="Loffe S">S Loffe</name></author><author><name sortKey="Szegedy, C" uniqKey="Szegedy C">C. Szegedy</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Tokui, S" uniqKey="Tokui S">S Tokui</name></author><author><name sortKey="Oono, K" uniqKey="Oono K">K Oono</name></author><author><name sortKey="Hido, S" uniqKey="Hido S">S Hido</name></author><author><name sortKey="Clayton, J" uniqKey="Clayton J">J. Clayton</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Meyer, O" uniqKey="Meyer O">O Meyer</name></author><author><name sortKey="Labarre, C" uniqKey="Labarre C">C Labarre</name></author><author><name sortKey="Dougados, M" uniqKey="Dougados M">M Dougados</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Vencovsk, J" uniqKey="Vencovsk J">J Vencovský</name></author><author><name sortKey="Machacek, S" uniqKey="Machacek S">S Machácek</name></author><author><name sortKey="Sedova, L" uniqKey="Sedova L">L Sedová</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Koga, T" uniqKey="Koga T">T Koga</name></author><author><name sortKey="Okada, A" uniqKey="Okada A">A Okada</name></author><author><name sortKey="Fukuda, T" uniqKey="Fukuda T">T Fukuda</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Xue, Y" uniqKey="Xue Y">Y Xue</name></author><author><name sortKey="Zhang, R" uniqKey="Zhang R">R Zhang</name></author><author><name sortKey="Deng, Y" uniqKey="Deng Y">Y Deng</name></author><author><name sortKey="Chen, K" uniqKey="Chen K">K Chen</name></author><author><name sortKey="Jiang, T" uniqKey="Jiang T">T. Jiang</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Spampinato, C" uniqKey="Spampinato C">C Spampinato</name></author><author><name sortKey="Palazzo, S" uniqKey="Palazzo S">S Palazzo</name></author><author><name sortKey="Giordano, D" uniqKey="Giordano D">D Giordano</name></author><author><name sortKey="Aldinucci, M" uniqKey="Aldinucci M">M Aldinucci</name></author><author><name sortKey="Leonardi, R" uniqKey="Leonardi R">R. Leonardi</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Larson, Db" uniqKey="Larson D">DB Larson</name></author><author><name sortKey="Chen, Mc" uniqKey="Chen M">MC Chen</name></author><author><name sortKey="Lungren, Mp" uniqKey="Lungren M">MP Lungren</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Lindsey, R" uniqKey="Lindsey R">R Lindsey</name></author><author><name sortKey="Daluiski, A" uniqKey="Daluiski A">A Daluiski</name></author><author><name sortKey="Chopra, S" uniqKey="Chopra S">S Chopra</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Liu, F" uniqKey="Liu F">F Liu</name></author><author><name sortKey="Zhou, Z" uniqKey="Zhou Z">Z Zhou</name></author><author><name sortKey="Samsonov, A" uniqKey="Samsonov A">A Samsonov</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Tack, A" uniqKey="Tack A">A Tack</name></author><author><name sortKey="Mukhopadhyay, A" uniqKey="Mukhopadhyay A">A Mukhopadhyay</name></author><author><name sortKey="Zachow, S" uniqKey="Zachow S">S. Zachow</name></author></analytic></biblStruct><biblStruct><analytic><author><name sortKey="Bien, N" uniqKey="Bien N">N Bien</name></author><author><name sortKey="Rajpurkar, P" uniqKey="Rajpurkar P">P Rajpurkar</name></author><author><name sortKey="Ball, Rl" uniqKey="Ball R">RL Ball</name></author></analytic></biblStruct></listBibl></div1></back></TEI><pmc article-type="research-article"><pmc-dir>properties open_access</pmc-dir>
<front><journal-meta><journal-id journal-id-type="nlm-ta">Rheumatol Adv Pract</journal-id><journal-id journal-id-type="iso-abbrev">Rheumatol Adv Pract</journal-id><journal-id journal-id-type="publisher-id">rheumap</journal-id><journal-title-group><journal-title>Rheumatology Advances in Practice</journal-title></journal-title-group><issn pub-type="epub">2514-1775</issn><publisher><publisher-name>Oxford University Press</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="pmid">31872173</article-id><article-id pub-id-type="pmc">6921374</article-id><article-id pub-id-type="doi">10.1093/rap/rkz047</article-id><article-id pub-id-type="publisher-id">rkz047</article-id><article-categories><subj-group subj-group-type="heading"><subject>Original Article</subject></subj-group></article-categories><title-group><article-title>Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis</article-title></title-group><contrib-group><contrib contrib-type="author"><contrib-id contrib-id-type="orcid" authenticated="false">http://orcid.org/0000-0001-8467-3154</contrib-id><name><surname>Hirano</surname><given-names>Toru</given-names></name><xref ref-type="corresp" rid="rkz047-cor1"></xref><xref ref-type="aff" rid="rkz047-aff1">1</xref><pmc-comment>thirano@imed3.med.osaka-u.ac.jp </pmc-comment>
</contrib><contrib contrib-type="author"><name><surname>Nishide</surname><given-names>Masayuki</given-names></name><xref ref-type="aff" rid="rkz047-aff1">1</xref></contrib><contrib contrib-type="author"><name><surname>Nonaka</surname><given-names>Naoki</given-names></name><xref ref-type="aff" rid="rkz047-aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Seita</surname><given-names>Jun</given-names></name><xref ref-type="aff" rid="rkz047-aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Ebina</surname><given-names>Kosuke</given-names></name><xref ref-type="aff" rid="rkz047-aff3">3</xref></contrib><contrib contrib-type="author"><name><surname>Sakurada</surname><given-names>Kazuhiro</given-names></name><xref ref-type="aff" rid="rkz047-aff2">2</xref></contrib><contrib contrib-type="author"><name><surname>Kumanogoh</surname><given-names>Atsushi</given-names></name><xref ref-type="aff" rid="rkz047-aff1">1</xref></contrib></contrib-group><aff id="rkz047-aff1"><label>1</label><institution>Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine</institution>, Osaka University, Suita, Osaka</aff><aff id="rkz047-aff2"><label>2</label><institution>Medical Sciences Innovation Hub Program</institution>, RIKEN, Yokohama, Kanagawa</aff><aff id="rkz047-aff3"><label>3</label><institution>Department of Musculoskeletal Regenerative Medicine, Graduate School of Medicine, Osaka University</institution>, Suita, Osaka,<country country="JP">Japan</country></aff><author-notes><corresp id="rkz047-cor1">Correspondence to: Toru Hirano, Department of Respiratory Medicine and Clinical Immunology, Internal Medicine, Graduate School of Medicine, Osaka University, 2-2 Yamadaoka, Suita, Osaka 565-0871, Japan. E-mail: <email>thirano@imed3.med.osaka-u.ac.jp</email></corresp></author-notes><pub-date pub-type="collection"><year>2019</year></pub-date><pub-date pub-type="epub" iso-8601-date="2019-11-22"><day>22</day><month>11</month><year>2019</year></pub-date><pub-date pub-type="pmc-release"><day>22</day><month>11</month><year>2019</year></pub-date><pmc-comment> PMC Release delay is 0 months and 0 days and was based on the
<volume>3</volume><issue>2</issue><elocation-id>rkz047</elocation-id><history><date date-type="received"><day>03</day><month>9</month><year>2019</year></date><date date-type="rev-recd"><day>4</day><month>11</month><year>2019</year></date></history><permissions><copyright-statement>© The Author(s) 2019. Published by Oxford University Press on behalf of the British Society for Rheumatology.</copyright-statement><copyright-year>2019</copyright-year><license license-type="cc-by-nc" xlink:href="http://creativecommons.org/licenses/by-nc/4.0/"><license-p>This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (<ext-link ext-link-type="uri" xlink:href="http://creativecommons.org/licenses/by-nc/4.0/">http://creativecommons.org/licenses/by-nc/4.0/</ext-link>), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com</license-p></license></permissions><self-uri xlink:href="rkz047.pdf"></self-uri><abstract><title>Abstract</title><sec id="s1"><title>Objective</title><p>The purpose of this research was to develop a deep-learning model to assess radiographic finger joint destruction in RA.</p></sec><sec id="s2"><title>Methods</title><p>The model comprises two steps: a joint-detection step and a joint-evaluation step. Among 216 radiographs of 108 patients with RA, 186 radiographs were assigned to the training/validation dataset and 30 to the test dataset. In the training/validation dataset, images of PIP joints, the IP joint of the thumb or MCP joints were manually clipped and scored for joint space narrowing (JSN) and bone erosion by clinicians, and then these images were augmented. As a result, 11 160 images were used to train and validate a deep convolutional neural network for joint evaluation. Three thousand seven hundred and twenty selected images were used to train machine learning for joint detection. These steps were combined as the assessment model for radiographic finger joint destruction. Performance of the model was examined using the test dataset, which was not included in the training/validation process, by comparing the scores assigned by the model and clinicians.</p></sec><sec id="s3"><title>Results</title><p>The model detected PIP joints, the IP joint of the thumb and MCP joints with a sensitivity of 95.3% and assigned scores for JSN and erosion. Accuracy (percentage of exact agreement) reached 49.3–65.4% for JSN and 70.6–74.1% for erosion. The correlation coefficient between scores by the model and clinicians per image was 0.72–0.88 for JSN and 0.54–0.75 for erosion.</p></sec><sec id="s4"><title>Conclusion</title><p>Image processing with the trained convolutional neural network model is promising to assess radiographs in RA.</p></sec></abstract><kwd-group><kwd>rheumatoid arthritis</kwd><kwd>joint destruction</kwd><kwd>artificial intelligence</kwd></kwd-group><funding-group><award-group award-type="grant"><funding-source><named-content content-type="funder-name">Japan Science and Technology Agency (JST)</named-content></funding-source></award-group><award-group award-type="grant"><funding-source><named-content content-type="funder-name">Osaka University</named-content><named-content content-type="funder-identifier">10.13039/501100004206</named-content></funding-source></award-group><award-group award-type="grant"><funding-source><named-content content-type="funder-name">Council for Science, Technology and Innovation (CSTI)</named-content></funding-source></award-group><award-group award-type="grant"><funding-source><named-content content-type="funder-name">Innovation AI Hospital System</named-content></funding-source></award-group><award-group award-type="grant"><funding-source><named-content content-type="funder-name">National Institute of Biomedical Innovation</named-content><named-content content-type="funder-identifier">10.13039/501100007077</named-content></funding-source></award-group><award-group award-type="grant"><funding-source><named-content content-type="funder-name">Health and Nutrition (NIBIOHN)</named-content></funding-source><award-id>JPMJIH1504</award-id></award-group></funding-group><counts><page-count count="8"></page-count></counts></article-meta></front></pmc><affiliations><list><country><li>Japon</li></country></list><tree><noCountry><name sortKey="Hirano, Toru" sort="Hirano, Toru" uniqKey="Hirano T" first="Toru" last="Hirano">Toru Hirano</name><name sortKey="Kumanogoh, Atsushi" sort="Kumanogoh, Atsushi" uniqKey="Kumanogoh A" first="Atsushi" last="Kumanogoh">Atsushi Kumanogoh</name><name sortKey="Nishide, Masayuki" sort="Nishide, Masayuki" uniqKey="Nishide M" first="Masayuki" last="Nishide">Masayuki Nishide</name><name sortKey="Nonaka, Naoki" sort="Nonaka, Naoki" uniqKey="Nonaka N" first="Naoki" last="Nonaka">Naoki Nonaka</name><name sortKey="Sakurada, Kazuhiro" sort="Sakurada, Kazuhiro" uniqKey="Sakurada K" first="Kazuhiro" last="Sakurada">Kazuhiro Sakurada</name><name sortKey="Seita, Jun" sort="Seita, Jun" uniqKey="Seita J" first="Jun" last="Seita">Jun Seita</name></noCountry><country name="Japon"><noRegion><name sortKey="Ebina, Kosuke" sort="Ebina, Kosuke" uniqKey="Ebina K" first="Kosuke" last="Ebina">Kosuke Ebina</name></noRegion></country></tree></affiliations></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Sante/explor/ChloroquineV1/Data/Pmc/Checkpoint
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000831 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Pmc/Checkpoint/biblio.hfd -nk 000831 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Sante
|area= ChloroquineV1
|flux= Pmc
|étape= Checkpoint
|type= RBID
|clé= PMC:6921374
|texte= Development and validation of a deep-learning model for scoring of radiographic finger joint destruction in rheumatoid arthritis
}}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/Pmc/Checkpoint/RBID.i -Sk "pubmed:31872173" \
| HfdSelect -Kh $EXPLOR_AREA/Data/Pmc/Checkpoint/biblio.hfd \
| NlmPubMed2Wicri -a ChloroquineV1
| This area was generated with Dilib version V0.6.33. |  |