1
|
Sun D, Nguyen TM, Allaway RJ, Wang J, Chung V, Yu TV, Mason M, Dimitrovsky I, Ericson L, Li H, Guan Y, Israel A, Olar A, Pataki BA, Stolovitzky G, Guinney J, Gulko PS, Frazier MB, Chen JY, Costello JC, Bridges SL. A Crowdsourcing Approach to Develop Machine Learning Models to Quantify Radiographic Joint Damage in Rheumatoid Arthritis. JAMA Netw Open 2022; 5:e2227423. [PMID: 36036935 PMCID: PMC9425151 DOI: 10.1001/jamanetworkopen.2022.27423] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 04/01/2022] [Accepted: 07/01/2022] [Indexed: 02/01/2023] Open
Abstract
Importance An automated, accurate method is needed for unbiased assessment quantifying accrual of joint space narrowing and erosions on radiographic images of the hands and wrists, and feet for clinical trials, monitoring of joint damage over time, assisting rheumatologists with treatment decisions. Such a method has the potential to be directly integrated into electronic health records. Objectives To design and implement an international crowdsourcing competition to catalyze the development of machine learning methods to quantify radiographic damage in rheumatoid arthritis (RA). Design, Setting, and Participants This diagnostic/prognostic study describes the Rheumatoid Arthritis 2-Dialogue for Reverse Engineering Assessment and Methods (RA2-DREAM Challenge), which used existing radiographic images and expert-curated Sharp-van der Heijde (SvH) scores from 2 clinical studies (674 radiographic sets from 562 patients) for training (367 sets), leaderboard (119 sets), and final evaluation (188 sets). Challenge participants were tasked with developing methods to automatically quantify overall damage (subchallenge 1), joint space narrowing (subchallenge 2), and erosions (subchallenge 3). The challenge was finished on June 30, 2020. Main Outcomes and Measures Scores derived from submitted algorithms were compared with the expert-curated SvH scores, and a baseline model was created for benchmark comparison. Performances were ranked using weighted root mean square error (RMSE). The performance and reproductivity of each algorithm was assessed using Bayes factor from bootstrapped data, and further evaluated with a postchallenge independent validation data set. Results The RA2-DREAM Challenge received a total of 173 submissions from 26 participants or teams in 7 countries for the leaderboard round, and 13 submissions were included in the final evaluation. The weighted RMSEs metric showed that the winning algorithms produced scores that were very close to the expert-curated SvH scores. Top teams included Team Shirin for subchallenge 1 (weighted RMSE, 0.44), HYL-YFG (Hongyang Li and Yuanfang Guan) subchallenge 2 (weighted RMSE, 0.38), and Gold Therapy for subchallenge 3 (weighted RMSE, 0.43). Bootstrapping/Bayes factor approach and the postchallenge independent validation confirmed the reproducibility and the estimation concordance indices between final evaluation and postchallenge independent validation data set were 0.71 for subchallenge 1, 0.78 for subchallenge 2, and 0.82 for subchallenge 3. Conclusions and Relevance The RA2-DREAM Challenge resulted in the development of algorithms that provide feasible, quick, and accurate methods to quantify joint damage in RA. Ultimately, these methods could help research studies on RA joint damage and may be integrated into electronic health records to help clinicians serve patients better by providing timely, reliable, and quantitative information for making treatment decisions to prevent further damage.
Collapse
Affiliation(s)
- Dongmei Sun
- University of Alabama at Birmingham, Birmingham
- Division of Rheumatology, Department of Medicine, Hospital for Special Surgery, New York, New York
| | | | | | - Jelai Wang
- University of Alabama at Birmingham, Birmingham
| | | | | | | | | | | | - Hongyang Li
- Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor
| | - Yuanfang Guan
- Department of Computational Medicine and Bioinformatics, University of Michigan, Ann Arbor
| | - Ariel Israel
- Leumit Health Services, Tel-Aviv, Israel
- Medical Solutions for Digital Medicine, Jerusalem, Israel
| | - Alex Olar
- Department of Complex Systems in Physics, Eötvös Loránd University, Budapest, Hungary
| | - Balint Armin Pataki
- Department of Complex Systems in Physics, Eötvös Loránd University, Budapest, Hungary
| | - Gustavo Stolovitzky
- T. J. Watson Research Center, IBM, Yorktown Heights, New York
- Now with Sema4, Stamford, Connecticut
| | | | - Percio S. Gulko
- Division of Rheumatology, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York
| | | | | | - James C. Costello
- Department of Pharmacology, University of Colorado Anschutz Medical Campus, Aurora
| | - S. Louis Bridges
- University of Alabama at Birmingham, Birmingham
- Division of Rheumatology, Department of Medicine, Hospital for Special Surgery, New York, New York
- Division of Rheumatology, Weill Cornell Medical College, New York, New York
| |
Collapse
|
2
|
Pfeil A, Oelzner P, Hoffmann T, Renz DM, Wolf G, Böttcher J. Sind röntgenologische Scoring-Methoden als Parameter zur
Verlaufsbeurteilung der rheumatoiden Arthritis noch
zeitgemäß? AKTUEL RHEUMATOL 2021. [DOI: 10.1055/a-1394-0299] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/21/2022]
Abstract
ZusammenfassungDie radiologische Progression beschreibt das Ausmaß der
Gelenkzerstörung im Verlauf einer rheumatoiden Arthritis. Zur
Quantifizierung der radiologischen Progression werden Scoring-Methoden
(z. B. van der Heijde Modifikation des Sharp-Score) eingesetzt. In
verschiedenen Studien zu biologischen- bzw. target-synthetischen Disease
Modifying Anti-Rheumatic Drugs gelang nur unzureichend eine Differenzierung
der radiologischen Progression. Zudem finden die Scores oft keinen
routinemäßigen Einsatz in der klinischen
Entscheidungsfindung. Durch die computerbasierte Analyse von
Handröntgenaufnahmen ist eine valide Quantifizierung der
radiologischen Progression und die zuverlässige Bewertung von
Therapieeffekten möglich. Somit stellen die computerbasierten
Methoden eine vielversprechende Alternative in der Quantifizierung der
radiologischen Progression dar.
Collapse
Affiliation(s)
- Alexander Pfeil
- Klinik für Innere Medizin III, Universitätsklinikum
Jena, Jena, Deutschland
| | - Peter Oelzner
- Klinik für Innere Medizin III, Universitätsklinikum
Jena, Jena, Deutschland
| | - Tobias Hoffmann
- Klinik für Innere Medizin III, Universitätsklinikum
Jena, Jena, Deutschland
| | - Diane M. Renz
- Institut für Diagnostische und Interventionelle Radiologie,
Medizinische Hochschule Hannover, Hannover, Deutschland
| | - Gunter Wolf
- Klinik für Innere Medizin III, Universitätsklinikum
Jena, Jena, Deutschland
| | - Joachim Böttcher
- Medizinische Fakultät, Friedrich-Schiller-Universität
Jena, Jena, Deutschland
| |
Collapse
|
3
|
Boesen M, Kubassova O, Sudoł-Szopińska I, Maas M, Hansen P, Nybing JD, Oei EH, Hemke R, Guermazi A. MR Imaging of Joint Infection and Inflammation with Emphasis on Dynamic Contrast-Enhanced MR Imaging. PET Clin 2018; 13:523-550. [PMID: 30219186 DOI: 10.1016/j.cpet.2018.05.007] [Citation(s) in RCA: 19] [Impact Index Per Article: 3.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/15/2023]
Abstract
Contrast-enhanced MR imaging (CE-MR imaging) is recommended for diagnosis and monitoring of infectious and most inflammatory joint diseases. CE-MR imaging clearly differentiates soft and bony tissue from fluid collections and infectious debris. To improve imaging information, a dynamic CE-MR imaging sequence (DCE-MR imaging) sequence can be applied using fast T1-weighted sequential image acquisition during contrast injection. Use of DCE-MR imaging allows robust extraction of quantitative information regarding blood flow and capillary permeability, especially when dedicated analysis methods and software are used to analyze contrast kinetics. This article describes principles of DCE-MR imaging for the assessment of infectious and inflammatory joint diseases.
Collapse
Affiliation(s)
- Mikael Boesen
- Department of Radiology, Bispebjerg and Frederiksberg Hospital, Bispebjerg Bakke 23, 2400, Copenhagen Nv, Denmark; Parker Institute, Bispebjerg and Frederiksberg Hospital, Nordrefasanvej 57, 2000 Copenhagen F, Denmark.
| | - Olga Kubassova
- Image Analysis Group (IAG), AQBC Minster House, 272-274 Vauxhall Bridge Road, SW1V 1BA, London, UK
| | - Iwona Sudoł-Szopińska
- Department of Radiology, National Institute of Geriatrics, Rheumatology and Rehabilitation, Warsaw, Poland; Department of Diagnostic Imaging, Warsaw Medical University, Warsaw, Poland
| | - Mario Maas
- Department of Radiology, Faculty of Medicine, Academic Medical Center (AMC) Amsterdam, University of Amsterdam, Amsterdam, The Netherlands; Department of Nuclear Medicine, Faculty of Medicine, Academic Medical Center (AMC) Amsterdam, University of Amsterdam, Amsterdam, The Netherlands
| | - Philip Hansen
- Department of Radiology, Bispebjerg and Frederiksberg Hospital, Bispebjerg Bakke 23, 2400, Copenhagen Nv, Denmark
| | - Janus Damm Nybing
- Department of Radiology, Bispebjerg and Frederiksberg Hospital, Bispebjerg Bakke 23, 2400, Copenhagen Nv, Denmark
| | - Edwin H Oei
- Department of Radiology, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands; Department of Nuclear Medicine, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands
| | - Robert Hemke
- Department of Radiology, Faculty of Medicine, Academic Medical Center (AMC) Amsterdam, University of Amsterdam, Amsterdam, The Netherlands; Department of Nuclear Medicine, Faculty of Medicine, Academic Medical Center (AMC) Amsterdam, University of Amsterdam, Amsterdam, The Netherlands
| | - Ali Guermazi
- Department of Radiology, Boston University School of Medicine, Boston, MA, USA
| |
Collapse
|
4
|
Riis RGC, Gudbergsen H, Henriksen M, Ballegaard C, Bandak E, Röttger D, Bliddal H, Hansen BB, Hangaard S, Boesen M. Synovitis assessed on static and dynamic contrast-enhanced magnetic resonance imaging and its association with pain in knee osteoarthritis: A cross-sectional study. Eur J Radiol 2016; 85:1099-108. [PMID: 27161058 DOI: 10.1016/j.ejrad.2016.03.017] [Citation(s) in RCA: 44] [Impact Index Per Article: 5.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/12/2016] [Revised: 03/01/2016] [Accepted: 03/18/2016] [Indexed: 01/11/2023]
Abstract
OBJECTIVES To investigate the association between pain and peripatellar-synovitis on static and dynamic contrast-enhanced MRI in knee osteoarthritis. METHODS In a cross-sectional setting, knee synovitis was assessed using 3-Tesla MRI and correlated with pain using the knee injury and osteoarthritis outcome score (KOOS). Synovitis was assessed in the peripatellar recesses with: (i) dynamic contrast-enhanced (DCE)-MRI, using both pharmacokinetic and heuristic models, (ii) contrast-enhanced (CE)-MRI, and (iii) non-CE-MRI. The DCE-MRI variable IRExNvoxel was chosen as the primary variable in the analyses. RESULTS Valid data were available in 94 persons with a mean age of 65 years, a BMI of 32.3kg/m(2) and a mean Kellgren-Lawrence grade of 2.5. IRExNvoxel showed a statically significant correlation with KOOS-Pain (r=-0.34; p=0.001), as was the case with all DCE-variables but one. Correlations between static MRI-variables and KOOS-Pain ranged between -0.21<r<-0.29 (p<0.040). Intraclass correlation coefficients ranged between 0.90-0.99 for the heuristic and 0.66-0.93 for the pharmacokinetic DCE-MRI variables. CONCLUSIONS The results confirm an association between peripatellar-synovitis and pain in KOA. Overall, DCE-MRI showed stronger correlations with KOOS-Pain compared to static MRI. DCE-MRI analyses were highly reproducible and have the potential to be used to further investigate the role of inflammation and perfusion in KOA.
Collapse
Affiliation(s)
- Robert G C Riis
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark; Department of Radiology, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Henrik Gudbergsen
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Marius Henriksen
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Christine Ballegaard
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Elisabeth Bandak
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Diana Röttger
- Image Analysis Ltd., QABC Minster House, 272-274 Vauxhall Bridge Road, London SW1 V 1BA, United Kingdom
| | - Henning Bliddal
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Bjarke Brandt Hansen
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Stine Hangaard
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark; Department of Radiology, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark
| | - Mikael Boesen
- The Parker Institute, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark; Department of Radiology, Copenhagen University Hospital, Bispebjerg-Frederiksberg, Nordre Fasanvej 57, DK-2000 Frederiksberg, Denmark.
| |
Collapse
|
5
|
Tripathi D, Agarwal V. Quantifying synovial inflammation: Emerging imaging techniques. World J Rheumatol 2014; 4:72-79. [DOI: 10.5499/wjr.v4.i3.72] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.1] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 04/12/2014] [Revised: 07/23/2014] [Accepted: 09/10/2014] [Indexed: 02/06/2023] Open
Abstract
Imaging techniques to assess synovial inflammation includes radiography, ultrasound, computed tomography, magnetic resonance imaging (MRI) and recently positron emission tomography. The ideal objective of imaging approaches are to quantify synovial inflammation by capturing features such as synovial hyperplasia, neo-angiogenesis and infiltration of immune cells in the synovium. This may enable clinicians to estimate response to therapy by measuring the improvement in the inflammatory signals at the level of synovium. Ultrasound can provide information regarding thickening of the synovial membrane and can reveal increased synovial blood flow using power Doppler technique. Bone marrow edema and synovial membrane thickness on MRI scan may serve as indicators for arthritis progression. Enhancement of the synovium on dynamic contrast MRI may closely mirror the inflammatory activity in the synovium. Diffusion tensor imaging is an advance MRI approach that evaluates the inflammation related to cell infiltration or aggregation in an inflamed synovium. In this review, we summarize the newer imaging techniques and their developments to evaluate synovial inflammation.
Collapse
|
6
|
Abstract
New MRI techniques have been developed to assess not only the static anatomy of synovial hyperplasia, bone changes and cartilage degradation in patients with rheumatoid arthritis (RA), but also the activity of the physiological events that cause these changes. This enables an estimation of the rate of change in the synovium, bone and cartilage as a result of disease activity or in response to therapy. Typical MRI signs of RA in the pre-erosive phase include synovitis, bone marrow edema and subchondral cyst formation. Synovitis can be assessed by T2-weighted imaging, dynamic contrast-enhanced MRI or diffusion tensor imaging. Bone marrow edema can be detected on fluid-sensitive sequences such as short-tau inversion recovery or T2-weighted fast-spin echo sequences. Detection of small bone erosions in the early erosive phase using T1-weighted MRI has sensitivity comparable to CT. Numerous MRI techniques have been developed for quantitative assessment of potentially pathologic changes in cartilage composition that occur before frank morphologic changes. In this Review, we summarize the advances and new directions in the field of MRI, with an emphasis on their current state of development and application in RA.
Collapse
Affiliation(s)
- Camilo G Borrero
- Department of Radiology, University of Pittsburgh, 200 Lothrop Street, Pittsburgh, PA 15213, USA
| | | | | |
Collapse
|