Skip to main content
Springer Nature Link
Log in

Differenzierung MR-tomographischer Arthrosephänotypen mithilfe von Techniken der Künstlichen Intelligenz

Differentiation of osteoarthritis phenotypes on MRI using artificial intelligence

  • Leitthema
  • Published:
Die Orthopädie Aims and scope Submit manuscript

Zusammenfassung

Hintergrund

Arthrose (Osteoarthritis [OA]) zählt mit mehr als 500 Mio. Betroffenen weltweit zu den häufigsten Gelenkerkrankungen. In den letzten Jahrzehnten gab es nur begrenzte Fortschritte bezüglich Diagnostik und Therapie. Während die OA lange Zeit primär als mechanisch bedingte Verschleißerkrankung betrachtet wurde, zeigen neuere Studien, dass es sich vielmehr um ein heterogenes Krankheitsbild handelt, welches sich in unterschiedlichen Phänotypen manifestiert. Obwohl die Künstliche Intelligenz (KI) in der medizinischen Forschung zunehmend an Bedeutung gewinnt, bleibt ihre konkrete Anwendung im Bereich der OA klinisch bislang noch wenig genutzt.

Ziel der Arbeit

Ziel dieser Übersichtsarbeit ist, die derzeitigen Ansätze zur Phänotypisierung der OA zusammenzufassen und dabei insbesondere die Rolle der KI bei der Identifikation und Klassifikation von OA-Phänotypen hervorzuheben.

Material und Methoden

Selektive Literaturrecherche

Ergebnisse

Es gibt verschiedene vielversprechende Anwendungen von KI in der Diagnose und Bewertung der OA, wie die automatisierte Beurteilung von Knorpelschäden oder die Vorhersage der späteren Notwendigkeit der endoprothetischen Versorgung. Eine enge Kooperation zwischen Orthopädie, Radiologie und KI-Expertinnen und Experten ist notwendig, um KI-Modelle in die Praxis zu integrieren.

Schlussfolgerung

Die Anwendung von KI zur Erkennung und Beurteilung OA-typischer Veränderungen bietet enormes Potenzial, um sowohl die diagnostische Bildgebung als auch die klinische Interpretation und Verlaufsvorhersage der Erkrankung entscheidend zu verbessern. Durch präzisere Diagnosen und individuellere Prognosen könnten KI-basierte Verfahren maßgeblich dazu beitragen, Therapieentscheidungen effektiver zu gestalten und damit die Patientenversorgung zu optimieren.

Abstract

Background

Osteoarthritis (OA) is one of the most common joint diseases, affecting more than 500 million people worldwide. In recent decades, there has only been limited progress in terms of diagnosis and treatment. For a long time, OA was considered to be primarily a mechanically induced degenerative disease. However, more recent work has shown that OA is a heterogeneous condition that manifests in different phenotypes. Although artificial intelligence (AI) is becoming increasingly important in medical research, its specific application in the field of OA remains limited in clinical use.

Objectives

The aim of this review is to summarize the current approaches to phenotyping OA and to highlight the role of AI in the identification and classification of OA phenotypes.

Materials and methods

Selective literature review

Results

There are several promising applications of AI in OA diagnosis and assessment, such as automated assessment of cartilage damage or prediction of the need for arthroplasty. Close cooperation between orthopaedics, radiology, and AI experts is necessary to integrate AI models into clinical practice.

Conclusions

The use of AI to detect and assess OA-typical changes offers major potential to improve diagnostic imaging, clinical interpretation, and disease prognosis. Through more precise diagnoses and individualized prognoses, AI-based methods could significantly contribute to making treatment decisions more effective and, thus, optimizing patient care overall.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+
from €37.37 /Month
  • Starting from 10 chapters or articles per month
  • Access and download chapters and articles from more than 300k books and 2,500 journals
  • Cancel anytime
View plans

Buy Now

Price includes VAT (Norway)

Instant access to the full article PDF.

Institutional subscriptions

Abb. 1
Abb. 2
Abb. 3

Similar content being viewed by others

Explore related subjects

Discover the latest articles, books and news in related subjects, suggested using machine learning.

Abbreviations

AUC:

„Area under the curve“

CNN:

„Convolutional neural networks“

DL:

„Deep learning“

KI:

Künstliche Intelligenz

Knie-TEP:

Knietotalendoprothese

ML:

Maschinelles Lernen („machine learning“)

OA:

Arthrose (Osteoarthritis)

ROC:

„Receiver operating characteristic“

WORMS:

Whole-Organ Magnetic Resonance Imaging Score

Literatur

  1. Kloppenburg M, Namane M, Cicuttini F (2025) Osteoarthritis. Lancet 405:71–85

    Article  PubMed  Google Scholar 

  2. Long H, Liu Q, Yin H et al (2022) Prevalence Trends of Site-Specific Osteoarthritis From 1990 to 2019: Findings From the Global Burden of Disease Study 2019. Arthritis Rheumatol 74:1172–1183

    Article  PubMed  PubMed Central  Google Scholar 

  3. Martel-Pelletier J, Barr AJ, Cicuttini FM et al (2016) Osteoarthritis. Nat Rev Dis Primers 2:16072

    Article  PubMed  Google Scholar 

  4. Mobasheri A, Loeser R (2024) Clinical phenotypes, molecular endotypes and theratypes in OA therapeutic development. Nat Rev Rheumatol 20:525–526

    Article  PubMed  Google Scholar 

  5. Roemer FW, Jarraya M, Collins JE et al (2023) Structural phenotypes of knee osteoarthritis: potential clinical and research relevance. Skelet Radiol 52:2021–2030

    Article  Google Scholar 

  6. Van Spil WE, Kubassova O, Boesen M et al (2019) Osteoarthritis phenotypes and novel therapeutic targets. Biochem Pharmacol 165:41–48

    Article  PubMed  Google Scholar 

  7. Dell’isola A, Steultjens M (2018) Classification of patients with knee osteoarthritis in clinical phenotypes: Data from the osteoarthritis initiative. Plos One 13:e191045

    Article  PubMed  PubMed Central  Google Scholar 

  8. Horii M, Kubo T, Hirasawa Y (2000) Radial MRI of the hip with moderate osteoarthritis. J Bone Joint Surg Br 82:364–368

    Article  CAS  PubMed  Google Scholar 

  9. Stelzeneder D, Mamisch TC, Kress I et al (2012) Patterns of joint damage seen on MRI in early hip osteoarthritis due to structural hip deformities. Osteoarthritis Cartilage 20:661–669

    Article  CAS  PubMed  Google Scholar 

  10. Lee S, Nardo L, Kumar D et al (2015) Scoring hip osteoarthritis with MRI (SHOMRI): A whole joint osteoarthritis evaluation system. J Magn Reson Imaging 41:1549–1557

    Article  PubMed  Google Scholar 

  11. Sergooris A, Verbrugghe J, Bonnechère B et al (2024) Beyond the Hip: Clinical Phenotypes of Hip Osteoarthritis Across the Biopsychosocial Spectrum. J Clin Med 13:6824

  12. Kellgren JH, Lawrence JS (1957) Radiological assessment of osteo-arthrosis. Ann Rheum Dis 16:494–502

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Roemer FW, Hunter DJ, Winterstein A et al (2011) Hip Osteoarthritis MRI Scoring System (HOAMS): reliability and associations with radiographic and clinical findings. Osteoarthritis Cartilage 19:946–962

    Article  CAS  PubMed  Google Scholar 

  14. Neumann G, Mendicuti AD, Zou KH et al (2007) Prevalence of labral tears and cartilage loss in patients with mechanical symptoms of the hip: evaluation using MR arthrography. Osteoarthritis Cartilage 15:909–917

    Article  CAS  PubMed  Google Scholar 

  15. Roemer FW, Guermazi A, Demehri S et al (2022) Imaging in Osteoarthritis. Osteoarthritis Cartilage 30:913–934

    Article  CAS  PubMed  Google Scholar 

  16. Fu MC, Buerba RA, Long WD 3rd et al (2014) Interrater and intrarater agreements of magnetic resonance imaging findings in the lumbar spine: significant variability across degenerative conditions. Spine J 14:2442–2448

    Article  PubMed  Google Scholar 

  17. Holzinger A, Langs G, Denk H et al (2019) Causability and explainability of artificial intelligence in medicine. Wiley Interdiscip Rev Data Min Knowl Discov 9:e1312

    Article  PubMed  PubMed Central  Google Scholar 

  18. Caliva F, Namiri NK, Dubreuil M et al (2022) Studying osteoarthritis with artificial intelligence applied to magnetic resonance imaging. Nat Rev Rheumatol 18:112–121

    Article  PubMed  Google Scholar 

  19. Choi RY, Coyner AS, Kalpathy-Cramer J et al (2020) Introduction to Machine Learning, Neural Networks, and Deep Learning. Transl Vis Sci Technol 9:14

    PubMed  PubMed Central  Google Scholar 

  20. Krizhevsky A, Sutskever I, Hinton GE (2017) ImageNet classification with deep convolutional neural networks. Commun ACM 60:84–90

    Article  Google Scholar 

  21. Zeng DY, Shaikh J, Holmes S et al (2019) Deep residual network for off-resonance artifact correction with application to pediatric body MRA with 3D cones. Magn Reson Med 82:1398–1411

    Article  PubMed  PubMed Central  Google Scholar 

  22. Smith CP, Harmon SA, Barrett T et al (2019) Intra- and interreader reproducibility of PI-RADSv2: A multireader study. J Magn Reson Imaging 49:1694–1703

    Article  PubMed  Google Scholar 

  23. Fernquest S, Park D, Marcan M et al (2018) Segmentation of hip cartilage in compositional magnetic resonance imaging: A fast, accurate, reproducible, and clinically viable semi-automated methodology. J Orthop Res

  24. Liukkonen MK, Mononen ME, Tanska P et al (2017) Application of a semi-automatic cartilage segmentation method for biomechanical modeling of the knee joint. Comput Methods Biomech Biomed Engin 20:1453–1463

    Article  PubMed  Google Scholar 

  25. Jaremko JL, Cheng RW, Lambert RG et al (2006) Reliability of an efficient MRI-based method for estimation of knee cartilage volume using surface registration. Osteoarthritis Cartilage 14:914–922

    Article  CAS  PubMed  Google Scholar 

  26. Gaj S, Yang M, Nakamura K et al (2020) Automated cartilage and meniscus segmentation of knee MRI with conditional generative adversarial networks. Magn Reson Med 84:437–449

    Article  PubMed  Google Scholar 

  27. Norman B, Pedoia V, Majumdar S (2018) Use of 2D U‑Net Convolutional Neural Networks for Automated Cartilage and Meniscus Segmentation of Knee MR Imaging Data to Determine Relaxometry and Morphometry. Radiology 288:177–185

    Article  PubMed  Google Scholar 

  28. Yang M, Colak C, Chundru KK et al (2022) Automated knee cartilage segmentation for heterogeneous clinical MRI using generative adversarial networks with transfer learning. Quant Imaging Med Surg 12:2620–2633

    Article  PubMed  PubMed Central  Google Scholar 

  29. Yang Z, Fripp J, Chandra SS et al (2015) Automatic bone segmentation and bone-cartilage interface extraction for the shoulder joint from magnetic resonance images. Phys Med Biol 60:1441–1459

    Article  PubMed  Google Scholar 

  30. Xia Y, Fripp J, Chandra SS et al (2013) Automated bone segmentation from large field of view 3D MR images of the hip joint. Phys Med Biol 58:7375–7390

    Article  PubMed  Google Scholar 

  31. Deniz CM, Xiang S, Hallyburton RS et al (2018) Segmentation of the Proximal Femur from MR Images using Deep Convolutional Neural Networks. Sci Rep 8:16485

    Article  PubMed  PubMed Central  Google Scholar 

  32. Brui E, Efimtcev AY, Fokin VA et al (2020) Deep learning-based fully automatic segmentation of wrist cartilage in MR images. NMR Biomed 33:e4320

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Ronneberger O, Fischer P, Brox T (2015) U‑net: Convolutional networks for biomedical image segmentation. In: Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference. Springer, Munich, Germany, S 234–241

    Google Scholar 

  34. Isensee F, Jaeger PF, Kohl SA et al (2021) nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 18:203–211

    Article  CAS  PubMed  Google Scholar 

  35. Hesamian MH, Jia W, He X et al (2019) Deep Learning Techniques for Medical Image Segmentation: Achievements and Challenges. J Digit Imaging 32:582–596

    Article  PubMed  PubMed Central  Google Scholar 

  36. Shin HC, Roth HR, Gao M et al (2016) Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning. IEEE Trans Med Imaging 35:1285–1298

    Article  PubMed  Google Scholar 

  37. Yosinski J, Clune J, Bengio Y et al (2014) How transferable are features in deep neural networks? Advances in Neural Information Processing Systems (NIPS) Bd 27

    Google Scholar 

  38. Akinci D’antonoli T, Berger LK, Indrakanti AK et al (2025) TotalSegmentator MRI: Robust Sequence-independent Segmentation of Multiple Anatomic Structures in MRI. Radiology 314:e241613

    Article  PubMed  Google Scholar 

  39. Cardoso MJ, Li W, Brown R et al (2022) Monai: An open-source framework for deep learning in healthcare. arXiv preprint arXiv:2211.02701

    Google Scholar 

  40. Peterfy CG, Guermazi A, Zaim S et al (2004) Whole-Organ Magnetic Resonance Imaging Score (WORMS) of the knee in osteoarthritis. Osteoarthritis Cartilage 12:177–190

    Article  CAS  PubMed  Google Scholar 

  41. Hunter DJ, Lo GH, Gale D et al (2008) The reliability of a new scoring system for knee osteoarthritis MRI and the validity of bone marrow lesion assessment: BLOKS (Boston Leeds Osteoarthritis Knee Score). Ann Rheum Dis 67:206–211

    Article  CAS  PubMed  Google Scholar 

  42. Hunter DJ, Guermazi A, Lo GH et al (2011) Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS (MRI Osteoarthritis Knee Score). Osteoarthritis Cartilage 19:990–1002

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  43. Tibrewala R, Ozhinsky E, Shah R et al (2020) Computer-Aided Detection AI Reduces Interreader Variability in Grading Hip Abnormalities With MRI. J Magn Reson Imaging 52:1163–1172

    Article  PubMed  PubMed Central  Google Scholar 

  44. Liu F, Zhou Z, Samsonov A et al (2018) Deep Learning Approach for Evaluating Knee MR Images: Achieving High Diagnostic Performance for Cartilage Lesion Detection. Radiology 289:160–169

    Article  PubMed  Google Scholar 

  45. Felfeliyan B, Wichuk S, Hareendranathan AR et al (2024) OMERACT validation of a deep learning algorithm for automated absolute quantification of knee joint effusion versus manual semi-quantitative assessment. Semin Arthritis Rheum 66:152420

    Article  PubMed  Google Scholar 

  46. Pedoia V, Norman B, Mehany SN et al (2019) 3D convolutional neural networks for detection and severity staging of meniscus and PFJ cartilage morphological degenerative changes in osteoarthritis and anterior cruciate ligament subjects. J Magn Reson Imaging 49:400–410

    Article  PubMed  Google Scholar 

  47. Schiratti JB, Dubois R, Herent P et al (2021) A deep learning method for predicting knee osteoarthritis radiographic progression from MRI. Arthritis Res Ther 23:262

    Article  PubMed  PubMed Central  Google Scholar 

  48. Tolpadi AA, Lee JJ, Pedoia V et al (2020) Deep Learning Predicts Total Knee Replacement from Magnetic Resonance Images. Sci Rep 10:6371

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Förderung

T. Rolvien erhält Förderung durch die Deutsche Forschungsgemeinschaft (DFG) im Rahmen der Klinischen Forschungsgruppe ProBone (RO 5925/5-1).

Author information

Authors and Affiliations

  1. Klinik für Unfallchirurgie und Orthopädie, Universitätsklinikum Hamburg-Eppendorf, Martinistraße 52, 20246, Hamburg, Deutschland

    Kirsten Schleid, Assil-Ramin Alimy, Tim Hoenig, Frank Timo Beil &  Tim Rolvien

  2. Klinik für Diagnostische und Interventionelle Radiologie, Uniklinik RWTH Aachen, Aachen, Deutschland

    Simon Westfechtel & Sven Nebelung

Authors
  1. Kirsten Schleid
    View author publications

    Search author on:PubMed Google Scholar

  2. Assil-Ramin Alimy
    View author publications

    Search author on:PubMed Google Scholar

  3. Tim Hoenig
    View author publications

    Search author on:PubMed Google Scholar

  4. Simon Westfechtel
    View author publications

    Search author on:PubMed Google Scholar

  5. Sven Nebelung
    View author publications

    Search author on:PubMed Google Scholar

  6. Frank Timo Beil
    View author publications

    Search author on:PubMed Google Scholar

  7. Tim Rolvien
    View author publications

    Search author on:PubMed Google Scholar

Corresponding author

Correspondence to Tim Rolvien.

Ethics declarations

Interessenkonflikt

K. Schleid, A.-R. Alimy, T. Hoenig, S. Westfechtel, S. Nebelung, F.T. Beil und T. Rolvien geben an, dass kein Interessenkonflikt besteht.

Für diesen Beitrag wurden von den Autor/-innen keine Studien an Menschen oder Tieren durchgeführt. Für die aufgeführten Studien gelten die jeweils dort angegebenen ethischen Richtlinien.

Additional information

Hinweis des Verlags

Der Verlag bleibt in Hinblick auf geografische Zuordnungen und Gebietsbezeichnungen in veröffentlichten Karten und Institutsadressen neutral.

figure qr

QR-Code scannen & Beitrag online lesen

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Schleid, K., Alimy, AR., Hoenig, T. et al. Differenzierung MR-tomographischer Arthrosephänotypen mithilfe von Techniken der Künstlichen Intelligenz. Orthopädie (2025). https://doi.org/10.1007/s00132-025-04751-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Version of record:

  • DOI: https://doi.org/10.1007/s00132-025-04751-3

Schlüsselwörter

Keywords