Zusammenfassung
Methoden der künstlichen Intelligenz (KI) haben in den letzten Jahren zunehmend Einzug in die Medizin gefunden. Einige Fachbereiche nutzen diese schon regelmäßig im klinischen Alltag. Die Anwendungsfelder sind weit, aber bisher noch nicht ausgeschöpft und in ihrer Vielfalt nicht ausreichend verstanden. Dieser Übersichtsbeitrag gibt einen Einblick in die Historie der KI und definiert die unterschiedlichen Begrifflichkeiten und Bereiche wie maschinelles Lernen (ML), neuronale Netze oder Deep Learning. Es werden die klassischen Schritte zur Entwicklung eines KI-Modells demonstriert sowie der Kreislauf der Datenbereinigung, -vorbereitung, das Training eines Modells bis hin zur Validierung und Umsetzung in der Praxis des klinischen Alltags erklärt. Bisherige Anwendungsfelder im muskuloskeletalen Fachbereich nutzen sowohl Methoden des ML als auch neuronaler Netze, z. B. zur Identifikation von Frakturen oder zur Klassifizierung. Prädikative Modelle anhand von Risikofaktoren mit dem Ziel der Komplikationsprävention finden erste Anwendung. Da Pseudarthrosen ein zwar seltenes, aber komplexes Krankheitsbild mit soziökonomischer Tragweite darstellen, ergeben sich viele noch offene Fragestellungen, die mithilfe der Methoden der KI zukünftig beantwortet werden könnten. Neue Forschungsfelder unter Nutzung von KI reichen von Prädikationsmodellen über Kostenanalysen bis hin zu personalisierter Therapie.
Abstract
Methods of artificial intelligence (AI) have found applications in many fields of medicine within the last few years. Some disciplines already use these methods regularly within their clinical routine. However, the fields of application are wide and there are still many opportunities to apply these new AI concepts. This review article gives an insight into the history of AI and defines the special terms and fields, such as machine learning (ML), neural networks and deep learning. The classical steps in developing AI models are demonstrated here, as well as the iteration of data rectification and preparation, the training of a model and subsequent validation before transfer into a clinical setting are explained. Currently, musculoskeletal disciplines implement methods of ML and also neural networks, e.g. for identification of fractures or for classifications. Also, predictive models based on risk factor analysis for prevention of complications are being initiated. As non-union in bone is a rare but very complex disease with dramatic socioeconomic impact for the healthcare system, many open questions arise which could be better understood by using methods of AI in the future. New fields of research applying AI models range from predictive models and cost analysis to personalized treatment strategies.
Literatur
Turing AM (1950) Computing machinery and intelligence. Mind 59:433–460. https://doi.org/10.1093/mind/LIX.236.433
McCarthy J, Minsky ML, Rochester N et al (1956) A proposal for the Dartmouth summer research project on artificial intelligence. Dartmouth Conference. Dartmouth College, Hanover, New Hampshire
Shortliffe E (1976) Computer-based medical consultations: MYCIN. Elsevier https://doi.org/10.1016/B978-0-444-00179-5.X5001-X
Lalehzarian SP, Gowd AK, Liu JN (2021) Machine learning in orthopaedic surgery. World J Orthop 12:685–699. https://doi.org/10.5312/wjo.v12.i9.685
Olczak J, Pavlopoulos J, Prijs J et al (2021) Presenting artificial intelligence, deep learning, and machine learning studies to clinicians and healthcare stakeholders: an introductory reference with a guideline and a Clinical AI Research (CAIR) checklist proposal. Acta Orthop 92:513–525. https://doi.org/10.1080/17453674.2021.1918389
Jordan MI, Mitchell TM (2015) Machine learning: trends, perspectives, and prospects. Science 349:255–260. https://doi.org/10.1126/science.aaa8415
Campbell M, Hoane AJ Jr, Hsu F (2002) Deep blue. Artif Intell 134:57–83. https://doi.org/10.1016/S0004-3702(01)00129-1
Ferrucci D, Brown E, Chu-Carroll J et al (2010) Building watson: an overview of the deepQA project. Artif Intell Mag 31(3):59–79. https://doi.org/10.1609/aimag.v31i3.2303
Silver D, Huang A, Maddison CJ et al (2016) Mastering the game of go with deep neural networks and tree search. Nature 529:484–489. https://doi.org/10.1038/nature16961
Wyatt JM, Booth GJ, Goldman AH (2021) Natural language processing and its use in orthopaedic research. Curr Rev Musculoskelet Med 14:392–396. https://doi.org/10.1007/s12178-021-09734-3
Thirukumaran CP, Zaman A, Rubery PT et al (2019) Natural language processing for the identification of surgical site infections in orthopaedics. J Bone Joint Surg Am 101:2167–2174. https://doi.org/10.2106/JBJS.19.00661
Phung VH, Rhee EJ (2019) A high-accuracy model average ensemble of convolutional neural networks for classification of cloud image patches on small datasets. Appl Sci. https://doi.org/10.3390/app9214500
Myers TG, Ramkumar PN, Ricciardi BF et al (2020) Artificial intelligence and orthopaedics: an introduction for clinicians. J Bone Joint Surg Am 102:830–840. https://doi.org/10.2106/JBJS.19.01128
Geron A (2019) Hands-on machine learning with Scikit—learn, Keras and tensor flow. O’Reilly Media
Martins LF (2014) IPython notebook essentials. Packt
Schmidhuber J (2015) Deep learning in neural networks: an overview. Neural Netw 61:85–117. https://doi.org/10.1016/j.neunet.2014.09.003
Zhao W, Davis CE (2011) A modified artificial immune system based pattern recognition approach—an application to clinical diagnostics. Artif Intell Med 52:1–9. https://doi.org/10.1016/j.artmed.2011.03.001
Kohonen T (2006) Self-organizing neural projections. Neural Netw 19:723–733. https://doi.org/10.1016/j.neunet.2006.05.001
LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444. https://doi.org/10.1038/nature14539
Borjali A, Chen AF, Muratoglu OK et al (2020) Detecting total hip replacement prosthesis design on plain radiographs using deep convolutional neural network. J Orthop Res 38:1465–1471. https://doi.org/10.1002/jor.24617
Erne F, Dehncke D, Herath SC et al (2021) Deep learning in the detection of rare fractures—development of a “deep learning convolutional network” model for detecting Acetabular fractures. Z Orthop Unfall. https://doi.org/10.1055/a-1511-8595
Olczak J, Emilson F, Razavian A et al (2021) Ankle fracture classification using deep learning: automating detailed AO Foundation/Orthopedic Trauma Association (AO/OTA) 2018 malleolar fracture identification reaches a high degree of correct classification. Acta Orthop 92:102–108. https://doi.org/10.1080/17453674.2020.1837420
Kruse C, Eiken P, Vestergaard P (2017) Machine learning principles can improve hip fracture prediction. Calcif Tissue Int 100:348–360. https://doi.org/10.1007/s00223-017-0238-7
Xue Y, Zhang R, Deng Y et al (2017) A preliminary examination of the diagnostic value of deep learning in hip osteoarthritis. PLoS ONE 12:e178992. https://doi.org/10.1371/journal.pone.0178992
Begg R, Kamruzzaman J (2005) A machine learning approach for automated recognition of movement patterns using basic, kinetic and kinematic gait data. J Biomech 38:401–408. https://doi.org/10.1016/j.jbiomech.2004.05.002
Joyseeree R, Abou Sabha R, Mueller H (2015) Applying machine learning to gait analysis data for disease identification. Stud Health Technol Inform 210:850–854
Sikka RS, Baer M, Raja A et al (2019) Analytics in sports medicine: implications and responsibilities that accompany the era of big data. J Bone Joint Surg Am 101:276–283. https://doi.org/10.2106/JBJS.17.01601
Ekegren CL, Edwards ER, de Steiger R et al (2018) Incidence, costs and predictors of non-union, delayed union and mal-union following long bone fracture. Int J Environ Res Public Health. https://doi.org/10.3390/ijerph15122845
McCoy TH Jr., Fragomen AT, Hart KL et al (2019) Genomewide association study of fracture nonunion using electronic health records. JBMR Plus 3:23–28. https://doi.org/10.1002/jbm4.10063
Mills LA, Aitken SA, Simpson A (2017) The risk of non-union per fracture: current myths and revised figures from a population of over 4 million adults. Acta Orthop 88:434–439. https://doi.org/10.1080/17453674.2017.1321351
Calori GM, Colombo M, Mazza EL et al (2014) Validation of the non-union scoring system in 300 long bone non-unions. Injury 45(Suppl 6):S93–S97. https://doi.org/10.1016/j.injury.2014.10.030
Calori GM, Phillips M, Jeetle S et al (2008) Classification of non-union: need for a new scoring system? Injury 39(Suppl 2):S59–S63. https://doi.org/10.1016/S0020-1383(08)70016-0
Santolini E, West RM, Giannoudis PV (2020) Leeds-Genoa Non-Union Index: a clinical tool for asessing the need for early intervention after long bone fracture fixation. Int Orthop 44:161–172. https://doi.org/10.1007/s00264-019-04376-0
Whelan DB, Bhandari M, Stephen D et al (2010) Development of the radiographic union score for tibial fractures for the assessment of tibial fracture healing after intramedullary fixation. J Trauma 68:629–632. https://doi.org/10.1097/TA.0b013e3181a7c16d
Karnuta JM, Navarro SM, Haeberle HS et al (2019) Bundled care for hip fractures: a machine-learning approach to an untenable patient-specific payment model. J Orthop Trauma 33:324–330. https://doi.org/10.1097/BOT.0000000000001454
Navarro SM, Wang EY, Haeberle HS et al (2018) Machine learning and primary total knee arthroplasty: patient forecasting for a patient-specific payment model. J Arthroplasty 33:3617–3623. https://doi.org/10.1016/j.arth.2018.08.028
Ramkumar PN, Haeberle HS, Bloomfield MR et al (2019) Artificial intelligence and arthroplasty at a single institution: real-world applications of machine learning to big data, value-based care, mobile health, and remote patient monitoring. J Arthroplasty 34:2204–2209. https://doi.org/10.1016/j.arth.2019.06.018
Ramkumar PN, Navarro SM, Haeberle HS et al (2019) Development and validation of a machine learning algorithm after primary total hip arthroplasty: applications to length of stay and payment models. J Arthroplasty 34:632–637. https://doi.org/10.1016/j.arth.2018.12.030
Beard DJ, Harris K, Dawson J et al (2015) Meaningful changes for the Oxford hip and knee scores after joint replacement surgery. J Clin Epidemiol 68:73–79. https://doi.org/10.1016/j.jclinepi.2014.08.009
Fontana MA, Lyman S, Sarker GK et al (2019) Can machine learning algorithms predict which patients Will achieve minimally clinically important differences from total joint arthroplasty? Clin Orthop Relat Res 477:1267–1279. https://doi.org/10.1097/CORR.0000000000000687
Keurentjes JC, Van Tol FR, Fiocco M et al (2012) Minimal clinically important differences in health-related quality of life after total hip or knee replacement: a systematic review. Bone Joint Res 1:71–77. https://doi.org/10.1302/2046-3758.15.2000065
Rupp M, Walter N, Pfeifer C et al (2021) Inzidenz von Frakturen in der Ewachsenenpopulation in Deutschland. Dtsch Arztebl Int 40:665–669
Danksagung
Wir möchten uns bei Dr. Dr. Matthias Reumann, Connected Health Insights International, für die fachliche Expertise zu künstlicher Intelligenz bedanken.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Interessenkonflikt
M.K. Reumann, B.J. Braun, M.M. Menger, F. Springer, J. Jazewitsch, T. Schwarz, A. Nüssler, T. Histing und M.F.R. Rollmann geben an, dass kein Interessenkonflikt besteht.
Für diesen Beitrag wurden von den Autoren keine Studien an Menschen oder Tieren durchgeführt. Für die aufgeführten Studien gelten die jeweils dort angegebenen ethischen Richtlinien.
Additional information
Redaktion
Benedikt J. Braun, Tübingen
Tina Histing, Tübingen
QR-Code scannen & Beitrag online lesen
Rights and permissions
About this article
Cite this article
Reumann, M.K., Braun, B.J., Menger, M.M. et al. Künstliche Intelligenz und Ausblick auf Anwendungsfelder in der Pseudarthrosentherapie. Unfallchirurgie 125, 611–618 (2022). https://doi.org/10.1007/s00113-022-01202-y
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00113-022-01202-y