The Journal of Hip Surgery 2024; 08(02): 104-112
DOI: 10.1055/s-0044-1800903
Original Article

Deep Learning in the Detection of Rare Fractures – Development of a “Deep Learning Convolutional Network” Model for Detecting Acetabular Fractures

Deep Learning in der Erkennung seltener Frakturen – Entwicklung eines „Deep Convolutional Neural Network“ (DCNN) am Beispiel der Azetabulumfraktur
Felix Erne*
1   Department of Trauma and Reconstructive Surgery, Occupational Accident Clinic Tübingen, Tübingen, Germany
,
Daniel Dehncke*
2   Department of Informatics, Methods in Medical Informatics, Eberhard Karls University of Tübingen Faculty of Mathematics and Natural Sciences, Tubingen, Germany
,
Steven C. Herath
1   Department of Trauma and Reconstructive Surgery, Occupational Accident Clinic Tübingen, Tübingen, Germany
,
Fabian Springer
3   Department of Diagnostic & Interventional Radiology, University Hospital Tübingen, Tübingen, Germany
4   Department of Radiology, Occupational Accident Clinic Tübingen, Tübingen, Germany
,
Nico Pfeifer
2   Department of Informatics, Methods in Medical Informatics, Eberhard Karls University of Tübingen Faculty of Mathematics and Natural Sciences, Tubingen, Germany
,
Ralf Eggeling**
2   Department of Informatics, Methods in Medical Informatics, Eberhard Karls University of Tübingen Faculty of Mathematics and Natural Sciences, Tubingen, Germany
,
1   Department of Trauma and Reconstructive Surgery, Occupational Accident Clinic Tübingen, Tübingen, Germany
› Institutsangaben

Abstract

Background Fracture detection by artificial intelligence and especially Deep Convolutional Neural Networks (DCNN) is a topic of growing interest in current orthopaedic and radiological research. As learning a DCNN usually needs a large amount of training data, mostly frequent fractures as well as conventional X-ray are used. Therefore, less common fractures like acetabular fractures (AF) are underrepresented in the literature. The aim of this pilot study was to establish a DCNN for detection of AF using computer tomography (CT) scans.

Methods Patients with an acetabular fracture were identified from the monocentric consecutive pelvic injury registry at the BG Trauma Center XXX from 01/2003–12/2019. All patients with unilateral AF and CT scans available in DICOM-format were included for further processing. All datasets were automatically anonymised and digitally post-processed. Extraction of the relevant region of interests was performed and the technique of data augmentation (DA) was implemented to artificially increase the number of training samples. A DCNN based on Med3D was used for autonomous fracture detection, using global average pooling (GAP) to reduce overfitting.

Results From a total of 2,340 patients with a pelvic fracture, 654 patients suffered from an AF. After screening and post-processing of the datasets, a total of 159 datasets were enrolled for training of the algorithm. A random assignment into training datasets (80%) and test datasets (20%) was performed. The technique of bone area extraction, DA and GAP increased the accuracy of fracture detection from 58.8% (native DCNN) up to an accuracy of 82.8% despite the low number of datasets.

Conclusion The accuracy of fracture detection of our trained DCNN is comparable to published values despite the low number of training datasets. The techniques of bone extraction, DA and GAP are useful for increasing the detection rates of rare fractures by a DCNN. Based on the used DCNN in combination with the described techniques from this pilot study, the possibility of an automatic fracture classification of AF is under investigation in a multicentre study.

Zusammenfassung

Hintergrund Die automatische Frakturerkennung mithilfe von künstlicher Intelligenz (KI) ist derzeit ein wichtiges Thema in der unfallchirurgischen und radiologischen Forschung. Vor allem bei der Nutzung von „Deep Convolutional Neural Networks“ (DCNN) ist allerdings üblicherweise eine große Anzahl an Trainingsdatensätzen erforderlich, um gute Erkennungsraten zu erzielen, weswegen diese Methoden bisher i.d.R. an sehr häufigen Frakturen und an Röntgenbildern angewendet werden. Demgegenüber stellen seltene Frakturen wie die Azetabulumfraktur (AF) eine Herausforderung für neuronale Netzwerke dar. Ziel dieser Pilotstudie war es daher, einen Algorithmus zu entwickeln, mit dessen Hilfe AF anhand von CT-Bildern zuverlässig automatisch erkannt werden können.

Methodik Patienten mit einer einseitigen AF wurden aus dem monozentrischen Beckenregister der BG Unfallklinik Tübingen über einen Zeitraum von 01/2003 bis 12/2019 identifiziert. Patienten, bei denen CT-Datensätze im DICOM-Format vorlagen, wurden weiterverarbeitet. Die automatische Weiterverarbeitung beinhaltete die Anonymisierung sowie verschiedene Methoden der Knochenextraktion, der „Data Augmentation“ (DA) sowie des Global Average Poolings (GAP), um die Datensätze für das Training des auf Basis von Med3D erstellten DCNN vorzubereiten.

Ergebnisse Von insgesamt 2340 Patienten mit einer Beckenfraktur erlitten 654 eine AF. Nach Weiterverarbeitung der Datensätze konnten 159 Datensätze für das Training verwendet werden. 80% davon wurden als Trainingsdatensätze randomisiert, 20% wurden als Testdatensätze randomisiert. Durch die Techniken der DA und des GAP konnte die Erkennungsrate des DCNN trotz der geringen Zahl an Trainingsdatensätzen von 58,8 auf 82,8% gesteigert werden.

Schlussfolgerung Die Erkennungsrate unseres DCNN ist trotz der im Vergleich niedrigen Zahl an Trainingsdatensätzen vergleichbar mit Literaturangaben bei häufigeren Frakturen. Die Techniken der Knochenextraktion, verbunden mit der DA und der GAP, helfen bei selten auftretenden Frakturen, die Erkennungsraten eines DCNN zu erhöhen. Mit den Techniken und dem DCNN aus dieser Pilotstudie soll in einer Multicenterstudie an einer größeren Anzahl an Datensätzen die Möglichkeit einer automatischen Klassifizierung von AF untersucht werden.

* shared authorship


** shared last authorship


* This article is a republished version of: Erne F, Dehncke D, Herath SC, Springer F, Pfeifer N, Eggeling R, Küper MA. Deep Learning in the Detection of Rare Fractures - Development of a “Deep Learning Convolutional Network” Model for Detecting Acetabular Fractures. Z Orthop Unfall 2023;161(1):42–50.




Publikationsverlauf

Artikel online veröffentlicht:
09. Januar 2025

© 2024. Thieme. All rights reserved.

Thieme Medical Publishers, Inc.
333 Seventh Avenue, 18th Floor, New York, NY 10001, USA

 
  • References

  • 1 Herath SC, Pott H, Rollmann MFR. et al. Geriatric Acetabular Surgery: Letournel's Contraindications Then and Now-Data From the German Pelvic Registry. J Orthop Trauma 2019; 33 (Suppl. 02) S8-S13
  • 2 Rollmann MF, Herath SC, Kirchhoff F. et al. Pelvic ring fractures in the elderly now and then – a pelvic registry study. Arch Gerontol Geriatr 2017; 71: 83-88
  • 3 Melhem E, Riouallon G, Habboubi K. et al. Epidemiology of pelvic and acetabular fractures in France. Orthop Traumatol Surg Res 2020; 106: 831-839
  • 4 Best MJ, Buller LT, Quinnan SM. Analysis of Incidence and Outcome Predictors for Patients Admitted to US Hospitals with Acetabular Fractures from 1990 to 2010. Am J Orthop (Belle Mead NJ) 2018;
  • 5 Andrich S, Haastert B, Neuhaus E. et al. Epidemiology of Pelvic Fractures in Germany: Considerably High Incidence Rates among Older People. PLoS One 2015; 10: e0139078
  • 6 Mandell JC, Weaver MJ, Khurana B. Computed tomography for occult fractures of the proximal femur, pelvis, and sacrum in clinical practice: single institution, dual-site experience. Emerg Radiol 2018; 25: 265-273
  • 7 Cabarrus MC, Ambekar A, Lu Y. et al. MRI and CT of insufficiency fractures of the pelvis and the proximal femur. AJR Am J Roentgenol 2008; 191: 995-1001
  • 8 Goodfellow I, Bengio Y, Courville A. Deep learning. An MIT Press Book. Im Internet (Stand: 26.03.2021): http://www.deeplearningbook.org
  • 9 Chen S, Ma K, Zheng Y. Med3D: Transfer Learning for 3D Medical Image Analysis. arXiv 2019 1904.00625v4 [cs.CV]. Im Internet (Stand: 29.03.2021): https://arxiv.org/abs/1904.00625
  • 10 Bengio Y, Courville A, Vincent P. Representation learning: a review and new perspectives. IEEE Trans Pattern Anal Mach Intell 2013; 35: 1798-1828
  • 11 Burrell J. How the machine ‘thinks’: Understanding opacity in machine learning algorithms. Big Data & Society 2016;
  • 12 Rowe M. An Introduction to Machine Learning for Clinicians. Acad Med 2019; 94: 1433-1436
  • 13 Cha MJ, Chung MJ, Lee JH. et al. Performance of Deep Learning Model in Detecting Operable Lung Cancer With Chest Radiographs. J Thorac Imaging 2019; 34: 86-91
  • 14 Huang Z, Liu D, Chen X. et al. Deep Convolutional Neural Network Based on Computed Tomography Images for the Preoperative Diagnosis of Occult Peritoneal Metastasis in Advanced Gastric Cancer. Front Oncol 2020; 10: 601869
  • 15 Tobler P, Cyriac J, Kovacs BK. et al. AI-based detection and classification of distal radius fractures using low-effort data labeling: evaluation of applicability and effect of training set size. Eur Radiol 2021;
  • 16 Kalmet PHS, Sanduleanu S, Primakov S. et al. Deep learning in fracture detection: a narrative review. Acta Orthop 2020; 91: 215-220
  • 17 Tomita N, Cheung YY, Hassanpour S. Deep neural networks for automatic detection of osteoporotic vertebral fractures on CT scans. Comput Biol Med 2018; 98: 8-15
  • 18 He K, Zhang X, Ren S. et al. Identity Mappings in Deep Residual Networks. arXiv 2016 1603.05027v3 [cs.CV]. Im Internet (Stand: 29.03.2021): https://arxiv.org/abs/1603.05027
  • 19 Srivastava N, Hinton GE, Krizhevsky A. et al. Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 2014; 15: 1929-1958
  • 20 Lin M, Chen Q, Yan S. Network In Network. arXiv 2013 1312.4400v3 [cs.NE]. Im Internet (Stand: 29.03.2021): https://arxiv.org/abs/1312.4400
  • 21 Shen J, Zhang CJP, Jiang B. et al. Artificial Intelligence Versus Clinicians in Disease Diagnosis: Systematic Review. JMIR Med Inform 2019; 7: e10010
  • 22 Olczak J, Fahlberg N, Maki A. et al. Artificial intelligence for analyzing orthopedic trauma radiographs. Acta Orthop 2017; 88: 581-586
  • 23 Gan K, Xu D, Lin Y. et al. Artificial intelligence detection of distal radius fractures: a comparison between the convolutional neural network and professional assessments. Acta Orthop 2019; 90: 394-400
  • 24 Lindsey R, Daluiski A, Chopra S. et al. Deep neural network improves fracture detection by clinicians. Proc Natl Acad Sci U S A 2018; 115: 11591-11596
  • 25 Langerhuizen DWG, Janssen SJ, Mallee WH. et al. What are the applications and limitations of artificial intelligence for fracture detection and classification in orthopaedic trauma imaging? A systematic review. Clin Orthop Relat Res 2019; 477: 2482-2491
  • 26 Pinto A, Berritto D, Russo A. et al. Traumatic fractures in adults: missed diagnosis on plain radiographs in the Emergency Department. Acta Biomed 2018; 89: 111-123
  • 27 Goddard P, Leslie A, Jones A. et al. Error in radiology. Br J Radiol 2001; 74: 949-951
  • 28 Pohlemann T, Tosounidis G, Bircher M. et al. The German Multicentre Pelvis Registry: a template for an European Expert Network?. Injury 2007; 38: 416-423
  • 29 Riouallon G, Sebaaly A, Upex P. et al. A New, Easy, Fast, and Reliable Method to Correctly Classify Acetabular Fractures According to the Letournel System. JB JS Open Access 2018; 3: e0032
  • 30 Ohashi K, El-Khoury GY, Abu-Zahra KW. et al. Interobserver agreement for Letournel acetabular fracture classification with multidetector CT: are standard Judet radiographs necessary?. Radiology 2006; 241: 386-391