Top

Published in:

Open Access 01-10-2019 | Imaging Informatics and Artificial Intelligence

Application of a deep learning algorithm for detection and visualization of hip fractures on plain pelvic radiographs

Authors: Chi-Tung Cheng, Tsung-Ying Ho, Tao-Yi Lee, Chih-Chen Chang, Ching-Cheng Chou, Chih-Chi Chen, I-Fang Chung, Chien-Hung Liao

Published in: European Radiology | Issue 10/2019

Abstract

Objective

To identify the feasibility of using a deep convolutional neural network (DCNN) for the detection and localization of hip fractures on plain frontal pelvic radiographs (PXRs).

Summary of background data

Hip fracture is a leading worldwide health problem for the elderly. A missed diagnosis of hip fracture on radiography leads to a dismal prognosis. The application of a DCNN to PXRs can potentially improve the accuracy and efficiency of hip fracture diagnosis.

Methods

A DCNN was pretrained using 25,505 limb radiographs between January 2012 and December 2017. It was retrained using 3605 PXRs between August 2008 and December 2016. The accuracy, sensitivity, false-negative rate, and area under the receiver operating characteristic curve (AUC) were evaluated on 100 independent PXRs acquired during 2017. The authors also used the visualization algorithm gradient-weighted class activation mapping (Grad-CAM) to confirm the validity of the model.

Results

The algorithm achieved an accuracy of 91%, a sensitivity of 98%, a false-negative rate of 2%, and an AUC of 0.98 for identifying hip fractures. The visualization algorithm showed an accuracy of 95.9% for lesion identification.

Conclusions

A DCNN not only detected hip fractures on PXRs with a low false-negative rate but also had high accuracy for localizing fracture lesions. The DCNN might be an efficient and economical model to help clinicians make a diagnosis without interrupting the current clinical pathway.

Key Points

• Automated detection of hip fractures on frontal pelvic radiographs may facilitate emergent screening and evaluation efforts for primary physicians.

• Good visualization of the fracture site by Grad-CAM enables the rapid integration of this tool into the current medical system.

• The feasibility and efficiency of utilizing a deep neural network have been confirmed for the screening of hip fractures.

Brauer CA, Coca-Perraillon M, Cutler DM, Rosen AB (2009) Incidence and mortality of hip fractures in the United States. JAMA 302:1573–1579. https://doi.org/10.1001/jama.2009.1462 CrossRefPubMedPubMedCentral

Riggs BL, Melton LJ III (1995) The worldwide problem of osteoporosis: insights afforded by epidemiology. Bone 17:S505–S511. https://doi.org/10.1016/8756-3282(95)00258-4 CrossRef

Leslie WD, O’Donnell S, Jean S et al (2009) Trends in hip fracture rates in Canada. JAMA 302:883–889. https://doi.org/10.1001/jama.2009.1231 CrossRefPubMed

Bliuc D, Nguyen ND, Milch VE et al (2009) Mortality risk associated with low-trauma osteoporotic fracture and subsequent fracture in men and women. JAMA 301:513–521. https://doi.org/10.1001/jama.2009.50 CrossRefPubMed

Center JR, Nguyen TV, Schneider D et al (1999) Mortality after all major types of osteoporotic fracture in men and women: an observational study. Lancet 353:878–882. https://doi.org/10.1016/S0140-6736(98)09075-8 CrossRefPubMed

Auron-Gomez M, Michota F (2008) Medical management of hip fracture. Clin Geriatr Med 24:701–19– ix. https://doi.org/10.1016/j.cger.2008.07.002 CrossRefPubMed

Marks R (2010) Hip fracture epidemiological trends, outcomes, and risk factors, 1970-2009. Int J Gen Med 3:1–17PubMedPubMedCentral

Michael Lewiecki E, Wright NC, Curtis JR et al (2018) Hip fracture trends in the United States, 2002 to 2015. Osteoporos Int 29:717–722. https://doi.org/10.1007/s00198-017-4345-0 CrossRefPubMed

Melton LJ, Gabriel SE, Crowson CS et al (2003) Cost-equivalence of different osteoporotic fractures. Osteoporos Int 14:383–388. https://doi.org/10.1007/s00198-003-1385-4 CrossRefPubMed

10.

Johnell O, Kanis J (2005) Epidemiology of osteoporotic fractures. Osteoporos Int 16(Suppl 2):S3–S7. https://doi.org/10.1007/s00198-004-1702-6 CrossRefPubMed

11.

Johnell O, Kanis JA (2006) An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int 17:1726–1733. https://doi.org/10.1007/s00198-006-0172-4 CrossRefPubMed

12.

Boonen S, Autier P, Barette M et al (2004) Functional outcome and quality of life following hip fracture in elderly women: a prospective controlled study. Osteoporos Int 15:87–94. https://doi.org/10.1007/s00198-003-1515-z CrossRefPubMed

13.

Grimes JP, Gregory PM, Noveck H et al (2002) The effects of time-to-surgery on mortality and morbidity in patients following hip fracture. Am J Med 112:702–709. https://doi.org/10.1016/S0002-9343(02)01119-1 CrossRefPubMed

14.

Hakkarinen DK, Banh KV, Hendey GW (2012) Magnetic resonance imaging identifies occult hip fractures missed by 64-slice computed tomography. J Emerg Med 43:303–307. https://doi.org/10.1016/j.jemermed.2012.01.037 CrossRefPubMed

15.

Chellam WB (2016) Missed subtle fractures on the trauma-meeting digital projector. Injury 47:674–676. https://doi.org/10.1016/j.injury.2015.11.001 CrossRefPubMed

16.

Tarrant SM, Hardy BM, Byth PL et al (2014) Preventable mortality in geriatric hip fracture inpatients. Bone Joint J 96-B:1178–1184. https://doi.org/10.1302/0301-620X.96B9.32814 CrossRefPubMedPubMedCentral

17.

Rehman H, Clement RGE, Perks F, White TO (2016) Imaging of occult hip fractures: CT or MRI? Injury 47:1297–1301. https://doi.org/10.1016/j.injury.2016.02.020 CrossRefPubMed

18.

Lauro GR, Cable W, Lesniak A et al (2013) Digital pathology consultations—a new era in digital imaging, challenges and practical applications. J Digit Imaging 26:668–677. https://doi.org/10.1007/s10278-013-9572-0 CrossRef

19.

Petrick N, Sahiner B, Armato SG et al (2013) Evaluation of computer-aided detection and diagnosis systems. Med Phys 40:087001. https://doi.org/10.1118/1.4816310 CrossRefPubMedPubMedCentral

20.

Kermany DS, Goldbaum M, Cai W et al (2018) Identifying medical diagnoses and treatable diseases by image-based deep learning. Cell 172:1122–1131.e9. https://doi.org/10.1016/j.cell.2018.02.010 CrossRefPubMed

21.

Prevedello LM, Erdal BS, Ryu JL et al (2017) Automated critical test findings identification and online notification system using artificial intelligence in imaging. Radiology 285:923–931. https://doi.org/10.1148/radiol.2017162664 CrossRefPubMed

22.

Stone JR, Wilde EA, Taylor BA et al (2016) Supervised learning technique for the automated identification of white matter hyperintensities in traumatic brain injury. Brain Inj 30:1458–1468. https://doi.org/10.1080/02699052.2016.1222080 CrossRefPubMed

23.

Arif Al SMMR, Knapp K, Slabaugh G (2018) Fully automatic cervical vertebrae segmentation framework for X-ray images. Comput Methods Prog Biomed 157:95–111. https://doi.org/10.1016/j.cmpb.2018.01.006 CrossRef

24.

Ho-Le TP, Center JR, Eisman JA et al (2017) Prediction of hip fracture in post-menopausal women using artificial neural network approach. Conf Proc IEEE Eng Med Biol Soc 2017:4207–4210. https://doi.org/10.1109/EMBC.2017.8037784 CrossRef

25.

Chung SW, Han SS, Lee JW et al (2018) Automated detection and classification of the proximal humerus fracture by using deep learning algorithm. Acta Orthop 89:468–473. https://doi.org/10.1080/17453674.2018.1453714 CrossRefPubMedPubMedCentral

26.

Urakawa T, Tanaka Y, Goto S et al (2018) Detecting intertrochanteric hip fractures with orthopedist-level accuracy using a deep convolutional neural network. Skeletal Radiol 41:1–6. https://doi.org/10.1007/s00256-018-3016-3 CrossRef

27.

Olczak J, Fahlberg N, Maki A et al (2017) Artificial intelligence for analyzing orthopedic trauma radiographs. Acta Orthop 88:581–586. https://doi.org/10.1080/17453674.2017.1344459 CrossRefPubMedPubMedCentral

28.

Kim DH, MacKinnon T (2018) Artificial intelligence in fracture detection: transfer learning from deep convolutional neural networks. Clin Radiol 73:439–445. https://doi.org/10.1016/j.crad.2017.11.015 CrossRefPubMed

29.

Gale W, Oakden-Rayner L, Carneiro G, et al (2017) Detecting hip fractures with radiologist-level performance using deep neural networks. arXiv:1711.06504

30.

Selvaraju RR, Cogswell M, Das A, et al (2017) Grad-CAM: visual explanations from deep networks via gradient-based localization. ICCV 618–626. arXiv:1610.02391

31.

LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444. https://doi.org/10.1038/nature14539 CrossRefPubMed

32.

Huang G, Liu Z, Van Der Maaten L, Weinberger KQ (2018) Densely connected convolutional networks. arXiv:1608.06993

33.

Gulshan V, Peng L, Coram M et al (2016) Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA 316:2402–2410. https://doi.org/10.1001/jama.2016.17216 CrossRefPubMed

34.

Esteva A, Kuprel B, Novoa RA et al (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542:115–118. https://doi.org/10.1038/nature21056 CrossRefPubMedPubMedCentral

35.

Bottle A, Aylin P (2006) Mortality associated with delay in operation after hip fracture: observational study. BMJ 332:947–951. https://doi.org/10.1136/bmj.38790.468519.55 CrossRefPubMedPubMedCentral

36.

Vidán MT, Sánchez E, Gracia Y et al (2011) Causes and effects of surgical delay in patients with hip fracture: a cohort study. Ann Intern Med 155:226–233. https://doi.org/10.7326/0003-4819-155-4-201108160-00006 CrossRefPubMed

37.

Shiga T, Wajima Z, Ohe Y (2008) Is operative delay associated with increased mortality of hip fracture patients? Systematic review, meta-analysis, and meta-regression. Can J Anaesth 55:146–154. https://doi.org/10.1007/BF03016088 CrossRefPubMed

38.

Gdalevich M, Cohen D, Yosef D, Tauber C (2004) Morbidity and mortality after hip fracture: the impact of operative delay. Arch Orthop Trauma Surg 124:334–340. https://doi.org/10.1007/s00402-004-0662-9 CrossRefPubMed

39.

Erhan D, Bengio Y, Courville A, et al (2009) Visualizing higher-layer features of a deep network. Available via https://pdfs.semanticscholar.org/65d9/94fb778a8d9e0f632659fb33a082949a50d3.pdf. Accessed on 28 Feb 2019

Title: Application of a deep learning algorithm for detection and visualization of hip fractures on plain pelvic radiographs
Authors: Chi-Tung Cheng
Tsung-Ying Ho
Tao-Yi Lee
Chih-Chen Chang
Ching-Cheng Chou
Chih-Chi Chen
I-Fang Chung
Chien-Hung Liao
Publication date: 01-10-2019
Publisher: Springer Berlin Heidelberg
Published in: European Radiology / Issue 10/2019
Print ISSN: 0938-7994
Electronic ISSN: 1432-1084
DOI: https://doi.org/10.1007/s00330-019-06167-y

Keynote webinar | Spotlight on medication adherence

Springer Medicine

Application of a deep learning algorithm for detection and visualization of hip fractures on plain pelvic radiographs

Abstract

Objective

Summary of background data

Methods

Results

Conclusions

Key Points

Keynote webinar | Spotlight on medication adherence

Springer Medicine

Abstract

Objective

Summary of background data

Methods

Results

Conclusions

Key Points

Please log in to get access to this content

Other articles of this Issue 10/2019

Application of deep learning–based computer-aided detection system: detecting pneumothorax on chest radiograph after biopsy

X-ray examination dose surveys: how accurate are my results?

Automatic classification of ultrasound breast lesions using a deep convolutional neural network mimicking human decision-making

Detection of liver metastases on gadobenate dimeglumine-enhanced MRI: systematic review, meta-analysis, and similarities with gadoxetate-enhanced MRI

Cervical necrotic lymphadenopathy: a diagnostic tree analysis model based on CT and clinical findings

Acetabular coverage differs between standing and supine positions: model-based assessment of low-dose biplanar radiographs and comparison with CT