{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,27]],"date-time":"2026-10-27T16:33:04Z","timestamp":1793118784003,"version":"4.5.0"},"reference-count":44,"publisher":"MDPI AG","issue":"8","license":[{"start":{"date-parts":[[2022,7,26]],"date-time":"2022-07-26T00:00:00Z","timestamp":1658793600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"AO Foundation and the Italian Ministry of Health (\u201cRicerca Corrente\u201d)"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Medicina"],"abstract":"<jats:p>Background and Objectives: Commonly being the first step in trauma routine imaging, up to 67% fractures are missed on plain radiographs of the thoracolumbar (TL) spine. The aim of this study was to develop a deep learning model that detects traumatic fractures on sagittal radiographs of the TL spine. Identifying vertebral fractures in simple radiographic projections would have a significant clinical and financial impact, especially for low- and middle-income countries where computed tomography (CT) and magnetic resonance imaging (MRI) are not readily available and could help select patients that need second level imaging, thus improving the cost-effectiveness. Materials and Methods: Imaging studies (radiographs, CT, and\/or MRI) of 151 patients were used. An expert group of three spinal surgeons reviewed all available images to confirm presence and type of fractures. In total, 630 single vertebra images were extracted from the sagittal radiographs of the 151 patients\u2014302 exhibiting a vertebral body fracture, and 328 exhibiting no fracture. Following augmentation, these single vertebra images were used to train, validate, and comparatively test two deep learning convolutional neural network models, namely ResNet18 and VGG16. A heatmap analysis was then conducted to better understand the predictions of each model. Results: ResNet18 demonstrated a better performance, achieving higher sensitivity (91%), specificity (89%), and accuracy (88%) compared to VGG16 (90%, 83%, 86%). In 81% of the cases, the \u201cwarm zone\u201d in the heatmaps correlated with the findings, suggestive of fracture within the vertebral body seen in the imaging studies. Vertebras T12 to L2 were the most frequently involved, accounting for 48% of the fractures. A4, A3, and A1 were the most frequent fracture types according to the AO Spine Classification. Conclusions: ResNet18 could accurately identify the traumatic vertebral fractures on the TL sagittal radiographs. In most cases, the model based its prediction on the same areas that human expert classifiers used to determine the presence of a fracture.<\/jats:p>","DOI":"10.3390\/medicina58080998","type":"journal-article","created":{"date-parts":[[2022,7,27]],"date-time":"2022-07-27T04:59:16Z","timestamp":1658897956000},"page":"998","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":33,"title":["Artificial Intelligence Accurately Detects Traumatic Thoracolumbar Fractures on Sagittal Radiographs"],"prefix":"10.3390","volume":"58","author":[{"given":"Guillermo S\u00e1nchez","family":"Rosenberg","sequence":"first","affiliation":[{"name":"AO Research Institute Davos, 7270 Davos, Switzerland"},{"name":"Department of Orthopedic and Trauma Surgery, University Hospital Basel, 4031 Basel, Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6016-6200","authenticated-orcid":false,"given":"Andrea","family":"Cina","sequence":"additional","affiliation":[{"name":"IRCCS Istituto Ortopedico Galeazzi, 20161 Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Giuseppe Rosario","family":"Schir\u00f3","sequence":"additional","affiliation":[{"name":"ASST GOM Niguarda, 20161 Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3682-1121","authenticated-orcid":false,"given":"Pietro Domenico","family":"Giorgi","sequence":"additional","affiliation":[{"name":"ASST GOM Niguarda, 20161 Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9795-115X","authenticated-orcid":false,"given":"Boyko","family":"Gueorguiev","sequence":"additional","affiliation":[{"name":"AO Research Institute Davos, 7270 Davos, Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0262-1412","authenticated-orcid":false,"given":"Mauro","family":"Alini","sequence":"additional","affiliation":[{"name":"AO Research Institute Davos, 7270 Davos, Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Peter","family":"Varga","sequence":"additional","affiliation":[{"name":"AO Research Institute Davos, 7270 Davos, Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Fabio","family":"Galbusera","sequence":"additional","affiliation":[{"name":"Spine Center, Schulthess Clinic, 8008 Zurich, Switzerland"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9287-9937","authenticated-orcid":false,"given":"Enrico","family":"Gallazzi","sequence":"additional","affiliation":[{"name":"UOC Patologia Vertebrale e Scoliosi, ASST Gaetano Pini-CTO, 20161 Milano, Italy"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2022,7,26]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"166","DOI":"10.1016\/j.injury.2008.06.040","article-title":"Epidemiology of traumatic spine fractures","volume":"40","author":"Leucht","year":"2009","journal-title":"Injury"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"e12190","DOI":"10.1097\/MD.0000000000012190","article-title":"National incidence of traumatic spinal fractures in China","volume":"97","author":"Liu","year":"2018","journal-title":"Medicine"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"866","DOI":"10.1111\/j.1553-2712.2001.tb01146.x","article-title":"Epidemiology of thoracolumbar spine injury in blunt trauma","volume":"8","author":"Holmes","year":"2001","journal-title":"Acad. Emerg. Med."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"383","DOI":"10.1016\/j.jor.2016.06.019","article-title":"The epidemiology of thoracolumbar trauma: A Meta-Analysis","volume":"13","author":"Katsuura","year":"2016","journal-title":"J. Orthop."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"492","DOI":"10.1097\/00007632-199602150-00016","article-title":"Epidemiology of Incident Spinal Fracture in a Complete Population","volume":"21","author":"Hu","year":"1966","journal-title":"Spine"},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"E37","DOI":"10.1097\/BRS.0000000000001177","article-title":"The Influence of Insurance Status on the Surgical Treatment of Acute Spinal Fractures","volume":"41","author":"Daly","year":"2016","journal-title":"Spine"},{"key":"ref_7","doi-asserted-by":"crossref","unstructured":"(2013). Advanced trauma life support (ATLS\u00ae): The ninth edition. J. Trauma Acute Care Surg., 74, 1363\u20131366.","DOI":"10.1097\/TA.0b013e31828b82f5"},{"key":"ref_8","unstructured":"(2016). Spinal Injury Assessment and Initial Management, National Institute for Health and Care Excellence."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"655","DOI":"10.1007\/s10140-019-01712-0","article-title":"Torso computed tomography in blunt trauma patients with normal vital signs can be avoided using non-invasive tests and close clinical evaluation","volume":"26","author":"Reitano","year":"2019","journal-title":"Emerg. Radiol."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"385","DOI":"10.1093\/heapol\/czq076","article-title":"Health technology diffusion in developing countries: A case study of CT scanners in Brazil","volume":"26","author":"Silva","year":"2011","journal-title":"Health Policy Plan"},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"99","DOI":"10.11604\/pamj.2016.25.99.9736","article-title":"Defining the diagnostic divide: An analysis of registered radiological equipment resources in a low-income African country","volume":"25","author":"Ngoya","year":"2016","journal-title":"Pan Afr. Med. J."},{"key":"ref_12","doi-asserted-by":"crossref","first-page":"S38","DOI":"10.1016\/j.afjem.2019.01.004","article-title":"Poor access to acute care resources to treat major trauma in low- and middle-income settings: A self-reported survey of acute care providers","volume":"9","author":"Alibhai","year":"2019","journal-title":"Afr. J. Emerg. Med."},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"1022","DOI":"10.1097\/01.TA.0000124304.68584.2C","article-title":"Computed Tomographic Scanning Reduces Cost and Time of Complete Spine Evaluation","volume":"56","author":"Brandt","year":"2004","journal-title":"J. Trauma Inj. Infect. Crit. Care"},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"153","DOI":"10.1016\/j.jemermed.2018.10.032","article-title":"Blunt Thoracolumbar-Spine Trauma Evaluation in the Emergency Department: A Meta-Analysis of Diagnostic Accuracy for History, Physical Examination, and Imaging","volume":"56","author":"VandenBerg","year":"2019","journal-title":"J. Emerg. Med."},{"key":"ref_15","doi-asserted-by":"crossref","first-page":"681","DOI":"10.1148\/radiol.2273020592","article-title":"Thoracolumbar spine fractures in patients who have sustained severe trauma: Depiction with multi-detector row CT","volume":"227","author":"Wintermark","year":"2003","journal-title":"Radiology"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"390","DOI":"10.1007\/s00586-011-1913-4","article-title":"Impact of magnetic resonance imaging on decision making for thoracolumbar traumatic fracture diagnosis and treatment","volume":"20","author":"Pizones","year":"2011","journal-title":"Eur. Spine J."},{"key":"ref_17","doi-asserted-by":"crossref","first-page":"451","DOI":"10.1097\/01.brs.0000199927.78531.b5","article-title":"Neurologic deterioration secondary to unrecognized spinal instability following trauma-A multicenter study","volume":"31","author":"Levi","year":"2006","journal-title":"Spine"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.actbio.2017.01.036","article-title":"Cell-laden hydrogels for osteochondral and cartilage tissue engineering","volume":"57","author":"Yang","year":"2017","journal-title":"Acta Biomater"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"1374","DOI":"10.1007\/s00586-017-4956-3","article-title":"ISSLS PRIZE IN BIOENGINEERING SCIENCE 2017: Automation of reading of radiological features from magnetic resonance images (MRIs) of the lumbar spine without human intervention is comparable with an expert radiologist","volume":"26","author":"Jamaludin","year":"2017","journal-title":"Eur. Spine J."},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Weng, C.-H., Wang, C.-L., Huang, Y.-J., Yeh, Y.-C., Fu, C.-J., Yeh, C.-Y., and Tsai, T.-T. (2019). Artificial Intelligence for Automatic Measurement of Sagittal Vertical Axis Using ResUNet Framework. J. Clin. Med., 8.","DOI":"10.3390\/jcm8111826"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"1069","DOI":"10.1007\/s11548-020-02173-4","article-title":"A convolutional neural network to detect scoliosis treatment in radiographs","volume":"15","author":"Vergari","year":"2020","journal-title":"Int. J. Comput. Assist. Radiol. Surg."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"694","DOI":"10.1097\/BRS.0000000000003353","article-title":"A Deep Convolutional Neural Network With Performance Comparable to Radiologists for Differentiating Between Spinal Schwannoma and Meningioma","volume":"45","author":"Maki","year":"2020","journal-title":"Spine"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"76","DOI":"10.1016\/j.media.2018.07.008","article-title":"Deep convolutional neural network-based segmentation and classification of difficult to define metastatic spinal lesions in 3D CT data","volume":"49","author":"Chmelik","year":"2018","journal-title":"Med. Image Anal."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"20031","DOI":"10.1038\/s41598-020-76866-w","article-title":"Artificial intelligence for the detection of vertebral fractures on plain spinal radiography","volume":"10","author":"Murata","year":"2020","journal-title":"Sci. Rep."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"84","DOI":"10.1145\/3065386","article-title":"ImageNet Classification with Deep Convolutional Neural Networks","volume":"60","author":"Krizhevsky","year":"2017","journal-title":"Commun. ACM"},{"key":"ref_26","doi-asserted-by":"crossref","first-page":"506","DOI":"10.2214\/AJR.19.21117","article-title":"Artificial Intelligence in Musculoskeletal Imaging: Current Status and Future Directions","volume":"213","author":"Gyftopoulos","year":"2019","journal-title":"AJR Am. J. Roentgenol."},{"key":"ref_27","unstructured":"Simonyan, K., and Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv."},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"He, K., Zhang, X., Ren, S., and Sun, J. (2016, January 27\u201330). Deep residual learning for image recognition. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"e1044","DOI":"10.1002\/jsp2.1044","article-title":"Artificial intelligence and machine learning in spine research","volume":"2","author":"Galbusera","year":"2019","journal-title":"JOR Spine"},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"1285","DOI":"10.1109\/TMI.2016.2528162","article-title":"Deep Convolutional Neural Networks for Computer-Aided Detection: CNN Architectures, Dataset Characteristics and Transfer Learning","volume":"35","author":"Shin","year":"2016","journal-title":"IEEE Trans. Med. Imaging"},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"781","DOI":"10.1007\/978-981-15-5971-6_83","article-title":"A Survey on Transfer Learning","volume":"194","author":"Panigrahi","year":"2021","journal-title":"Smart Innov. Syst. Technol."},{"key":"ref_32","unstructured":"Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., and Antiga, L. (2019). PyTorch: An imperative style, high-performance deep learning library. arXiv."},{"key":"ref_33","doi-asserted-by":"crossref","unstructured":"Sidey-Gibbons, J.A.M., and Sidey-Gibbons, C.J. (2019). Machine learning in medicine: A practical introduction. BMC Med. Res. Methodol., 19.","DOI":"10.1186\/s12874-019-0681-4"},{"key":"ref_34","doi-asserted-by":"crossref","first-page":"581","DOI":"10.1080\/17453674.2017.1344459","article-title":"Artificial intelligence for analyzing orthopedic trauma radiographs: Deep learning algorithms\u2014Are they on par with humans for diagnosing fractures?","volume":"88","author":"Olczak","year":"2017","journal-title":"Acta Orthop."},{"key":"ref_35","doi-asserted-by":"crossref","first-page":"27","DOI":"10.1111\/1754-9485.12828","article-title":"Computer vs. human: Deep learning versus perceptual training for the detection of neck of femur fractures","volume":"63","author":"Adams","year":"2019","journal-title":"J. Med. Imaging Radiat. Oncol."},{"key":"ref_36","doi-asserted-by":"crossref","first-page":"468","DOI":"10.1080\/17453674.2018.1453714","article-title":"Automated detection and classification of the proximal humerus fracture by using deep learning algorithm","volume":"89","author":"Chung","year":"2018","journal-title":"Acta Orthop."},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"439","DOI":"10.1016\/j.crad.2017.11.015","article-title":"Artificial intelligence in fracture detection: Transfer learning from deep convolutional neural networks","volume":"73","author":"Kim","year":"2018","journal-title":"Clin. Radiol."},{"key":"ref_38","doi-asserted-by":"crossref","first-page":"1454","DOI":"10.1007\/s00586-016-4687-x","article-title":"The radiologic assessment of posterior ligamentous complex injury in patients with thoracolumbar fracture","volume":"26","author":"Chen","year":"2017","journal-title":"Eur. Spine J."},{"key":"ref_39","doi-asserted-by":"crossref","first-page":"713.e17","DOI":"10.1016\/j.crad.2020.05.021","article-title":"Diagnostic accuracy of deep learning in orthopaedic fractures: A systematic review and meta-analysis","volume":"75","author":"Yang","year":"2020","journal-title":"Clin. Radiol."},{"key":"ref_40","doi-asserted-by":"crossref","first-page":"1463","DOI":"10.1007\/s00586-016-4623-0","article-title":"The value of CT and MRI in the classification and surgical decision-making among spine surgeons in thoracolumbar spinal injuries","volume":"26","author":"Rajasekaran","year":"2017","journal-title":"Eur. Spine J."},{"key":"ref_41","doi-asserted-by":"crossref","first-page":"S326","DOI":"10.1097\/TA.0b013e31827559b8","article-title":"Screening for thoracolumbar spinal injuries in blunt trauma: An eastern association for the surgery of trauma practice management guideline","volume":"73","author":"Sixta","year":"2012","journal-title":"J. Trauma Acute Care Surg."},{"key":"ref_42","doi-asserted-by":"crossref","unstructured":"Balogh, E.P., Miller, B.T., and Ball, J.R. (2016). Improving Diagnosis in Health Care, National Academies Press.","DOI":"10.17226\/21794"},{"key":"ref_43","doi-asserted-by":"crossref","first-page":"246","DOI":"10.1016\/j.ejrad.2018.06.020","article-title":"The past, present and future role of artificial intelligence in imaging","volume":"105","author":"Fazal","year":"2018","journal-title":"Eur. J. Radiol."},{"key":"ref_44","doi-asserted-by":"crossref","first-page":"11591","DOI":"10.1073\/pnas.1806905115","article-title":"Deep neural network improves fracture detection by clinicians","volume":"115","author":"Lindsey","year":"2018","journal-title":"Proc. Natl. Acad. Sci. USA"}],"container-title":["Medicina"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1648-9144\/58\/8\/998\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,10]],"date-time":"2025-10-10T23:57:05Z","timestamp":1760140625000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1648-9144\/58\/8\/998"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,7,26]]},"references-count":44,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2022,8]]}},"alternative-id":["medicina58080998"],"URL":"https:\/\/doi.org\/10.3390\/medicina58080998","relation":{},"ISSN":["1648-9144"],"issn-type":[{"value":"1648-9144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,7,26]]}}}