{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,16]],"date-time":"2026-10-16T13:12:14Z","timestamp":1792156334126,"version":"4.4.0"},"reference-count":37,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JMIR Med Inform"],"abstract":"Abstract<\/jats:title>\n \n Background<\/jats:title>\n Osteoporotic vertebral fractures (OVFs) are common in older adults and often lead to disability if not properly diagnosed and classified. With the increased use of computed tomography (CT) imaging and the development of radiomics and deep learning technologies, there is potential to improve the classification accuracy of OVFs.<\/jats:p>\n <\/jats:sec>\n \n Objective<\/jats:title>\n This study aims to evaluate the efficacy of a deep learning radiomics model, derived from CT imaging, in accurately classifying OVFs.<\/jats:p>\n <\/jats:sec>\n \n Methods<\/jats:title>\n The study analyzed 981 patients (aged 50\u201095 years; 687 women, 294 men), involving 1098 vertebrae, from 3 medical centers who underwent both CT and magnetic resonance imaging examinations. The Assessment System of Thoracolumbar Osteoporotic Fractures (ASTLOF) classified OVFs into Classes 0, 1, and 2. The data were categorized into 4 cohorts: training (n=750), internal validation (n=187), external validation (n=110), and prospective validation (n=51). Deep transfer learning used the ResNet-50 architecture, pretrained on RadImageNet and ImageNet, to extract imaging features. Deep transfer learning\u2013based features were combined with radiomics features and refined using Least Absolute Shrinkage and Selection Operator (LASSO) regression. The performance of 8 machine learning classifiers for OVF classification was assessed using receiver operating characteristic metrics and the \u201cOne-vs-Rest\u201d approach. Performance comparisons between RadImageNet- and ImageNet-based models were performed using the DeLong test. Shapley Additive Explanations (SHAP) analysis was used to interpret feature importance and the predictive rationale of the optimal fusion model.<\/jats:p>\n <\/jats:sec>\n \n Results<\/jats:title>\n \n Feature selection and fusion yielded 33 and 54 fused features for the RadImageNet- and ImageNet-based models, respectively, following pretraining on the training set. The best-performing machine learning algorithms for these 2 deep learning radiomics models were the multilayer perceptron and Light Gradient Boosting Machine (LightGBM). The macro-average area under the curve (AUC) values for the fused models based on RadImageNet and ImageNet were 0.934 and 0.996, respectively, with DeLong test showing no statistically significant difference (\n P<\/jats:italic>\n =2.34). The RadImageNet-based model significantly surpassed the ImageNet-based model across internal, external, and prospective validation sets, with macro-average AUCs of 0.837 versus 0.648, 0.773 versus 0.633, and 0.852 versus 0.648, respectively (\n P<\/jats:italic>\n <.05). Using the binary \u201cOne-vs-Rest\u201d approach, the RadImageNet-based fused model achieved superior predictive performance for Class 2 (AUC=0.907, 95% CI 0.805\u20100.999), with Classes 0 and 1 following (AUC\/accuracy=0.829\/0.803 and 0.794\/0.768, respectively). SHAP analysis provided a visualization of feature importance in the RadImageNet-based fused model, highlighting the top 3 most influential features: cluster shade, mean, and large area low gray level emphasis, and their respective impacts on predictions.\n <\/jats:p>\n <\/jats:sec>\n \n Conclusions<\/jats:title>\n The RadImageNet-based fused model using CT imaging data exhibited superior predictive performance compared to the ImageNet-based model, demonstrating significant utility in OVF classification and aiding clinical decision-making for treatment planning. Among the 3 classes, the model performed best in identifying Class 2, followed by Class 0 and Class 1.<\/jats:p>\n <\/jats:sec>","DOI":"10.2196\/75665","type":"journal-article","created":{"date-parts":[[2025,8,29]],"date-time":"2025-08-29T22:20:14Z","timestamp":1756506014000},"page":"e75665-e75665","source":"Crossref","is-referenced-by-count":6,"title":["Deep Learning Radiomics Model Based on Computed Tomography Image for Predicting the Classification of Osteoporotic Vertebral Fractures: Algorithm Development and Validation"],"prefix":"10.2196","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-4148-2348","authenticated-orcid":false,"given":"Jiayi","family":"Liu","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-9626-4809","authenticated-orcid":false,"given":"Lincen","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3352-9640","authenticated-orcid":false,"given":"Yousheng","family":"Yuan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0001-0849-4290","authenticated-orcid":false,"given":"Jun","family":"Tang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-0994-2443","authenticated-orcid":false,"given":"Yongkang","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0456-1943","authenticated-orcid":false,"given":"Liang","family":"Xia","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6666-8703","authenticated-orcid":false,"given":"Jun","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2025,8,29]]},"reference":[{"issue":"10","key":"R1","doi-asserted-by":"publisher","first-page":"1789","DOI":"10.1002\/jbmr.3757","article-title":"The prevalence of osteoporosis in China, a nationwide, multicenter DXA survey","volume":"34","author":"Zeng","journal-title":"J Bone Miner Res"},{"issue":"11","key":"R2","doi-asserted-by":"publisher","first-page":"e1175","DOI":"10.1210\/clinem\/dgad256","article-title":"New insights in the pathophysiology, epidemiology, and response to treatment of osteoporotic vertebral fractures","volume":"108","author":"Skj\u00f8dt","journal-title":"J Clin Endocrinol Metab"},{"issue":"2 Suppl","key":"R3","doi-asserted-by":"publisher","first-page":"46S","DOI":"10.1177\/2192568217717972","article-title":"Classification of osteoporotic thoracolumbar spine fractures: recommendations of the spine section of the German Society for Orthopaedics and Trauma (DGOU)","volume":"8","author":"Schnake","journal-title":"Global Spine J"},{"issue":"12","key":"R4","doi-asserted-by":"publisher","first-page":"1819","DOI":"10.1016\/j.acra.2022.02.020","article-title":"Deep learning classification of spinal osteoporotic compression fractures on radiographs using an adaptation of the genant semiquantitative criteria","volume":"29","author":"Dong","journal-title":"Acad Radiol"},{"issue":"S02","key":"R5","doi-asserted-by":"publisher","first-page":"S85","DOI":"10.1007\/s00198-004-1723-1","article-title":"The current treatment--a survey of osteoporotic fracture treatment. Osteoporotic spine fractures: the spine surgeon\u2019s perspective","volume":"16 Suppl 2","author":"Heini","journal-title":"Osteoporos Int"},{"issue":"9","key":"R6","doi-asserted-by":"publisher","first-page":"2431","DOI":"10.1007\/s00586-022-07201-2","article-title":"Inter- and intraobserver reliabilities and critical analysis of the osteoporotic fracture classification of osteoporotic vertebral body fractures","volume":"31","author":"Sch\u00f6nrogge","journal-title":"Eur Spine J"},{"issue":"1","key":"R7","doi-asserted-by":"publisher","DOI":"10.1038\/s41598-018-26562-7","article-title":"The analysis of MSTMOVCF (Multi-segment thoracolumbar mild osteoporotic fractures surgery or conservative treatment) based on ASTLOF (the assessment system of thoracolumbar osteoporotic fracture)","volume":"8","author":"Du","journal-title":"Sci Rep"},{"issue":"1","key":"R8","doi-asserted-by":"publisher","DOI":"10.1186\/s13018-020-01882-5","article-title":"Reliability and application of the new morphological classification system for chronic symptomatic osteoporotic thoracolumbar fracture","volume":"15","author":"Hao","journal-title":"J Orthop Surg Res"},{"key":"R9","doi-asserted-by":"publisher","DOI":"10.1016\/j.wneu.2018.03.122","article-title":"Surgery for multisegment thoracolumbar mild osteoporotic fractures: revised assessment system of thoracolumbar osteoporotic fracture","volume":"114","author":"Du","journal-title":"World Neurosurg"},{"issue":"8","key":"R10","doi-asserted-by":"publisher","first-page":"998","DOI":"10.3390\/medicina58080998","article-title":"Artificial intelligence accurately detects traumatic thoracolumbar fractures on sagittal radiographs","volume":"58","author":"Rosenberg","journal-title":"Med Bogota Colomb"},{"issue":"1","key":"R11","doi-asserted-by":"publisher","DOI":"10.1186\/s12939-018-0869-y","article-title":"Equity assessment of the distribution of CT and MRI scanners in China: a panel data analysis","volume":"17","author":"He","journal-title":"Int J Equity Health"},{"issue":"5","key":"R12","doi-asserted-by":"publisher","first-page":"2207","DOI":"10.1007\/s00330-018-5846-8","article-title":"Vertebral body insufficiency fractures: detection of vertebrae at risk on standard CT images using texture analysis and machine learning","volume":"29","author":"Muehlematter","journal-title":"Eur Radiol"},{"issue":"1","key":"R13","doi-asserted-by":"publisher","DOI":"10.1186\/s12859-022-04596-z","article-title":"A hierarchical opportunistic screening model for osteoporosis using machine learning applied to clinical data and CT images","volume":"23","author":"Liu","journal-title":"BMC Bioinformatics"},{"issue":"1","key":"R14","doi-asserted-by":"publisher","first-page":"165","DOI":"10.1186\/s12891-023-06281-5","article-title":"Differentiation of acute and chronic vertebral compression fractures using conventional CT based on deep transfer learning features and hand-crafted radiomics features","volume":"24","author":"Zhang","journal-title":"BMC Musculoskelet Disord"},{"key":"R15","doi-asserted-by":"publisher","DOI":"10.3389\/fendo.2022.1093508","article-title":"Preoperative prediction of residual back pain after vertebral augmentation for osteoporotic vertebral compression fractures: initial application of a radiomics score based nomogram","volume":"13","author":"Ge","journal-title":"Front Endocrinol (Lausanne)"},{"issue":"4","key":"R16","doi-asserted-by":"publisher","DOI":"10.1148\/ryai.2019180066","article-title":"Urinary stone detection on CT images using deep convolutional neural networks: evaluation of model performance and generalization","volume":"1","author":"Parakh","journal-title":"Radiol Artif Intell"},{"key":"R17","doi-asserted-by":"publisher","DOI":"10.3389\/fendo.2024.1370838","article-title":"Exploring deep learning radiomics for classifying osteoporotic vertebral fractures in X-ray images","volume":"15","author":"Zhang","journal-title":"Front Endocrinol"},{"issue":"40","key":"R18","doi-asserted-by":"publisher","first-page":"670","DOI":"10.3238\/arztebl.m2021.0295","article-title":"Osteoporotic fractures of the thoracic and lumbar vertebrae: diagnosis and conservative treatment","volume":"118","author":"Spiegl","journal-title":"Dtsch Arztebl Int"},{"issue":"5","key":"R19","doi-asserted-by":"publisher","first-page":"1156","DOI":"10.1007\/s10278-021-00506-6","article-title":"Benchmarking various radiomic toolkit features while applying the image biomarker standardization initiative toward clinical translation of radiomic analysis","volume":"34","author":"Lei","journal-title":"J Digit Imaging"},{"key":"R20","unstructured":"Pyradiomics. Python Software Foundation. URL: http:\/\/pypi.org\/project\/pyradiomics [Accessed 17-05-2023]"},{"key":"R21","unstructured":"PyRadiomics documentation. pyradiomics. URL: http:\/\/pyradiomics.readthedocs.io [Accessed 11-12-2016]"},{"issue":"1","key":"R22","doi-asserted-by":"publisher","first-page":"53","DOI":"10.1148\/radiol.2019182023","article-title":"Validation of a method to compensate multicenter effects affecting CT radiomics","volume":"291","author":"Orlhac","journal-title":"Radiology"},{"key":"R23","doi-asserted-by":"publisher","DOI":"10.1016\/j.clineuro.2015.09.006","article-title":"An assessment system for evaluating the severity of thoracolumbar osteoporotic fracture and its clinical application: a retrospective study of 381 cases","volume":"139","author":"Xu","journal-title":"Clin Neurol Neurosurg"},{"issue":"1","key":"R24","doi-asserted-by":"publisher","first-page":"534","DOI":"10.1007\/s10278-024-01204-9","article-title":"RadImageNet and ImageNet as datasets for transfer learning in the assessment of dental radiographs: a comparative study","volume":"38","author":"Okazaki","journal-title":"J Digit Imaging Inform med"},{"key":"R25","doi-asserted-by":"publisher","DOI":"10.2196\/54944","article-title":"Combining clinical-radiomics features with machine learning methods for building models to predict postoperative recurrence in patients with chronic subdural hematoma: retrospective cohort study","volume":"26","author":"Fang","journal-title":"J Med Internet Res"},{"issue":"1","key":"R26","doi-asserted-by":"publisher","first-page":"56","DOI":"10.1186\/s12880-024-01517-9","article-title":"Deep learning models for differentiating three sinonasal malignancies using multi-sequence MRI","volume":"25","author":"Wang","journal-title":"BMC Med Imaging"},{"key":"R27","unstructured":"SHAP. GitHub. URL: https:\/\/github.com\/slundberg\/shap [Accessed 20-04-2018]"},{"key":"R28","doi-asserted-by":"publisher","DOI":"10.1016\/j.compbiomed.2024.109168","article-title":"Polynomial-SHAP analysis of liver disease markers for capturing of complex feature interactions in machine learning models","volume":"182","author":"Ejiyi","journal-title":"Comput Biol Med"},{"issue":"4","key":"R29","doi-asserted-by":"publisher","first-page":"152","DOI":"10.1016\/j.afos.2022.11.001","article-title":"Lower SMI is a risk factor for dysphagia in Japanese hospitalized patients with osteoporotic vertebral and hip fracture: a retrospective study","volume":"8","author":"Suseki","journal-title":"Osteoporos Sarcopenia"},{"issue":"8","key":"R30","doi-asserted-by":"publisher","first-page":"3242","DOI":"10.1007\/s00586-024-08235-4","article-title":"Development and validation of a predictive model for vertebral fracture risk in osteoporosis patients","volume":"33","author":"Zhang","journal-title":"Eur Spine J"},{"issue":"4","key":"R31","doi-asserted-by":"publisher","first-page":"646","DOI":"10.1177\/2192568220964051","article-title":"Relationship between the OF classification and radiological outcome OF osteoporotic vertebral fractures after kyphoplasty","volume":"12","author":"Palmowski","journal-title":"Global Spine J"},{"issue":"7","key":"R32","doi-asserted-by":"publisher","first-page":"1598","DOI":"10.1097\/CORR.0000000000001685","article-title":"Can a deep-learning model for the automated detection of vertebral fractures approach the performance level of human subspecialists?","volume":"479","author":"Li","journal-title":"Clin Orthop Relat Res"},{"key":"R33","doi-asserted-by":"publisher","DOI":"10.2196\/54641","article-title":"Development and validation of a prediction model using sella magnetic resonance imaging\u2013based radiomics and clinical parameters for the diagnosis of growth hormone deficiency and idiopathic short stature: cross-sectional, multicenter study","volume":"26","author":"Song","journal-title":"J Med Internet Res"},{"issue":"1","key":"R34","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-023-37720-5","article-title":"Interstitial lung disease diagnosis and prognosis using an AI system integrating longitudinal data","volume":"14","author":"Mei","journal-title":"Nat Commun"},{"key":"R35","doi-asserted-by":"publisher","DOI":"10.3389\/fonc.2021.654114","article-title":"CT radiomics model for discriminating the risk stratification of gastrointestinal stromal tumors: a multi-class classification and multi-center study","volume":"11","author":"Chen","journal-title":"Front Oncol"},{"key":"R36","doi-asserted-by":"publisher","DOI":"10.2196\/69328","article-title":"Detection and analysis of circadian biomarkers for metabolic syndrome using wearable data: cross-sectional study","volume":"13","author":"Kim","journal-title":"JMIR Med Inform"},{"issue":"4","key":"R37","doi-asserted-by":"publisher","first-page":"376","DOI":"10.1016\/j.hbpd.2023.04.001","article-title":"A deep learning model based on contrast-enhanced computed tomography for differential diagnosis of gallbladder carcinoma","volume":"23","author":"Xiang","journal-title":"Hepatobiliary Pabcreat Dis Int"}],"container-title":["JMIR Medical Informatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/medinform.jmir.org\/2025\/1\/e75665","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T18:06:48Z","timestamp":1782842808000},"score":1,"resource":{"primary":{"URL":"https:\/\/medinform.jmir.org\/2025\/1\/e75665"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,8,29]]},"references-count":37,"URL":"https:\/\/doi.org\/10.2196\/75665","relation":{},"ISSN":["2291-9694"],"issn-type":[{"value":"2291-9694","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,8,29]]},"article-number":"v13i10e75665"}}