Introduction:. Most of the published papers on AI based diagnosis have focused on the algorithm's diagnostic performance in a ‘binary’ setting (i.e. disease vs no disease). However, no study evaluated the actual value for the clinicians of an AI based approach in diagnostic. Detection of Traumatic thoracolumbar (TL) fractures is challenging on planar radiographs, resulting in significant rates of missed diagnoses (30-60%), thus constituting a field in which a performance improvement is needed. Aim of this study is therefore to evaluate the value provided by AI generated saliency maps (SM), i.e. the maps that highlight the AI identified region of interests. Methods:. An AI model aimed at identifying TL fractures on plain radiographs was trained and tested on 567 single vertebrae images. Three expert spine surgeons established the Ground Truth (GT) using CT and MRI to confirm the presence of the fracture. From the test set, 12 cases (6 with a GT of fracture and 6 with a GT of no fracture, associated with varying levels of algorithm confidence) were selected and the corresponding SMs were generated and shown to 7 independent evaluators with different grade of experience; the evaluators were requested to: (1) identify the presence or absence of a fracture before and after the saliency map was shown; (2) grade, with a score from 1 (low) to 6 (high) the pertinency (correlation between the map and the human diagnosis), and the utility (the perceived utility in confirming or not the initial diagnosis) of the SM. Furthermore, the usefulness of the SM was evaluated through the rate of correct change in diagnosis after the maps had been shown. Finally, the obtained scores were correlated with the algorithm confidence for the specific case. Results:. Of the selected maps, 8 had an agreement between the AI diagnosis and the GT, while in 4 the diagnosis was discordant (67% accuracy). The pertinency of the map was found higher when the AI diagnosis was the same as the GT and the human diagnosis (respectively p-value = .021 and <.000). A positive and significant correlation between the AI confidence score and the perceived utility (Spearman: 27%, p-value=.0-27) was found. Furthermore, evaluator with experience < 5 year found the maps more useful than the experts (z-score=2.004; p-value=.0455). Among the 84 evaluation we found 12 diagnostic errors in respect to the GT, 6 (50%) of which were reverted after the saliency map evaluation (z statistic = 1.25 and p-value = .21). Discussion:. The perceived utility of AI generated SM correlate with the model confidence in the diagnosis. This highlights the fact that to be considered helpful, the AI must provide not only the diagnosis but also the case specific confidence. Furthermore, the perceived utility was higher among less experienced users, but overall, the SM were useful in improving the human diagnostic accuracy. Therefore, in this setting, the AI enhanced approach provides value in improving the human performance

Introduction and Objective. Up to 30% of thoracolumbar (TL) fractures are missed in the emergency room. Failure to identify these fractures can result in neurological injuries up to 51% of the casesthis article aimed to clarify the incidence and risk factors of traumatic fractures in China. The China National Fracture Study (CNFS. Obtaining sagittal and anteroposterior radiographs of the TL spine are the first diagnostic step when suspecting a traumatic injury. In most cases, CT and/or MRI are needed to confirm the diagnosis. These are time and resource consuming. Thus, reliably detecting vertebral fractures in simple radiographic projections would have a significant impact. We aim to develop and validate a deep learning tool capable of detecting TL fractures on lateral radiographs of the spine. The clinical implementation of this tool is anticipated to reduce the rate of missed vertebral fractures in emergency rooms. Materials and Methods. We collected sagittal radiographs, CT and MRI scans of the TL spine of 362 patients exhibiting traumatic vertebral fractures. Cases were excluded when CT and/or MRI where not available. The reference standard was set by an expert group of three spine surgeons who conjointly annotated (fracture/no-fracture and AO Classification) the sagittal radiographs of 171 cases. CT and/or MRI were used confirm the presence and type of the fracture in all cases. 302 cropped vertebral images were labelled “fracture” and 328 “no fracture”. After augmentation, this dataset was then used to train, validate, and test deep learning classifiers based on the ResNet18 and VGG16 architectures. To ensure that the model's prediction was based on the correct identification of the fracture zone, an Activation Map analysis was conducted. Results. Vertebras T12 to L2 were the most frequently involved, accounting for 48% of the fractures. Accuracies of 88% and 84% were obtained with ResNet18 and VGG16 respectively. The sensitivity was 89% with both architectures but ResNet18 had a significantly higher specificity (88%) compared to VGG16 (79%). The fracture zone used was precisely identified in 81% of the heatmaps. Conclusions. Our AI model can accurately identify anomalies suggestive of TL vertebral fractures in sagittal radiographs precisely identifying the fracture zone within the vertebral body

Impacted bone allograft is often used in revision joint replacement. Hydroxyapatite granules have been suggested as a substitute or to enhance morcellised bone allograft. We hypothesised that adding osteogenic protein-1 to a composite of bone allograft and non-resorbable hydroxyapatite granules (ProOsteon) would improve the incorporation of bone and implant fixation. We also compared the response to using ProOsteon alone against bone allograft used in isolation. We implanted two non-weight-bearing hydroxyapatite-coated implants into each proximal humerus of six dogs, with each implant surrounded by a concentric 3 mm gap. These gaps were randomly allocated to four different procedures in each dog: 1) bone allograft used on its own; 2) ProOsteon used on its own; 3) allograft and ProOsteon used together; or 4) allograft and ProOsteon with the addition of osteogenic protein-1.

After three weeks osteogenic protein-1 increased bone formation and the energy absorption of implants grafted with allograft and ProOsteon. A composite of allograft, ProOsteon and osteogenic protein-1 was comparable, but not superior to, allograft used on its own.

ProOsteon alone cannot be recommended as a substitute for allograft around non-cemented implants, but should be used to extend the volume of the graft, preferably with the addition of a growth factor.