Objectives : Trauma chest radiographs may contain subtle and time-critical pathology. Artificial intelligence (AI) may aid in accurate reporting, timely identification and worklist prioritisation. However, few AI programs have been externally validated. This study aimed to evaluate the performance of a commercially available deep convolutional neural network - Annalise CXR V1.2 (Annalise.ai)- for detection of traumatic injuries on supine chest radiographs. Methods: Chest radiographs with a CT performed within 24 h in the setting of trauma were retrospectively identified at a level one adult trauma centre between January 2009 and June 2019. Annalise.ai assessment of the chest radiograph was compared to the radiologist report of the chest radiograph. Contemporaneous CT report was taken as the ground truth. Agreement with CT was measured using Cohen’s κ and sensitivity/specificity for both AI and radiologists were calculated. Results: There were 1404 cases identified with a median age of 52 (IQR 33–69) years, 949 male. AI demonstrated superior performance compared to radiologists in identifying pneumothorax (p = 0.007) and segmental collapse (p = 0.012) on chest radiograph. Radiologists performed better than AI for clavicle fracture (p = 0.002), humerus fracture (p < 0.0015) and scapula fracture (p = 0.014). No statistical difference was found for identification of rib fractures and pneumomediastinum. Conclusion: The evaluated AI performed comparably to radiologists in interpreting chest radiographs. Further evaluation of this AI program has the potential to enable it to be safely incorporated in clinical processes. Advances in knowledge: Clinically useful AI programs represent promising decision support tools.

Purpose To assess the efficacy of conservative management and embolisation in patients with spontaneous retroperitoneal haemorrhage. Methods Single-centre retrospective case–control study of patients with spontaneous retroperitoneal haemorrhage treated conservatively or with embolisation. Patients aged ≥ 18 years were identified from CT imaging reports stating a diagnosis of retroperitoneal haemorrhage or similar and images reviewed for confirmation. Exclusion criteria included recent trauma, surgery, retroperitoneal vascular line insertion, or other non-spontaneous aetiology. Datapoints analysed included treatment approach (conservative or embolisation), technical success, clinical success, and mortality outcome. Results A total of 54 patients met inclusion criteria, who were predominantly anticoagulated (74%), male (72%), older adults (mean age 69 years), with active haemorrhage on CT (52%). Overall mortality was 15%. Clinical success was more likely with conservative management (36/38) than embolisation (9/16; p < 0.01), and all-cause (1/38 vs 7/16; p < 0.01) and uncontrolled primary bleeding (1/38 vs 5/16; p < 0.01) mortality were higher with embolisation. However, embolised patients more commonly had active bleeding on CT (15/38 vs 13/16; p < 0.01), shock (5/38 vs 6/16; p < 0.04), and higher blood transfusion volumes (mean 2.2 vs 5.9 units; p < 0.01). After one-to-one propensity score matching, differences in clinical success (p = 0.04) and all-cause mortality (p = 0.01) remained; however, difference in uncontrolled primary bleeding mortality did not (p = 0.07). Conclusion Conservative management of SRH is likely to be effective in most patients, even in those who are anticoagulated and haemodynamically unstable, with variable success seen after embolisation in a more unstable patient group, supporting the notion that resuscitation and optimisation of coagulation are the most vital components of treatment.

No abstract