Artificial intelligence-based detection of paediatric appendicular skeletal fractures: performance and limitations for common fracture types and locations.

BACKGROUND: Research into artificial intelligence (AI)-based fracture detection in children is scarce and has disregarded the detection of indirect fracture signs and dislocations. OBJECTIVE: To assess the diagnostic accuracy of an existing AI-tool for the detection of fractures, indirect fracture signs, and dislocations. MATERIALS AND METHODS: An AI software, BoneView (Gleamer, Paris, France), was assessed for diagnostic accuracy of fracture detection using paediatric radiology consensus diagnoses as reference. Radiographs from a single emergency department were enrolled retrospectively going back from December 2021, limited to 1,000 radiographs per body part. Enrolment criteria were as follows: suspected fractures of the forearm, lower leg, or elbow; age 0-18 years; and radiographs in at least two projections. RESULTS: Lower leg radiographs showed 607 fractures. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were high (87.5%, 87.5%, 98.3%, 98.3%, respectively). Detection rate was low for toddler's fractures, trampoline fractures, and proximal tibial Salter-Harris-II fractures. Forearm radiographs showed 1,137 fractures. Sensitivity, specificity, PPV, and NPV were high (92.9%, 98.1%, 98.4%, 91.7%, respectively). Radial and ulnar bowing fractures were not reliably detected (one out of 11 radial bowing fractures and zero out of seven ulnar bowing fractures were correctly detected). Detection rate was low for styloid process avulsions, proximal radial buckle, and complete olecranon fractures. Elbow radiographs showed 517 fractures. Sensitivity and NPV were moderate (80.5%, 84.7%, respectively). Specificity and PPV were high (94.9%, 93.3%, respectively). For joint effusion, sensitivity, specificity, PPV, and NPV were moderate (85.1%, 85.7%, 89.5%, 80%, respectively). For elbow dislocations, sensitivity and PPV were low (65.8%, 50%, respectively). Specificity and NPV were high (97.7%, 98.8%, respectively). CONCLUSIONS: The diagnostic performance of BoneView is promising for forearm and lower leg fractures. However, improvement is mandatory before clinicians can rely solely on AI-based paediatric fracture detection using this software.

The effect of the COVID-19 pandemic on pediatric emergency department utilization in three regions in Switzerland.

PURPOSE: The COVID-19 pandemic was associated with a decrease in emergency department (ED) visits. However, contradictory, and sparse data regarding children could not yet answer the question, how pediatric ED utilization evolved throughout the pandemic. Our objectives were to investigate the impact of the pandemic in three language regions of Switzerland by analyzing trends over time, describe regional differences, and address implications for future healthcare. METHODS: We conducted a retrospective, longitudinal cohort study at three Swiss tertiary pediatric EDs (March 1st, 2018-February 28th, 2022), analyzing the numbers of ED visits (including patients` age, triage categories, and urgent vs. non-urgent cases). The impact of COVID-19 related non-pharmaceutical interventions (NPIs) on pediatric ED utilization was assessed by interrupted time series (ITS) modelling. RESULTS: Based on 304'438 ED visits, we found a drop of nearly 50% at the onset of NPIs, followed by a gradual recovery. This primarily affected children 0-4 years, and both non-urgent and urgent cases. However, the decline in urgent visits appeared to be more pronounced in two centers compared to a third, where also hospitalization rates did not decrease significantly during the pandemic. A subgroup analysis showed a significant decrease in respiratory and gastrointestinal diseases, and an increase in the proportion of trauma patients during the pandemic. CONCLUSIONS: The COVID-19 pandemic had substantial effects on number and reasons for pediatric ED visits, particularly among children 0-4 years. Despite equal regulatory conditions, the utilization dynamics varied markedly between the three regions, highlighting the multifactorial modification of pediatric ED utilization during the pandemic. Furthermore, future policy decisions should take regional differences into account.

Diagnostic potential of IL6 and other blood-based inflammatory biomarkers in mild traumatic brain injury among children.

Objectives: Inflammatory biomarkers, as indicators of biological states, provide a valuable approach for accurate and reproducible measurements, crucial for the effective management of mild traumatic brain injury (mTBI) in pediatric patients. This study aims to assess the diagnostic utility of blood-based inflammatory markers IL6, IL8, and IL10 in children with mTBI, including those who did not undergo computed tomography (CT) scans. Methods: A prospective multicentric cohort study involving 285 pediatric mTBI patients was conducted, stratified into CT-scanned and non-CT-scanned groups within 24 h post-trauma, alongside 74 control subjects. Biomarker levels were quantitatively analyzed using ELISA. Sensitivity and specificity metrics were calculated to determine the diagnostic efficacy of each biomarker. Results: A total of 223 mTBI patients (78%) did not undergo CT scan examination but were kept in observation for symptoms monitoring at the emergency department (ED) for more than 6 h (in-hospital-observation patients). Among CT-scanned patients (n = 62), 14 (23%) were positive (CT+). Elevated levels of IL6 and IL10 were found in mTBI children compared to controls. Within mTBI patients, IL6 was significantly increased in CT+ patients compared to both CT- and in-hospital-observation patients. No significant differences were observed for IL8 among the compared groups. IL6 yielded a specificity of 48% in identifying CT- and in-hospital-observation patients, with 100% sensitivity in excluding all CT+ cases. These performances were maintained whether IL6 was measured within 6 h or within 24 h after the trauma. Conclusion: The inflammatory marker IL6 emerges as a robust biomarker, showing promising stratification value for pediatric mTBI patients undergoing CT scans or staying in observation in a pediatric ED.

Management of Pediatric Mild Traumatic Brain Injury Patients: S100b, Glial Fibrillary Acidic Protein, and Heart Fatty-Acid-Binding Protein Promising Biomarkers.

Children are highly vulnerable to mild traumatic brain injury (mTBI). Blood biomarkers can help in their management. This study evaluated the performances of biomarkers, in discriminating between children with mTBI who had intracranial injuries (ICIs) on computed tomography (CT+) and (1) patients without ICI (CT-) or (2) both CT- and in-hospital-observation without CT patients. The aim was to rule out the need of unnecessary CT scans and decrease the length of stay in observation in the emergency department (ED). Newborns to teenagers (≤16 years old) with mTBI (Glasgow Coma Scale > 13) were included. S100b, glial fibrillary acidic protein (GFAP), and heart fatty-acid-binding protein (HFABP) performances to identify patients without ICI were evaluated through receiver operating characteristic curves, where sensitivity was set at 100%. A total of 222 mTBI children sampled within 6 h since their trauma were reported. Nineteen percent (n = 43/222) underwent CT scan examination, whereas the others (n = 179/222) were kept in observation at the ED. Sixteen percent (n = 7/43) of the children who underwent a CT scan had ICI, corresponding to 3% of all mTBI-included patients. When sensibility (SE) was set at 100% to exclude all patients with ICI, GFAP yielded 39% specificity (SP), HFABP 37%, and S100b 34% to rule out the need of CT scans. These biomarkers were even more performant: 52% SP for GFAP, 41% for HFABP, and 39% for S100b, when discriminating CT+ versus both in-hospital-observation and CT- patients. These markers can significantly help in the management of patients in the ED, avoiding unnecessary CT scans, and reducing length of stay for children and their families.