2D/3D ultrasound diagnosis of pediatric distal radius fractures by human readers vs artificial intelligence

Abstract

Wrist trauma is common in children and generally requires radiography for exclusion of fractures, subjecting children to radiation and long wait times in the emergency department. Ultrasound (US) has potential to be a safer, faster diagnostic tool. This study aimed to determine how reliably US could detect distal radius fractures in children, to contrast the accuracy of 2DUS to 3DUS, and to assess the utility of artificial intelligence for image interpretation. 127 children were scanned with 2DUS and 3DUS on the affected wrist. US scans were then read by 7 blinded human readers and an AI model. With radiographs used as the gold standard, expert human readers obtained a mean sensitivity of 0.97 and 0.98 for 2DUS and 3DUS respectively. The AI model sensitivity was 0.91 and 1.00 for 2DUS and 3DUS respectively. Study data suggests that 2DUS is comparable to 3DUS and AI diagnosis is comparable to human experts.

2D vs. 3D Ultrasound Diagnosis of Pediatric Supracondylar Fractures

“…It may also be useful to target a narrower age group for investigation in future studies as the anatomy of the elbow varies greatly between the ages of 0 and 17 and could have led to some difficulty for readers in determining expected normal anatomy, especially when blinded to all clinical data [18]. Similar to other MSK ultrasound applications, we expect that the elbow ultrasound data that we have collected will be highly amenable to real-time analysis and interpretation with AI [6]. In future research, we plan to investigate whether AI can decrease variability between readers and increase the accuracy and reproducibility of supracondylar fracture diagnosis using 2D POCUS.…”

“…Kappa values showed moderate agreement between readers, with the sensitivity of fracture detection by 2D POCUS and 3DUS ranging from 0.64 to 0.91 and 0.46 to 0.73, respectively. Given the operator-dependent nature of a US, this variability is likely to increase even further with limited user experience, as it does in similar MSK applications of POCUS [6].…”

“…Though this approach can be helpful in preventing complications of missed fractures, inevitably, some children are immobilized unnecessarily, and their daily activities are affected [5]. Another limitation is that children often have to wait for long periods of time in the emergency departments (ED) and often move between departments to undergo their radiographs as a result of the increased demand for healthcare services and the immobility of diagnostic imaging services [6]. Furthermore, radiographic images may not be interpreted promptly, compounding the in-hospital wait times families experience before children are discharged in the absence of fracture or receive care for their fracture [6].…”

“…Recent studies have demonstrated that POCUS performed by pediatric emergency doctors with appropriate training in musculoskeletal sonography can have a high sensitivity and specificity for supracondylar fractures compared to radiographs [8][9][10]. Additionally, highly portable handheld POCUS probes that can easily be connected to a portable smartphone or tablet device have become readily available, which has led to a growing opportunity for the adoption of POCUS in the ED setting and has made POCUS accessible to physicians or healthcare professionals of varying skill levels [6]. One major drawback to ultrasound is that it is entirely user-dependent, and the required training and repeated exposure that is necessary for medical professionals to use it effectively and interpret the findings is not broadly accessible, limiting its utility [6].…”