Popular Boards
Abith Ganesh Kamath, Saran Singh Gill, Hussayn Shinwari | EFORT Open Reviews | (2026)
Key Takeaways
Plain English Takeaway
Computer programs that can learn from data are getting very good at helping doctors spot hip problems in children, but they still need more testing and should be used alongside expert doctors for now.
Study Aim
The main goal of this paper is to systematically review and summarize how well artificial intelligence (AI) and machine learning (ML) models perform when used to diagnose developmental dysplasia of the hip (DDH) in children using medical images. The authors aim to compare these AI-based methods to traditional ways of reading images and to highlight both the strengths and weaknesses of current research in this area.
Simply put: The study wants to find out if computer programs can help doctors better spot hip problems in kids using scans.
Study Design
The authors conducted a systematic review following PRISMA guidelines. They searched several medical databases for studies published since 1980 that used AI or machine learning to diagnose DDH in children using imaging methods like ultrasound or X-rays. Only English-language, peer-reviewed studies with full text and human pediatric subjects were included. Nineteen studies, covering a total of 36,907 patients, were analyzed. The researchers extracted data on study design, sample size, AI model type, imaging method, and diagnostic accuracy measures such as sensitivity (how well the test finds true cases), specificity (how well it avoids false alarms), and area under the receiver operating characteristic curve (AUROC, a measure of overall test performance). They used descriptive statistics and qualitative analysis to summarize the results, and assessed study quality and risk of bias using the QUADAS-2 tool.
Simply put: The researchers looked at many studies where computers helped doctors read hip scans in kids, checking how well these systems worked.
Findings
The review found that most AI models, especially those using deep learning (a type of computer program that learns patterns from lots of data), showed high accuracy, sensitivity, and specificity in diagnosing DDH—often matching or even beating human experts. Some models, like those using the mask R-CNN framework, were especially good at spotting key features in hip images. However, the studies varied widely in their methods, the types of images used, and the populations studied, making direct comparisons difficult. Some AI systems struggled with poor-quality images or unusual cases, and most worked best as helpers rather than replacements for expert doctors. The authors recommend using AI as a support tool for radiologists and for training new doctors, but stress the need for larger, more diverse studies and standard ways to measure performance. They also note that AI could help reduce mistakes and speed up diagnosis, but expert oversight is still needed for safety.
Simply put: Computer programs are very good at helping doctors find hip problems in kids, but they still need more testing and should not replace expert doctors yet.
Abstract
AI technologies hold significant potential for enhancing the diagnostic accuracy of DDH. However, existing variability and bias across studies highlight the need for further standardisation and validation.
Referenced In
Mercedes C.
2 months ago
Created: Apr 11, 2026