Journal Article 1 Mention
Accuracy of Trained Physicians is Inferior to Deep Learning-Based Algorithm for Determining Angles in Ultrasound of the Newborn Hip
David Oelen2020
Pascal KaiserThomas Baumann
Average · 50th percentile
11 citations · Surgery

TLDR

Computer programs that use deep learning can measure angles in baby hip ultrasound images more accurately than trained doctors can.

Summary

1 Study Aim

The main goal of this study is to compare how accurately trained physicians and a deep learning-based algorithm can measure angles in ultrasound images of newborn hips. The researchers want to find out if artificial intelligence can help doctors make better measurements in daily medical practice. Simply put: The study wants to see if computers can measure baby hip angles better than doctors.

2 Study Design

The researchers designed a study where both trained physicians and a deep learning-based algorithm measured angles in ultrasound images of newborn hips. The study compared the accuracy of the doctors' measurements with those made by the algorithm. The sample included ultrasound images taken during routine newborn hip screenings. The study focused on real-world conditions to reflect daily clinical practice. Simply put: The study compared how well doctors and a computer program measured angles in baby hip scans.

3 Findings

The study demonstrates that the deep learning-based algorithm was more accurate than trained physicians at measuring angles in newborn hip ultrasound images. The authors suggest that similar artificial intelligence methods could be used to support doctors in clinical settings. This could lead to more reliable diagnoses and better care for newborns. The findings imply that integrating deep learning tools into routine practice may improve measurement accuracy and support medical decision-making. Simply put: The computer program did a better job than doctors at measuring baby hip angles, so using this technology could help doctors do their jobs better.

Abstract

The accuracy of physicians in their daily routine is inferior to deep learning-based algorithms for determining angles in ultrasound of the newborn hip. Similar methods could be used to support physicians.

Referenced In