Related Experiment Video
Updated: Jun 15, 2025
Protocol and Guidelines for Point-of-Care Lung Ultrasound in Diagnosing Neonatal Pulmonary Diseases Based on International Expert Consensus
Published on: March 6, 2019
Development and testing of a deep learning algorithm to detect lung consolidation among children with pneumonia using
David Kessler1, Meihua Zhu2, Cynthia R Gregory2
1Department of Emergency Medicine, Columbia University Vagelos College of Physicians & Surgeons, New York Presbyterian Morgan Stanley Children's Hospital, NY, NY, United States of America.
Insights
An artificial intelligence algorithm accurately detects pneumonia in children using lung ultrasounds. This AI tool offers crucial diagnostic support for severe pneumonia in resource-limited settings.
Area of Science:
- Artificial Intelligence in Medical Diagnostics
- Pediatric Pulmonology
- Point-of-Care Ultrasound
Background:
- Severe pneumonia is a leading cause of mortality in young children globally.
- Access to advanced diagnostic imaging is limited in many regions, hindering timely diagnosis.
- There is a need for accessible and accurate diagnostic tools for pediatric pneumonia.
Purpose of the Study:
- To develop and validate an artificial intelligence (AI) algorithm for detecting pulmonary consolidation on point-of-care lung ultrasounds.
- To assess the accuracy of the AI algorithm in hospitalized children with suspected pneumonia.
Main Methods:
- A prospective, multicenter study enrolled 107 pediatric participants (18 months to 17 years) with suspected lower respiratory tract infections.
- Bedside lung ultrasounds were performed using a handheld device, with standardized protocols for data collection.
- An AI algorithm was trained, tuned, and tested on video data to detect consolidation, with adult data augmenting the training set.
Main Results:
- The AI algorithm achieved an overall accuracy of 88.5% for identifying and localizing consolidation.
- Specific performance metrics included 88% sensitivity, 89% specificity, 89% positive predictive value, and 87% negative predictive value.
- The model was developed using data from 107 pediatric participants, comprising 604 positive and 589 negative videos for consolidation.
Conclusions:
- The developed AI algorithm demonstrates high accuracy in identifying consolidation on pediatric chest ultrasounds for pneumonia diagnosis.
- Automated diagnostic support via ultraportable point-of-care devices holds significant potential for global health, especially in resource-limited environments.
- This technology can improve pneumonia detection and management in children, particularly in austere settings lacking advanced imaging capabilities.
Background And Objectives:
Severe pneumonia is the leading cause of death among young children worldwide, disproportionately impacting children who lack access to advanced diagnostic imaging. Here our objectives were to develop and test the accuracy of an artificial intelligence algorithm for detecting features of pulmonary consolidation on point-of-care lung ultrasounds among hospitalized children.
Methods:
This was a prospective, multicenter center study conducted at academic Emergency Department and Pediatric inpatient or intensive care units between 2018-2020. Pediatric participants from 18 months to 17 years old with suspicion of lower respiratory tract infection were enrolled. Bedside lung ultrasounds were performed using a Philips handheld Lumify C5-2 transducer and standardized protocol to collect video loops from twelve lung zones, and lung features at both the video and frame levels annotated. Data from both affected and unaffected lung fields were split at the participant level into training, tuning, and holdout sets used to train, tune hyperparameters, and test an algorithm for detection of consolidation features. Data collected from adults with lower respiratory tract disease were added to enrich the training set. Algorithm performance at the video level to detect consolidation on lung ultrasound was determined using reference standard diagnosis of positive or negative pneumonia derived from clinical data.
Results:
Data from 107 pediatric participants yielded 117 unique exams and contributed 604 positive and 589 negative videos for consolidation that were utilized for the algorithm development process. Overall accuracy for the model for identification and localization of consolidation was 88.5%, with sensitivity 88%, specificity 89%, positive predictive value 89%, and negative predictive value 87%.
Conclusions:
Our algorithm demonstrated high accuracy for identification of consolidation features on pediatric chest ultrasound in children with pneumonia. Automated diagnostic support on an ultraportable point-of-care device has important implications for global health, particularly in austere settings.
