Advancements in AI for Pediatric Radiology

Advancements in AI for Pediatric Radiology

Table of Contents:

1. Introduction
2. The Current State of AI in Pediatric Radiology 2.1. PubMed Search Results 2.2. Reasons for the Lag in Pediatric Radiology AI Development
3. Well-Known Uses of AI in Pediatric Radiology 3.1. Bone Age Assessment 3.2. Commercially Available Software 3.3. Survey Results
4. Meaningful Uses of AI in Pediatric Radiology 4.1. Short-Term Applications 4.1.1. Providing Better and Safer Exams for Children 4.1.2. Shortening MRI Exam Length 4.1.3. Reducing Radiation Exposure in CT and Fluoro 4.1.4. Reducing Dose of Gadolinium in Contrast Brain MRIs 4.2. Medium-Term Applications 4.2.1. Questioning the Current State of Radiology 4.2.2. Integration with Clinical Data
5. Summary of AI Development in Pediatric Radiology
6. Conclusion

Introduction

Hello everyone! First of all, I would like to thank Jordan for the invitation and each one of you for attending this session. My name is Marcelo, and I am a pediatric radiologist based in Sao Paulo, Brazil. I divide my time between clinical practice and working with AI in radiology. Today, I will be discussing the current state and meaningful applications of AI in pediatric imaging.

The Current State of AI in Pediatric Radiology

When it comes to AI in pediatric radiology, the development has been slower compared to other radiology areas. A PubMed search demonstrates a notable difference in the number of publications on AI and radiology between the general radiology and pediatric radiology fields. The limited adoption of AI in pediatric radiology can be attributed to several reasons.

Well-Known Uses of AI in Pediatric Radiology

Despite the slower development, there are some well-established applications of AI in pediatric radiology. One such application is bone age assessment, which involves classifying HAND and wrist X-rays based on an Atlas. Commercially available software, like Bone Expert, has streamlined and improved the efficiency of bone age assessment. A recent survey showed a significant reduction in the time taken to evaluate bone age when using AI assistance.

Meaningful Uses of AI in Pediatric Radiology

In the short-term, AI can offer better and safer exams for children. Techniques such as denoising and super resolution can be employed to shorten MRI exam length and reduce motion artifacts. Similarly, in CT and fluoroscopy, AI algorithms can help reduce radiation exposure while maintaining diagnostic quality. The use of AI in reducing the dose of gadolinium in contrast brain MRIs is also gaining attention.

In the medium-term, there are two significant areas where AI can make a meaningful impact. Firstly, it is important to question the current state of radiology and move away from relying solely on standardized atlases. Researchers have explored using specific data from the population being analyzed to provide more accurate assessments. Secondly, integrating clinical data with imaging data can enhance the diagnostic process. By analyzing large datasets of patients with clinically defined conditions, the AI algorithms can learn to identify Patterns that may not be visible to human radiologists.

Summary of AI Development in Pediatric Radiology

In summary, the development of AI in pediatric radiology has been slower compared to other subspecialties. However, there are notable applications of AI in areas such as bone age assessment. In the short-term, AI can contribute to better and safer exams for children. In the medium-term, the focus should be on questioning current radiology practices and integrating clinical data with imaging data.

Conclusion

AI has the potential to revolutionize pediatric radiology in terms of diagnostic accuracy, efficiency, and patient care. While there are challenges specific to pediatric radiology, such as the heterogeneity of the pediatric population and varying disease presentations, AI can still play a valuable role. By actively exploring and implementing AI Tools, we can improve clinical workflows and enhance the overall quality of pediatric imaging.

Highlights:

• AI development in pediatric radiology has been slower compared to other subspecialties.
• Bone age assessment is a well-established application of AI in pediatric radiology.
• Short-term applications of AI include providing better and safer exams for children and reducing radiation exposure.
• Medium-term applications involve questioning the current state of radiology practices and integrating clinical data with imaging data.

FAQ

Q: What is the current state of AI in pediatric radiology? A: AI development in pediatric radiology has been slower compared to other subspecialties.

Q: Is bone age assessment an established application of AI in pediatric radiology? A: Yes, bone age assessment is a well-known use of AI in pediatric radiology.

Q: Can AI help reduce radiation exposure in pediatric imaging? A: Yes, AI algorithms can aid in reducing radiation exposure in CT and fluoroscopy exams.

Q: What are the medium-term applications of AI in pediatric radiology? A: Medium-term applications include questioning the current state of radiology practices and integrating clinical data with imaging data.