Machine Learning and Deep Learning in Cardiothoracic Imaging: A Scoping Review.

Area of Science:

• Medical imaging analysis
• Artificial intelligence in healthcare

Background:

• Machine-learning (ML) and deep-learning (DL) algorithms excel at identifying complex patterns in data through training.
• These AI techniques have seen increased application in medical image processing over the last decade.
• Cardiothoracic imaging has been an early adopter, with the COVID-19 pandemic further accelerating research in ML/DL applications.

Purpose of the Study:

• To systematically review and quantitatively analyze the use of ML/DL in cardiothoracic imaging.
• To provide a comprehensive overview of current ML/DL applications and their nuances within this specialty.
• To offer suggestions for advancing ML/DL research from proof-of-concept to clinical adoption.

Main Methods:

• Systematic search of peer-reviewed medical literature.
• Quantitative extraction of key data elements from selected studies.
• Scoping review methodology to synthesize findings.

Main Results:

• ML/DL algorithms are increasingly utilized for complex information extraction in cardiothoracic imaging.
• The field has seen significant research growth, particularly post-COVID-19.
• Diverse applications of ML/DL have been identified within cardiothoracic imaging.

Conclusions:

• ML/DL holds significant potential to enhance cardiothoracic imaging analysis.
• Bridging the gap between research and clinical practice requires strategic development.
• Further research should focus on facilitating the clinical integration of these powerful AI tools.