AI and Machine Learning in Radiology:
Practical Considerations for Implementation in the Era of COVID-19
Date: Saturday, 24 April 2021
Time: 15:00 -17:00 GMT +8 / 07:00 – 09:00 GMT
Rapid advances in artificial intelligence (AI), machine learning (ML) and deep learning (DL) have led to promising real world applications in numerous fields, including Radiology.
Join us as we cover the basics concepts of AI, ML and DL, with emphasis on the clinical and technical considerations when building an AI algorithm in healthcare. Gain insights into challenges and opportunities in developing AI algorithms for clinical tasks. Learn how AI can augment radiologists in the diagnosis and prognostication of COVID-19 pneumonia, to facilitate clinical decision making during the global pandemic. See how AI can aid in the interpretation of chest CT scans for tasks such as lung cancer detection, classification and quantification of diffuse lung diseases like emphysema, and evaluation of cardiovascular and respiratory risks.
- Learn basic concepts of artificial intelligence (AI), machine learning (ML) and deep learning (DL)
- Learn clinical and technical considerations, as well as challenges and opportunities for AI algorithm development in medicine
- Learn about current developments in AI for chest imaging
- Learn radiologists’ experience in AI use for early detection, accurate diagnosis and follow up decision support for COVID-19 in France and Singapore.
- Understand the potential pitfalls for COVID-19 AI tools and future development.
- Understand the latest technical achievements and potential of AI in radiological imaging.
- Understand basic principles of performance evaluation of AI software devices.
Dr Daniel Ting
The real-world application of AI, ML and DL has generated significant interest throughout the computer science and medical communities in recent years. DL has revolutionized the computer vision field and achieved substantial jumps in diagnostic performance for image recognition, speech recognition, and natural language processing. In medicine, this technique has shown promising diagnostic performance in the detection of eye, skin or chest diseases for different imaging modalities, including Radiology. This talk aims to cover the basic concepts of artificial intelligence (AI), machine learning (ML) and deep learning (DL), with the clinical/technical considerations when building an AI algorithm in healthcare.
Dr Ting Yong Han
The talk will cover the process in developing and implementing RadiLogic, an algorithm used in the prioritisation of screening chest X-rays performed in NCID for suspected COVID patients. In addition, the challenges and lessons in machine learning (ML) development will be discussed.
Prof Philippe Grenier
While there have been many reports of deep learning solutions providing automatic detection of multiple common chest abnormalities on radiography, AI has also been applied to CT. Numerous algorithms to detect, segment and quantify various chest lesions are currently running in many institutions or are still under development. Deep learning algorithms have also been designed to classify and quantify diffuse lung disease (COPD, interstitial lung disease) to improve follow-up assessment or to predict outcomes. We will present two specific applications. The first consists of a multi-task AI-based solution able to provide automatic detection and volumetric quantifications of solid pulmonary nodules, pulmonary emphysema, cardiac calcifications, and thoracic aorta diameters perpendicular to its central axis, to run a multidisciplinary platform for a personalized evaluation of cardiovascular and respiratory risks and detection of lung cancer on low-dose CT scans in smokers. In the second example, we used an automated 3D AI-driven CT quantification of lung disease on initial chest CT scan in patients hospitalized for COVID-19 pneumonia to predict severe outcomes (clinical deterioration leading to a transfer in intensive care unit, or death). Both independent and incremental (association with comorbidities and biological data) prognostic CT values were demonstrated.
- Dr Daniel Ting
Associate Professor in Ophthalmology, Duke-NUS Medical School, Singapore
- Dr Ting Yong Han
Consultant, Diagnostic Radiology, Tan Tock Seng Hospital, Singapore
- Prof Philippe Grenier
Former Professor of Radiology and Chairman, Sorbonne University. AI Implementation Lead, Hôpital Foch, France
- Dr Charles Goh
Consultant, Nuclear Medicine and Molecular Imaging, Singapore General Hospital, Singapore
- Dr Andrew Makmur
Consultant, Department of Diagnostic Imaging, National University Hospital, Singapore
Who Should Attend:
- Radiologists with an interest in AI or chest imaging
- Radiographers with an interest in AI
- Radiology Trainees
- Clinicians with an interest in AI or COVID-19 pneumonia imaging
- CMIOs, CTOs, and HODs