Can Artificial Intelligence Enhance Syncope Management?

Figures

• 

  Figure 1. Syncope in the Context of Transient Loss of Consciousness Adapted from the 2018 European Society of Cardiology Guidelines for the Diagnosis and Management of Syncope.3 TIA = transient ischemic attack; TLOC = transient loss of consciousness.

• 

  Figure 2. Short- and Long-Term Risk Factors in Syncope Patients Adapted from the 2017 ACC/AHA/HRS Guideline for the Evaluation and Management of Patients with Syncope.1

• 

  Figure 3. Steps Involved in Supervised Machine Learning (Top row) A supervised classifier is trained on different types of vehicles (eg, car, truck, and motorcycle). (Bottom row) When given a new set of vehicles of different colors/model types, the trained classifier can assign the correct vehicle label.

• 

  Central Illustration. Harnessing AI to Improve Syncope Management This diagram outlines clinical objectives and proposed steps of AI-based initiatives to improve syncope management. Created with Biorender.com. AI = artificial intelligence; OH = orthostatic hypotension; TLOC = transient loss of consciousness; VVS = vasovagal syncope.

• 

  Figure 4. Developing a ML Model for Syncope Proposed steps to build a ML model for syncope. (First row) Emphasizing importance of large scale, prospective, and international datasets. (Second row) Potential inputs (left) and outputs (right) of a ML model to address the main clinical objectives in syncope management while recognizing underlying ethical principles. (Third row) ML pipeline showing sequential stages of a supervised ML project. (Fourth row) Descriptions of each stage in the ML pipeline. Created with Flaticon.com and Biorender.com. TLOC = transient loss of consciousness.