In 2014, it was estimated that there were just 450 anatomic phantoms in the world. Today, based on advanced models of breast anatomy, an infinite number of models exist. As such, it is possible to simulate individuals and specific pathologies from the population of all humans with increasingly higher accuracy. This, together with advanced models of image simulation, image processing and image reconstruction, means that we can create arbitrarily large databases of simulated images. At the same time, advances in machine observer methods mean that it is possible to conduct virtual clinical trials (VCT) using the simulated images, together with simulations of medical displays, human optical perception and cognition.
The logistics of conducting VCT with thousands of patients is similar to the logistics of organizing the data from clinical trials of similar size. As such, we have developed a standards document outlining methods for conducting VCT, storing VCT results (intermediate and final), and communicating these image data and associate metadata between VCT components. In this course, we will use our experience in conducting large-scale VCT to encourage those new to the field to adopt VCT methods and to aid those already conducting VCT. The course will have applicability to VCT for designing new medical imaging equipment and methods, to use VCT data for prototyping and/or complementing the conduct of real clinical trials, and for preparing VCT data for regulatory approvals of new systems and methods.
- describe the roles and methods for conducting VCT
- identify the necessary constituent software components for conducting VCT
- name the standards relevant for conducting VCT, including DICOM, ASME, IEEE, AAPM, etc.
- construct and Design examples of VCTs to illustrate there usage
- demonstrate existing use cases
- explain the underlying statistical considerations for conducting VCT