The Rise of Virtual Imaging Trials
A faster, safer and cost-effective alternative to using only imaging trials
Providing a way to understand the most effective and safe means for preventing, diagnosing and treating a disease, clinical trials have long been a cornerstone of health care. In radiology, imaging trials are used similarly to evaluate and optimize the design and clinical use of medical imaging devices and methods.
These trials, however, tend to be expensive, time-consuming, and they often require a large patient cohort. They might also carry some risk, such as when trials require the use of ionizing radiation.
“In medicine, there’s a significant limitation to our science. We often do not know the exact ground truth of patients, meaning, we do not know the exact status of the disease, its extent, its biology or morphology. In VIT, we know those details, because they are precise and realistic models in the computational space,” said Ehsan Samei, PhD, director and principal investigator at Duke University’s Center for Virtual Imaging Trials.
By means of realistic and validated simulations, virtual imaging trials (VITs) offer an effective avenue for assessing and optimizing the performance of imaging protocols and technologies.
“This is due to the benefits of VITs, in terms of duration, cost, and availability of appropriate patient population. Imaging systems development routinely includes repeated cycles in which various combinations of system parameters are tested and optimized. Repeating patient images for each parameter combination is prohibitive and costly. Similarly, repeating animal studies to that goal is also costly–and not humane,” explained Predrag R. Bakic, PhD, associate professor of diagnostic radiology at Lund University in Sweden.
Moreover, finding the desired group of clinical patients to adequately test the system might take too long, or just not be possible, according to Dr. Bakic.
“VITs are flexible to vary the characteristics of virtual patients and repeat VITs as many times as needed, for the cost of computer time only,” he said. “Lastly, physical phantoms are designed to test only specific characteristics of imaging systems, and, thus, cannot capture variations in shape, size, and internal tissue composition as seen in clinical patients. Again, virtual patients allow for varying those characteristics and repeating VITs as needed.”
The Anatomy of a Virtual Imaging Trial
Virtual imaging trials are based on the simulation of human anatomy, imaging modalities and image interpretation, which parallel the components of real-life imaging trials. Depending on the imaging task to be assessed, each component will include various models to be simulated.
“Anatomy simulation serves to define a ground truth against which we assess the results from the analysis of simulated medical images,” said Dr. Bakic, who co-authored an article in Radiation Protection Dosimetry on the evaluation and optimization of VITs.
Dr. Bakic noted that because the imaging modalities are being simulated based on the physical principles used to generate images from human anatomy, the simulation needs to be as detailed as necessary for the simulation task in question. For example, the spatial/temporal resolution, or the quantification of the analyzed image descriptors, should include the details needed to perform the required detection or quantification task.
This point touches on one of VITs shortcomings: ensuring that patient models, particularly digital twins, are realistic interpretations of real patients and offer the appropriate level of detail. Another shortcoming is a lack of diversity.
“Realistic VIT requires not only realistic digital twins of the patient’s anatomy and physiology, but also those of diagnostic and therapeutic methodologies so a clinical scenario can be realistically replicated,” Dr. Samei said. “Such details are essential to ensure our answers would be reliable and relevant.”
Virtual Imaging Trials Offer Multiple Benefits
Despite any shortcomings, Dr. Bakic noted that the potential benefits of VITs are multifold.
“In addition to their time and cost saving potential, VITs can accelerate the development of medical imaging systems and help bring novel imaging methods to patients quicker,” he said. “They can also be used to test and optimize digital twins for personalized therapy research.”
Furthermore, VITs can be used to design effective clinical trials. For example, VITs could discern exactly what severity of a disease and how many cases of each type would need to be recruited in a clinical trial to most efficiently discern the utility of particular medical intervention at a desired level of accuracy.
To illustrate, Dr. Samei points to assessing the AI technologies used in radiology.
“Virtual imaging trials provide data sets with precise ground truth—not approximate ones—so we can effectually and proficiently evaluate the claims being made by AI innovations,” he said.
Leveraging the Use of Virtual Trials in Imaging Research
Both Drs. Samei and Bakic are adamant in their belief that radiology should fully leverage the use—and benefits—of virtual trials in imaging research.
“Radiology would be remiss not to take full advantage of virtual imaging trials,” Dr. Samei said. Such advantages, he offers, include offering a significant ability to massively reduce the time and costs of a trial; prototyping medical innovations cost-effectively, establishing digital twin individuals and populations; reducing the need for animal trials; and driving definitive and quantitative endpoints for medical interactions.
“By increasing access to medical imaging research for a wider range of academic and industrial labs—even those without direct access to clinical facilities and databases—VITs make our field more democratic, accelerate development and, hopefully, may improve patient care,” Dr. Bakic concluded.
For More Information
Access the Radiation Protection Dosimetry article at academic.oup.com/rpd.
Access Duke University’s Center for Virtual Imaging.
First International Summit on VIT Held
The prospect and opportunity of VIT in medicine was highlighted during VITM24, the first international summit on Virtual Imaging Trials in Medicine.
Held at Duke University and chaired by Dr. Samei, VITM24 brought over 120 participants across North America, Europe, Asia and Oceania together to discuss and chart the future course of the technology and its position to advance radiology and health care.
Already planned as an annual event, the next summit—VITM25—will take place June 2-4, 2025, in Manchester, UK.