Medical simulation, specifically in radiology, involves four facets. The original intent was for quality assurance and patient safety. More recently, the ability to practice image-guided interventions and modality-specific hands-on experience in diagnostic imaging has advanced the role medical simulation plays in residency programs as well as training for advanced practice professionals and medical students.

This review article intends to provide both background and supposition as to the utility of simulation, as it pertains to both diagnostic and interventional radiology.

Simulation in Quality Assurance & Patient Safety

It has been a long-standing practice to use anatomic substitutes, or phantoms when performing mandated quality assurance tests with cross-sectional modalities such as computed tomography (CT) as well as with magnetic resonance imaging (MRI). This is meant to ensure patient exposure to harmful ionizing radiation, and radiofrequency energy deposition falls within allowable limits while maintaining acceptable image quality.

A crucial component of cross-sectional imaging is the addition of intravenous contrast in the forms of organic compound bound iodine with CT, or gadolinium as is the case with MRI. Unfortunately, the IV administration of such compounds presents the potential for serious and, at times, potentially-fatal allergic reactions which radiology departments must be equipped to handle. As physicians in training, most residency programs require proficiency in basic and advanced life support (BLS/ACLS). Also, diagnostic radiology residents must receive education specific to the management of contrast-related allergic reactions.[1]

This situation provides the first, albeit foundational, opportunity for simulation. Mock imaging environments can be created with the use of pre-existing equipment or in a laboratory setting to provide a comprehensive education for radiology residents on how to approach acute emergencies, including contrast reactions, but also more broad skills such as placement of electrocardiography leads among others.[2] This type of training has shown not only to be useful to residents in training but also for fellows and even senior faculty, providing an avenue for life long learning. The frequency of such a refresher course has been suggested to be biannual rather than once a year.[3]

Simulation in Teaching Image-Guided Intervention

Perhaps the most fertile ground for simulated experience has been in the realm of procedures requiring image-guidance. There is an ever-growing weight placed on patient safety, and training programs must provide avenues for attaining proficiency without unnecessarily patient harm. Hands-on workshops have become more popular and have been made available at earlier training levels.[4] This evolution of the classic ‘master-and-apprentice’ model has had the added benefit of providing a more objective reproducible learning experience, in addition to alleviating some of the burden incurred by faculty having to serve an ever-increasing clinical burden.[5]

Given the full breadth of image-related procedures, it seems only logical to start with the more rudimentary modalities usually introduced in the early years of residency utilizing X-rays. Real-time fluoroscopy is not only used in diagnostics of the gastrointestinal and genitourinary tracts but also in needle placement with arthrography and the intrathecal sampling of cerebrospinal fluid (CSF). The main drawback of fluoroscopy would be the exposure of ionizing radiation, so one of the goals and achievements of simulation has been the quantifiable dose reduction after simulated training with phantoms[6]. An analogous modality would be fluoroscopic CT when performing cross-sectional needle placement for biopsies and drain placements. This more advanced technology has also shown to benefit from the introduction of simulation, leading to truncated total procedural time as well as dose reduction.[7][8] Perhaps the most advanced fluoroscopic technique would be digital subtracted angiography, which requires the utmost operator confidence, as the complications carry a more significant consequence. We find yet again that simulation exercises have led to a reduction of catheter misplacements in addition to a shortened overall procedural time.[9]

The ultrasound would be the next modality to which radiology resident procedural education would center around. It is safe, cost-efficient, and portable similar to C-arm fluoroscopy, however without necessitating additional safety equipment such as leaded aprons or goggles. Tissue characterization is also superior to x-ray based modalities. This allows for the identification of small abnormalities within superficial organs such as the thyroid or breast, as well as to deeper viscera such as the liver and kidney. While non-biologic phantoms offer the benefit of being reusable, they do have a limited immersion factor when it comes to simulating realistic tissue resistance.[10] More creative solutions have been offered depending on the model organ, such as the use of poultry breasts substituting for human breast and muscle.[11] A more specific simulation model is required for the replication of human-like viscera such as the liver given its characteristic anatomy, so porcine and even bovine analogs have been implemented with measurable success.[12]

Clinical productivity is frequently at odds with resident training. Given increasing volumes, the clinical landscape has drastically changed in recent times, specifically with the welcome addition of advanced practice professionals (APP) to assist in optimizing faculty efficiency. A radiological APP may have specific training in an imaging modality with more advanced procedural training as with the radiology assistant (RA), but more often, physician assistants (PA) and nurse practitioners (NP) have also found the field of radiology to be professionally rewarding and intriguing. While APP’s may not have identical backgrounds in the basic sciences as resident physicians, simulation-based training has demonstrated to be still quite helpful when focused on procedural education.[13]

Simulation in Training of Diagnostic Imaging

The thrust of radiology training mostly focuses on the interpretation of diagnostic imaging. Skills in optimizing image quality and acquisition are, however, crucial in providing an acceptable final image suitable for interpretation and are frequently overlooked. While this task is mainly delegated to modality-specific technologists, the radiologist is viewed as the one ultimately responsible for image-quality. As such, all radiologists need to demonstrate a level of familiarity with the acquisition as it pertains to troubleshooting artifacts and optimizing clarity. No other imaging modality is as operator-dependent as ultrasound.  The sonographic field-of-view is variable, and the imaging window can be affected by a variety of medical states (NPO, intubated, etc.).

A larger body habitus can have a profound effect on the acoustic window and limit utility. One of the most essential and most frequently performed diagnostic imaging tests is the transvaginal ultrasound (TVUS). It is critical in the assessment of abdominopelvic pain in the female patient, whether pregnant or not, and can diagnose various life- or organ-threatening pathology. It is also able to assess for diagnoses for which there are not ideal modality-alternatives. It is because of all these reasons that TVUS is the prime candidate for simulation education, especially before residents taking overnight call.[14][15] 

Simulation training has been compared directly to conventional education models, including slide and video presentations, but the addition of virtual reality and haptic feedback further enhances the education benefit, albeit at the cost of specialized equipment that may not be widely available.[16] Perhaps one of the most fundamental diagnostic skill is in the interpretation of chest radiography. Training modules exist for teaching the basics of X-ray interpretation that have proven useful even when applied to non-radiology residents.[17]

While computer-aided detection (CAD) has existed now for decades, it has had limited utility and has mainly been reserved for identifying abnormalities in screening mammography and screening lung CT. Artificial intelligence (AI) has future potential in aiding diagnosis on more complex modalities, but presently human perception remains the mainstay for imaging diagnosis. Even in scenarios where AI provides an initial report, human oversight over the final diagnosis will remain essential. Currently, physicians in training, as well as established faculty, suffer from an overall lack of experience with newer applications of AI. This is a potential area where simulation modules can assist in the future.