Investigation of an Ex vivo Perfusion Model for Teaching Angiographic Procedures to Novice Trainees

Current training models for angiographic procedures are limited to expensive simulations, simplistic models, and animal studies. This study hypothesized that trainees allowed to partake in a hands-on training session with an ex vivo perfusion model would result in superior subjective and objective improvements in procedural indicators compared to an introductory lecture alone.

Materials and Methods:

This project was granted IRB exemption at our institution. Twelve volunteer medical students took a survey assessing confidence and familiarity with embolizations and then viewed a lecture covering angiography and embolization techniques. Volunteers then watched a physician perform an embolization on the model using a porcine kidney obtained from a local butcher perfused with normal saline. A randomly selected experimental group (n=6) then performed an embolization on the model during a practice session. The control group (n=6) did not have the opportunity to practice. One week later, each participant was then given a target subsegmental vessel on the model and performed an embolization independently. A physician recorded the following variables: total procedure time, non-target embolization, delivery to contrast stasis, and instrument familiarity. A post-survey assessed the utility of hands-on practice and each participant’s confidence and familiarity with the procedure. Survey results and procedural factors were compared for study group differences.

Results:

Average procedure time was 21.2 + 3.8 min for experimental group (n=6) and 27.5 + 6.0 min for control group (n=6) (p=0.026, one tail T-test). The control group caused non-target damage in 5/6 cases compared to 1/6 in the experimental group (p< 0.01, 2 proportion test). All participants delivered embolic to the point of contrast stasis. Physician evaluation rated the control group as being less familiar with the materials, requiring guidance to complete the procedure. Comparing experimental and control groups, respectively, there was a 218% + 39% versus 171% + 69% increase in confidence for selecting vessels and 260% + 39% versus 193% + 58% increase in confidence performing angioembolization (p=0.18, 0.06, two tailed t-tests).

Conclusion:

Subjective and objective results demonstrate the utility of an ex vivo perfusion model as an inexpensive and valuable training tool that simulates the tactile feedback of an IR procedure. An ex vivo perfusion model for IR education may serve as an effective, low risk option to decrease use of live animal models while improving trainee skill and confidence in a low-risk setting.