3486 - AI-Generated Image-Based Mnemonics: Enhancing Learning Efficiency in Radiation Oncology

Purpose/Objective(s): Image-based mnemonic resources, such as SketchyMedical^©, have become popular in medical education and use the “method of loci” memorization technique to effectively promote retention. We used AI image generators to create a similar platform for radiation oncology (RO). Here we report the results of our survey evaluating residents’ opinions of the platform.

Materials/

Methods: Five images were generated using the AI software DALL-E^© and Midjourney^© and refined with editing software. The mnemonics symbolized the RAPIDO and UK Fast Forward trials, treatment eligibility for accelerated partial breast irradiation, OAR constraints for conventional CNS and H&N treatments, and low energy photon interactions. A 15 question survey was made using a 5-point Likert Scale and sent to 648 PGY1-5 RO residents to assess educational experience and perceived usefulness of the mnemonics.

Results: 41 responses were recorded. Median age was 31 (IQR=4) years. Training level comprised of 2.4% PGY1, 24.4% PGY2, 19.5% PGY3, 31.7% PGY4, and 22.0% PGY5.

Full results are shown in Table 1. All respondents were familiar with image-based mnemonics and 82.9% used them in medical school. Respondents strongly or somewhat agreed that memorizing details related to clinical trials (87.8%), treatment regimens (87.8%), OAR constraints (87.8%), and physics concepts (82.9%) is time-consuming. They found the mnemonics helpful for learning clinical trials (70.7%), treatment regimens (70.7%), OAR constraints (63.4%), and physics concepts (58.5%). Most (87.8%) wished this platform previously existed, and 65.9% could envision using it. 58.5% felt the detail in the mnemonics was sufficient and 92.7% felt they were most useful when accompanied by a story. 53.7% could see themselves contributing to our platform. Written feedback suggested decreasing text and adding consistent symbols across images.

Conclusion: Our findings strongly suggest residents in radiation oncology endorse the efficacy of AI image-based mnemonics and express a desire for their integration into educational frameworks.