Congress:
ECR25
Poster Number:
C-26871
Type:
Poster: EPOS Radiologist (scientific)
Authorblock:
A. Bose1, Z. Q. Lee2, B. Reid3, S. Feng4; 1Auckland/NZ, 2Brisbane/AU, 3Palmerston North/NZ, 4Melbourne/AU
Disclosures:
Aabharika Bose:
Nothing to disclose
Zhi Qi Lee:
Nothing to disclose
Ben Reid:
Nothing to disclose
Sijing Feng:
Nothing to disclose
Keywords:
Anatomy, Artificial Intelligence, CT, eLearning, Education and training
MCQs generated with the assistance of a customised LLM was successful in replicating RANZCR style phase 1 anatomy MCQs in several domains. Registrars evaluated it highly in terms of using it as a study resource and recommending it to their peers for examination preparation. It is efficient in terms of question creation thereby reducing barriers and enriching availability of good quality exam practice materials. In the future, the question generation system can be reviewed and modified to provide the actual board examination questions, significantly reducing the time and effort required by volunteer radiologists manually creating these questions.