Comparative Performance of Seven Mainstream Large Language Models on the 2022 American College of Radiology Diagnostic Imaging In-Training Examination | Synapse