Assessing the Quality of Multiple-Choice Questions Using GPT-4 and Rule-Based Methods