Generative Pre-trained Transformer (GPT)-assisted multimodal visual language model (VLM) diagnostic support