Sure, it can see obvious stuff in images, but as far as I'm aware it is not designed for (or tested on) performing the kind microscopic analysis that radiology involves

Throw a chess board on there. See how it does. It always gets pieces and positions completely wrong because it’s terrible at analyzing images.

Claude is the worst FM at image understanding. Prior to gpt-5.4 the only usable models were Gemini and Qwen.