Search papers, labs, and topics across Lattice.
1
0
3
Despite their promise, even the best multimodal LLM (GPT-4o) achieves only 26% accuracy in grading knee osteoarthritis from radiographs, revealing a significant gap in clinical reliability.
