Current SOTA model in MRI

[reeered]

I am currently researching medical image segmentation, specifically focusing on MRI scans, and I'd like to learn about the SOTA segmentation models beyond nnU‑Net. My current nnU‑Net model achieves a Dice score of about 0.86, which is not sufficient for my use case. Could anyone please recommend any MRI segmentation approaches or SOTA models that can surpass nnU‑Net in performance?