3d CAM or visualization of activations for 3d data

Hi all, I am looking for a way to visualize 3D activations from a classification based network. If I try to use CAM, I would get one attention map for for every one of the slices of my input tensor (3d medical data) but that’s not really indicative of which slice and which spatial co-ordinates activated the classification label. This may be an open research question but I’m wondering if there is any way to visualize this that is common in the community - any help appreciated!

PS - I checked out the most relevant topic here in this post which only tried to visualize filters but thats not my exact use case.