I think the anodyne to this is - and I admit the degree to which this is indicative of my biases! - more data, especially early on. Getting a good baseline before you have really any significant chance of most cancers to be able to do within-individual diffs, effectively, might be a big deal.

It might also reveal that every MRI shows ghost artifacts a half a dozen times that make it longitudinally useless, of course. I'm not foolish enough to think that epidemiologists haven't thought of this.