Quantitative measurements bring no value and likely are misleading

[fedorov]

In the data, quantitative measurements are tied to the segmentations of the specific structures.

In the portal, quantitative measurements of the specific kind are lumped into one set of numbers across all structures.

I cannot think of a use case where this would be helpful, and I can easily see how this can be very confusing.

I suggest we remove quantitative measurements from the portal, or disable them whenever we see that there are multiple segmented structures in the current cohort.

[G-White-ISB]

I have also wondered about this.

[s-paquette]

@fedorov Regarding multiple segmented structures, what would be the best way to detect that? Adding to this, how SHOULD we really present them? We've discussed re-approaching how we do the derived data in the past, but never actually sat down to do so; maybe we should.

[fedorov]

@s-paquette we definitely SHOULD. But with the rate of issues and progress in implementation, it is hard to tell when this will happen. In the meantime, I suggest we should remove this strange feature until we can implement it properly.

[fedorov]

@G-White-ISB to follow up on the discussion today at the meeting, the problem with the current slider here

is that it shows the range of all measurements across the studies present in the current cohort selection. It does not show the measurements corresponding to the selected segments.

For example, if I select "Liver", the cohort will include all studies that contain liver segmentation. And quantitative measurements will include all those from the cohort. Those measurements will include volumes of Liver along with volumes of pancreas and heart and whatever, which I don't think can be helpful in any scenario. Because of this I think we should remove quantitative measurements.

Note that the other places where we have sliders, such as in "Slice thickness", are separate from this issue.