PAM50 - Spinal levels under T12


(I am using sct v4.3 on Linux).
I noticed that on the PAM50 template data the spinal levels only go until T12. I wonder if you had tried lower (the lumbar and sacral sections), and if so how is it. I know that due to the variability between subjects this is quite tricky.



Hi Sergio,

Thank you for reaching out. We have not extrapolated the spinal level below T12 because as you rightfully mention, there is not a linear relationship between the levels above T12 and the levels below and we didn’t want to generate false information.

Originally, the spinal level mapping was based on information from spinal levels C3 to C8 (code here), manually-identified on twenty adult subjects; more details in [Cadotte et al. AJNR 2015]. So, it is important to keep in mind that even levels below C8 are extrapolated from actual measures, and could also be wrong.


Dear Julien,

I am coming back to this issue (which I see also coming back in several posts:

A possible solution would then be to collect spinal level information (mu, sigma) and insert that information there. I was curious if you have already gone through this.

Agreed. We have not tried this, although my intuition is that the precision of spinal level estimation would degrade as we go towards lower levels. I could be wrong though.