Calibration for optimised lung parenchyma analysis


Chest analysis is more clinically accurate if the hounsfield units are calibrated prior to analysis, is this possible or already built in to the CIP (specifically the parenchyma analysis module)?

Eg. air outside patient set to -1000HU and blood in aorta set to 50HU



That is an excellent suggestion. Unfortunately, we don’t currently have an autocalibration approach within CIP to perform a linear correction of HU values based on reference values. We will incorporate this into our wish list of features.

In the meantime, you could use a combination of the editor and label statistics module to gather mean values in the trachea and aorta to perform a correction.



Thank you for adding this to the wish list. Just an update that may help any developers - when I tried doing the calibration manually it didn’t give a linear difference (this may not be the correct terminology). For example the tracheal air would be 14 HU more than -1000 and the aortic blood would be 63 HU more than 50. Therefore I couldn’t simply adjust by adding/subtracting.
There would also be different differences between various slices from the same patient.
So it may be most accurate and reproducible to segment internal sections of the trachea and aorta for calibration rather than one slice.