Is it an important issue?

"HU values observed for a given material also depend on the dimensions of the phantom and on the positioning of the phantom, especially the presence and the type of patient support or location on/off axis. Also, a change in position of the tissue-equivalent insert in the phantom can lead to a difference of up to 80 HU."

Because the phantoms were placed on the table in a single line for the first scan but I placed them in triplet lines for the second scan.

Good morning.

I have noticed during various measurements that, on some Siemens CT scanners, the HU of a given material also depends on the algorithm used for the image reconstruction, which, in the case of Siemens scanners, is called kernel (if I am not mistaken). This also affects the image noise (std of the measurement). You should make sure that you always use the same kernel for all your measurements.

All the best for the new year,

George.

I agree with George,

Kernel is important parameter for measuring the value of HU, but if you use the same CT protocols for scanning should get the same results, because CT uses tables HU units of the calibration protocol. If all this is done without error difference may be due to re-scan the material but in the second spatial point. It is important that the material and shape of the container with an agar gel (glass, plastic, etc.), which can cause strike artifacts affecting measurement. The number of projection in one rotation is also important. When You use spiral CT you reconstruct CT slices from RAW data (acquisition slices in mm and recon slices in mm could be different) and CT use reconstruction alghoritam for this slice (wide, slim etc). Also important parameters are PITCH  and INCREMENT (overlapping of  RAW data slice) which may influence the measurement unit HU. Also, if you measure homogeneus material I suggest using ROI measurement (not only one or two pixel) with the same size and and you'll get the minimum and maximum values within the ROI as well as average and standard deviation which indicates the homogeneity of the material.

after all

I assume that the problems in a different place can begin scanning and shape and material of the container for a gel which may make it strike artifacts (if you can see in the picture, change Windov/Level parameters).

Best regard

Frane

The reproducible may due to CT machine or within the analysis you do using gel phantom. The reproducibleof the machine based on the variation of the tube current and independent of HV. So you may study the output dose by a standard teqinec befor your use of gel phantom to get the repruceabilti due to th CT machine

The reproducible may due to CT machine or within the analysis you do using gel phantom. The reproducible of the machine based on the variation of the tube current and independent of HV. So you may study the output dose by a standard teqinec befor your use of gel phantom to get the repruceabilti due to th CT machine

The problem is I am using totally the same CT scanner and same protocol (also same gels). Same kVp and same slice thickness... I don't think the container affect the results. I will try to use AEC off scan. Thank you for the suggestions. 

Ahmed, I am trying to characterize my gel phantoms. That's why, I want to be sure about their HU units, I don't deal with dose or any parameter for the scanner.

Happy new year, thank you for the good wishes.

Regards,

Gizem

Dear Gizem

I suggested this to be sure that the variation is due to the gel phantom and then you take in your consideration the contribution of the reproduceabilty of the machine if any.

Also if the standard deviation is within acceptable value it will be ok.

Don't forget that site of measurement of density is crucial.

Measurement done at different distance from center of phantom prodices different value (artifact induced by beam hardening),