Hello

it’s only counts except for quantification examinations (xSPECT quant for Siemens, you will have Bq/ml units)

best regards

Hi, the voxel/pixel value in spect containing a total counts over the scan duration of the study. Indeed there is no info regarding the conversion factors within the dicom tag. Depending on the system nowadays the vendors start to implementing experiment based conversion factor value from the volumetric/ planar sensitivity using uniform phantom. That’s doesn’t included in the dicom tag but in the special option in the software itself so that the software can used for converting any counts to the activity concentration as per used for SUV calculation. Hope this helps you. Thanks

Syahir

You are lucky if the vendor has put that information in a 'private' dicom tag. This is done frequently until the dicom police has made a decision on encoding standards. You need to ask the vendor if this is the case, or reverse-engineer the private tags.

Good luck!

If you are interested to have an activity map in MATLAB then you could achieve it by dividing all the pixel values (you can save this info in units of counts per second) of SPECT image by the calibration factor (in units of cps/MBq). The calibration factor is the ratio of the total reconstructed counts to the true activity. Generally phantom studies are used to calculate the calibration factor in SPECT imaging.

Dear Masud, you are looking for a dicom file with pt extension (which includes MBq/cc) on a SPECT device which typically ONLY shows the Dicom SPECT format. At least three manufacturers DO create SuV (thus MBq/cc) from their recent systems as there are: GE,Siemens with XSPect and Hermes with the application SUV Spect. At that very moment you of course will need to calibrate your SPECT (CT) system according to manufacturers prescriptions first, for the used Isotope, and need to enter EXACT time after injection, kg and MBq injected to the patient. This Quantitative SPECT will than alow you to perform easy Dosimetry. More details can be given in case you like to have such in a "personal" discussion, i hope this helps, peter

I agree with Peter van Loon

Dear Masud, again we have to refer to the earlier answer by Mohd Syahir Mansor . He is completely right, the pixel value is either 8 or 16 bit , and the value is just how many counts have been sampled (over whatever period), so says nothing except indicates the relative amount versus other pixels around. Only when (as said) there is more info available (which is not the case) one could derive MBq/ml. I hope this aligns all other great answers, peter