Twilite Calibration

In experiments using the twilite and a PET system it is necessary to cross-calibrate the two devices. This implies the derivation of a calibration factor for the twilite that converts the coincidence count rate for a solution of known activity concentration to the kBq/cc activity concentration measured for the same solution by the PET. Note that this additionally implies that the PET has been calibrated to return kBq/cc voxel values, typically using the manufacturer’s recommended procedures and corrections (e.g. normalization of detector efficiency, system dead time correction, correction for random coincidences, scatter and attenuation correction).

The following calibration procedure is recommended:

1.PET tracer is added to water such that the activity concentration is in the range of 200-500 kBq/cc. Complete mixing should be carefully ensured.
Note that ideally the same isotope should be used for calibration as in the actual experiments. The reason for this is variability in the branching ratio between isotopes, namely the ratio of positron decays to other means of decay such as electron capture. Branching ratio is typically accounted for during the calculation of the tracer concentration by the PET system (i.e. post-processing/calibration). While the most common PET isotopes (F-18, C-11, O-15) have branching ratios above 0.95, other isotopes used in research can differ substantially, e.g. Cu-64 (0.174) and Ga-68 (0.891). Information about the branching ratio is available from Turku PET Center.
Note: Branching ratio correction is available in the PSAMPLE Correction tool.

2.A catheter identical to the type used in the actual experiment is filled with the tracer solution from the phantom. Extreme care is required to avoid the presence of air bubbles in the catheter. Note that tap water may contain large amounts of dissolved air that can result in air bubble formation in the catheter over time.

3.3) A suitable phantom is filled with the same (well-mixed) fluid. Complete mixing should be carefully ensured. For PET/CT a suitable phantom is a 250 or 500 ml plastic drinks bottle. This is readily mixed and can be well sealed. Large air bubbles should be avoided. For small animal PET scanners a 20 ml syringe, 20 or 50 ml Falcon tube may alternatively be used. For PET/MR it may be necessary to use a larger phantom for which the mu-map for attenuation correction is available. In this case the activity concentration may be reduced (lower limit 50 kBq/cc), and care should be taken that the total activity in the scanner field-of-view does not exceed the limit for noise-equivalent count rate (NECR). In all cases it is necessary to confirm that the tracer does not stick to the plastic walls of the phantom, as this will lead to incorrect determination of the activity concentration in solution.

4.After ensuring that the system clocks of the twilite acquisition computer and PET scanner are synchronized, twilite acquisition is started without the catheter guide inserted. Background counts should be acquired for at least two minutes.

5.Without stopping the data acquisition, the guide with filled catheter is inserted into the detector head, and counts acquired for at least two minutes. The calibration data curve thus resembles a step function.

6.In parallel, the activity in the phantom is measured in the PET system. Typically a protocol such as 10 minute static FDG brain is used. The data should be corrected and reconstructed in the same way as in the actual experiment, resulting in a tracer concentration in kBq/cc. It is not necessary to synchronize PET and twilite acquisitions as the PSAMPLE Correction tool performs decay/decay correction to match the scan start time.

7.Calibration processing is performed in the PSAMPLE Correction tool as outlined below. The activity concentration measured from the PET image of the phantom and PET scan start time (DICOM Acquisition Time) are required. For example, these parameters can be measured/taken from another PMOD module. The calibration factor is finally calculated by diving the PET concentration by the twilite count-rate:

F = CPET/Rtwilite

and has units (kBq/cc)/(counts/sec).

In practice, the calibration has to cope with the following challenges:

•The twilite (with LYSO detector head) has an intrinsic number of background counts due to the radioactivity in the LYSO scintillators. This background has to be subtracted from all measurements. The background can either be measured in a separate twilite acquisition without catheter in place (or with an empty catheter in place), or by generating a step function as described above.

•Continual decay of the radioactivity must be compensated for. PET systems correct this decay to the start of the acquisition. For this reason, the twilite activity is also corrected to the PET scan start time. This is particularly relevant if there is an offset between the PET and the twilite calibration measurements.

Recommendation: Although the calibration factor is stable over time, it is recommended that calibration be performed regularly as part of a quality control procedure.