Pre-therapeutic (124)I PET(/CT) dosimetry confirms low average absorbed doses per administered (131)I activity to the salivary glands in radioiodine therapy of differentiated thyroid cancer

Abstract

Purpose: Salivary gland impairment following high activity radioiodine therapy of differentiated thyroid cancer (DTC) is a severe side effect. Dosimetric calculations using planar gamma camera scintigraphy (GCS) with (131)I and ultrasonography (US) provided evidence that the average organ dose per administered (131)I activity (ODpA) is too low to account for observed radiation damages to the salivary glands. The objective of this work was to re-estimate the ODpA using (124)I PET(/CT) as a more reliable approach than (131)I GCS/US.

Methods: Ten DTC patients underwent a series of six (or seven) PET scans and one PET/CT scan after administration of approximately 23 MBq (124)I-iodide. Volumes of interest (VOIs) drawn on the CT and serial PET images were used to determine the glandular volumes and the imaged (124)I activities. To enable identical VOIs to be drawn on serial PET images, each PET was co-registered with the CT image. To correct for partial volume effect and for the artificial bias in the activity concentration due to cascading gamma coincidences occurring in (124)I decay, the imaged activity was effectively corrected using isovolume recovery coefficients (RCs) based on recovery phantom measurements. A head-neck phantom, which contained (124)I-filled spheres, was manufactured to validate the isovolume recovery correction method with a realistic patient-based phantom geometry and for a range of activity concentration regimes. The mean+/-standard deviation (range) ODpA projected for (131)I was calculated using the absorbed dose fraction method.

Results: The ODpAs (in Gy/GBq) for the submandibular and parotid glands were 0.32 +/- 0.13 (0.18-0.55) and 0.31 +/- 0.10 (0.13-0.46), respectively. No significant differences (p> 0.2) in the mean ODpA between (124)I PET(/CT) and (131)I GCS/US dosimetry was found. The validation experiment showed that the percentage deviations between RC-corrected and true activity concentrations were <10%.

Conclusion: (124)I PET(/CT) dosimetry also corroborates the low ODpAs to the salivary glands. A voxel-based calculation taking into account the nonuniform activity distributions in the glands is necessary to possibly explain the radiation-induced salivary gland damage.

Fig. 1

Head-neck phantom (left) containing five ¹²⁴I-filled spheres, designated with the symbols S1, S2, P1, P2, and R, and three hollow objects simulating maxillary and sphenoidal sinuses. The corresponding CT topogram (right) illustrates the objects in the interior of the phantom cavity

Fig. 2

Isovolume recovery curves of ¹²⁴I (solid line) and of ¹⁸F (dashed line) for EXACT HR⁺ PET scanners measured with a cylindrical phantom containing radionuclide-filled spheres and rotational ellipsoids. The ratio of the prepared objects’ to the background activity concentration was 12. The data points were fitted using a three-parameter sigmoid function (solid line). The dashed line was obtained by multiplying the ¹²⁴I fitting curve with a scaling factor of 1.13. The dotted lines represent the mean submandibular and parotid gland volumes observed in this study

Fig. 3

Deviation between RC-corrected and true activity concentration in percentage of the true activity concentration for the spheres S1 and S2 (left), spheres P1 and P2 (middle), and sphere R (right) at different activity concentration regimes. The individual deviations of the three scans acquired with the clinical acquisition protocol are shown

Fig. 4

CT image (top), PET emission image (middle) 4 h after ¹²⁴I administration, and fused image (bottom) of the submandibular (left) and parotid gland (right) with ROIs (in blue) delineating the boundaries of the salivary glands. The gray (CT) and color scale range (PET) were −125 to 225 HU and 0 to 2 kBq/ml, respectively

Fig. 5

Representative examples of ¹²⁴I uptake curve for the submandibular (top) and parotid gland (bottom) in percentage of administered activity per kilogram of gland tissue. The (mass-normalized) ¹²⁴I time-uptake curve was selected as an equivalent illustration instead of the commonly used time-activity curve