Radioiodine uptake measurement on planar scintigraphic images: an automatic process reducing thyroid volume effect
Résumé
Purpose
Dosimetry-based therapy of benign thyroid diseases requires the assessment of the radioiodine uptake (RIU) at least at one time-point before therapy. For that purpose, a planar scintigraphic image can be acquired with a gamma-camera after administration of a tracer activity of I-123. The aims of this systematic study is to investigate the accuracy of a standard RIU calibration method using a clinical routine phantom compared with a set of five realistic thyroid phantoms.
Materials and methods
The planar images were acquired with two gamma-cameras, Siemens Symbia T2 and S, with LEHR collimators. The thyroid phantoms, with volumes ranging from 3 to 30 mL and the routine phantom were filled with I-123 solution (10 MBq). The phantoms were imaged at two measuring distances (20 and 30 cm). In this systematic study, six measurements of reproducibility were performed for each geometry. The influence of the following parameters has been evaluated: distance, phantom type, thyroid volume, image processing and calibration protocol. A new automatic method, independent of the region of interest (ROI) chosen by the user was evaluated.
Results
The standard calibration procedure generally underestimated the thyroid activity when compared to the realistic phantoms calibration factor. Precisely, at 30 cm, the difference in activity was at most 12% for Symbia T2 and 20% for Symbia S. Experiments showed limited impact of thyroid volume (<6%) and distance on uptake assessment but the standard image processing method influenced greatly RUI estimates. Therefore, the systematic study enabled implementing a new automatic procedure based on a fixed thresholding of planar images. For I-123, an optimum threshold, independent of the gamma-camera and measuring distance, and minimizing the thyroid volume effect was obtained with value of 10%. Using this image processing, the bias on measured activity was 3% for Symbia T2 and 2% for Symbia S.
Conclusion
In the investigated clinical practice, the RIU was often inaccurately determined at any thyroid volume. An underestimation of 20% of the uptake indicated that the therapeutic activity given to patients could have been reduced by that much. The automatic calibration method considerably reduced the thyroid volume effect. Currently, this method and realistic thyroid phantoms are used in a French multi-centric study to evaluate the accuracy of thyroid uptake calibration. In a future work, the quantification of RIU for patients with multinodular diseases will be studied with a new pathologic phantom including two nodules enabling to mimic hyper- or hypo-uptake.