Activity quantification combining conjugate-view planar scintigraphies and SPECT/CT data for patient-specific 3-D dosimetry in radionuclide therapy
- 491 Downloads
Three-dimensional dosimetry based on quantitative SPECT/CT has potential advantages over planar approaches, but may be impractical due to acquisition durations. We combine one SPECT/CT with improved quantification of multiple planar scintigraphies to shorten acquisitions.
A hybrid 2-D/3-D quantification technique is proposed, using SPECT/CT information for robust planar image quantification and creating virtual SPECTs out of conjugate-view planar scintigraphies; these are included in a 3-D absorbed dose calculation. A projection model simulates photon attenuation and scatter as well as camera and collimator effects. Planar and SPECT calibration techniques are described, offering multiple pathways of deriving calibration factors for hybrid quantification. Model, phantom and patient data are used to validate the approach on a per-organ basis, and the similarity of real and virtual SPECTs, and of planar images and virtual SPECT projections, is assessed using linear regression analysis.
Organ overlap, background activity and organ geometry are accounted for in the algorithm. Hybrid time-activity curves yield the same information as those derived from a conventional SPECT evaluation. Where correct values are known, hybrid quantification errors are less than 16% for all but two compartments (SPECT/CT 23%). Under partial volume effects, hybrid quantification can provide more robust results than SPECT/CT. The mean correlation coefficient of 3-D data is 0.962 (2-D 0.934). As a consequence of good activity quantification performance, good agreement of absorbed dose estimates and dose-volume histograms with reference results is achieved.
The proposed activity quantification method for 2-D scintigraphies can speed up SPECT/CT-based 3-D dosimetry without losing accuracy.
KeywordsRadiobiology/dosimetry Radionuclide therapy SPECT/CT Conjugate views Activity quantification
The authors wish to thank Prof. Felix Mottaghy for access to the scanning facilities at the University Hospital Maastricht; furthermore, Dr. John Humm for making patient data available for use in this work.
Conflicts of interest
Yannick Berker was a diploma student with Philips; Andreas Goedicke is an employee of Philips; Bernd Schweizer was an employee of Philips.
- 6.Loevinger R, Budinger TF, Watson EE. MIRD primer for absorbed dose calculations, revised. New York: Society of Nuclear Medicine; 1991.Google Scholar
- 13.Ferrer L, Malek E, Bodet-Milin C, Legouill S, Prangère T, Robu D, et al. Comparison of dosimetric approaches for non-Hodgkin lymphoma in RIT [abstract]. Eur J Nucl Med Mol Imaging 2010;37 Suppl 2:S238.Google Scholar
- 16.Schweizer B, Schaefer A, Donsch P, Kremp S, Gouverneur E, Farmakis G, et al. Evaluation of voxel-based dosimetry for targeted radionuclide therapies in phantom studies [abstract]. Eur J Nucl Med Mol Imaging 2009;36 Suppl 2:S428.Google Scholar
- 17.Segars W. Development and application of the new dynamic NURBS-based cardiac-torso (NCAT) phantom, University of North Carolina at Chapel Hill; 2001.Google Scholar
- 20.Schweizer B, Goedicke A. Validation of GEANT 4 for accurate modeling of 111In SPECT acquisition. In: Deserno TM, Handels H, Meinzer HP, Tolxdorff T, editors. Bildverarbeitung für die Medizin 2010; 2010 Mar 18–20; Aachen, Germany. Berlin: Springer, 2010. p. 345–9. Available via: http://ceur-ws.org/Vol-574/bvm2010_70.pdf.