Individual and Simultaneous Imaging of 99 m Tc and 177Lu with a Preclinical Broad Energy-Spectrum CZT-based SPECT

Abstract

Radiopharmaceutical therapy has demonstrated a high efficacy in the treatment of various tumor types. One of the radionuclides already used in the clinic is ¹⁷⁷Lu, a beta emitter that also emits several photons imageable with SPECT. Quantitative imaging of ¹⁷⁷Lu is critical for developing new radiopharmaceuticals. Energy resolution is an important factor when imaging multiple photon emissions. Solid-state detectors offer a superior performance over scintillators, that are commonly used in commercially-available preclinical SPECT scanners. This study demonstrates the feasibility of ^99m Tc and ¹⁷⁷Lu quantitative imaging in mouse phantoms, individually and simultaneously, with a SPECT prototype built with four CdZnTe (CZT) detector heads and a custom-designed and energy-optimized parallel-hole tungsten collimator. With a custom implementation of the one-step late (OSL) image reconstruction algorithm, the system is capable of imaging energies from ~70 keV to 250 keV. Above 250 keV, images were significantly affected by septal penetration, consistent with the collimator design. A recovery coefficient within 25% was obtained for activities as low as 2 kBq/mL for ^99m Tc and 45% for ¹⁷⁷Lu. Compared to a commercial NaI-based preclinical SPECT (VECTor4/CT), our prototype showed a superior energy resolution (< 5% at 140 keV), a similar uniformity with a high-compact design.

Keywords: GPU Image Reconstruction; Low activity imaging; Radiopharmaceutical therapy; SPECT; Theranostics.

Fig. 1.

a) UCSF SPECT prototype scanning the mouse phantom. b) detector IDEAS head with a mounted Tungsten collimator. c) diagram with the shape of an individual CZT pixel coupled to a collimator hole

Fig. 2.

Processes of image acquisition, system calibration, and image reconstruction design for the prototype

Fig. 3.

a) BIOEMTECH mouse phantom. b) Uniform phantom

Fig. 4.

a) System calibrations - 2D histogram depicting ADC values vs. Detector pixels samples from an acquisition of the uniform phantom filled with ¹⁷⁷Lu. b) 2D histogram illustrating detector pixel number versus calibrated energy, demonstrating the energy resolution improvement. c) Comparative analysis of a single detector head’s response to 112.9 keV and 208 keV $\gamma$-ray emissions from a uniform ¹⁷⁷Lu-filled source. At the top and right we shown line profiles depicting the sum of responses along the z and x direction for both energies; and at the bottom we show the heat map representations of CZT pixel responses, providing the sensitivity response variation of the detector for both energies. The green stripes highlight the sensitivity drops due to the smaller CZT channel size and the pink stripes are caused by the combination of this effect with the radioactive source size. d) System energy resolution across ¹⁷⁷Lu γ-ray emission energies. e) Energy spectrum of measured ¹⁷⁷Lu, 4 of the 6 main $\gamma$ -ray peaks are clearly visible at 112.9, 208.4, 249.7, and 321.3 keV. f) Maximum intensity projection (MIP) of the sum of system matrix calculated taking into consideration the system response to 112.9 keV emission of ¹⁷⁷Lu.

Fig. 5.

Pyramidal projector configuration and detector configuration scheme. The diagram illustrates the construction of the pyramidal projector using four equation planes. Focal point (red dot) is determined by the intersection of lines defined by the vertices of the CZT pixel (gold color) and opposite vertices of the collimator (blue dots).

Fig. 6.

RC and uniformity for different number of iterations for a spherical ROI located on the liver section of a mouse phantom filled with ^99mTc. The green region represents the RC STD.

Fig. 7.

¹⁷⁷Lu-filled mouse phantom showing 6 (from 71.6 keV to 321.3 keV) emitted $\gamma$ images obtained by weighted average of the slices within a 20mm thickness centred at the phantom centre. Images were reconstructed with the MRP with 20 iterations, a beta value of 0.15, and a kernel size of 3 and a voxel size of 0.8 × 0.8 × 0.8 mm³. 3D render of the 112.9 keV tomographic image.

Fig. 8.

^99mTc-filled mouse phantom imaging analysis. a) Coronal image projections at different activity concentrations. b) Recovery coefficient for different organs as a function of activity concentration c) Relative uncertainty or uniformity for each organ and activity concentrations.

Fig. 9.

¹⁷⁷Lu-filled mouse phantom imaging analysis. a) Coronal image projections at different activity concentrations. b) Recovery coefficient for different organs as a function of activity concentration c) Relative uncertainty or uniformity for each organ and activity concentrations.

Fig. 10.

Uniformity comparison between our CZT system (solid bars) and the commercial NaI multipinhole SPECT (dashed bars).

Fig. 11.

Simultaneous ^99mTc and ¹⁷⁷Lu imaging. a) Comparison of energy spectra obtained with our CZT system and the commercial NaI one. b) Images obtained with our system compared to the commercial one. Colorbar scale is the same for all the images to ease comparison. It represents the percentage of the maximum value in each individual image.

Fig. 12.

Energy spectrum of ^99mTc. The orange shading indicates energy window applied to 140 keV $\gamma$ peak