Development of Simultaneous Multi-Radionuclide Imaging with a Novel SiPM-based Preclinical SPECT Scanner

Abstract

Multi-radionuclide single-photon emission computed tomography (SPECT) is becoming increasingly important in nuclear medicine investigations for radiopharmaceutical drug development and procedural advancement. Crosstalk is the primary challenge, where gamma-rays from one radionuclide become incorrectly attributed to the competing radionuclide(s), leading to a mixing of acquired signals. This degrades image quality and accuracy, and can negatively impact interpretation of nuclear medicine images and studies. This work presents the development of a novel crosstalk correction technique, referred to as "spectral unmixing", that separates radiopharmaceutical distributions into their respective images using spectral information during reconstruction. It was implemented at the Biomedical MRI Research laboratory (BMRL) using the Cubresa Spark silicon photomultiplier (SiPM)-based preclinical SPECT scanner.

First, the Spark's performance was characterized with one radionuclide, namely Tc-99m. This work was published in the article "NEMA NU 1-2018 performance characterization and Monte Carlo model validation of the Cubresa Spark SiPM-based preclinical SPECT scanner" in EJNMMI Physics. In tandem, open-source pinhole-SPECT reconstruction software was developed and integrated into the Software for Tomographic Image Reconstruction (STIR). The software, as demonstrated in the publication "Integration of advanced 3D SPECT modelling for pinhole collimators into the open-source STIR framework" in Front. Nucl. Med., was licensed to University College London (UCL) and is the first configurable platform for pinhole collimators. The extension of the pinhole-SPECT library from STIR to the Synergistic Image Reconstruction Framework (SIRF) established the basis for spectral unmixing. The final manuscript, "Spectral unmixing of multi-radionuclide SPECT acquisitions using the open-source SIRF and CIL frameworks", was developed in collaboration with researchers at UCL. Complex multi-radionuclide SPECT acquisitions using Tc-99m/I-123 and Tc-99m/In-111 were measured and simulated, and spectral unmixing was found to have superior image quality and quantitative accuracy compared to conventional crosstalk correction methods.

The spectral unmixing crosstalk correction methodology can be readily implemented with different SPECT systems, and its modular construction is suitable for more versatile advancements. Spectral unmixing crosstalk correction has the potential to lead to novel molecular imaging abilities and technologies, as well as accelerated studies offering unprecedented insight into the complexities of human physiology and disease progression.