Multiplexed imaging of radionuclides

Abstract

Nuclear imaging provides non-invasive and near-quantitative insight into the biodistribution of radiolabelled compounds, and it does so with exceptional sensitivity and practically unlimited penetration depth. These properties make nuclear imaging highly valuable for monitoring the pharmacokinetics, biodistribution and in vivo stability of therapeutics. Moreover, the diversity of radioactive probes allows for detailed insight into cell dynamics, metabolism, epigenetics and other biological processes. However, nuclear imaging remains largely limited to single-tracer studies, or to the sequential imaging of each tracer. Tracking only a single probe or compound at a time limits the insight that can be gained. Here we discuss the applications and clinical feasibility of established and upcoming strategies for the simultaneous imaging of multiple radiotracers.

Fig. 1 ∣. Discriminating between radionuclides on the basis of photon energies.

a, Collimators enable SPECT cameras to localize γ-ray emitters by detecting their radiation. b, Energy windows help reduce interference when performing SPECT imaging. c, The use of two γ-ray cameras, each operating with a different energy window, allows for dual-tracer SPECT imaging. d, The photons emitted by different types of SPECT radionuclides can be distinguished by applying different energy windows. However, downscatter (owing primarily to Compton scattering, which results in a reduction in photon energy) can affect the results. e, An early example of dual-tracer SPECT imaging, in which ^81mKr gas and^99mTc-labelled microspheres were used to monitor lung volume and perfusion, respectively. The images show transaxial slices and their position in the raw tomographic images. f, A patient with endocarditis affecting a mitral-valve prosthesis was imaged with CT, conventional SPECT using ¹¹¹In-oxine (left column), and with dual-tracer SPECT using a CZT detector and ¹¹¹In-oxine and ^99mTc-sestamibi radiotracers (two rightmost columns). ¹¹¹In-oxine is a tracer for white blood cells (WBCs), and^99mTc-sestamibi informs on perfusion. The arrows indicate a focal point adjacent to the implant, in line with endocarditis. HLA, horizontal long axis; SA, short axis; VLA, vertical long axis. Panels adapted with permission from: e, ref. , IOP Publishing; f, ref. , Oxford Univ. Press.

Fig. 2 ∣. Using dynamic modelling to discriminate between radionuclides.

a, Two or more radiotracers can be tracked simultaneously by staggering their administration and performing several imaging sessions. Kinetic modelling is typically required to correct for radiochemical decay of the tracer and for changes in biodistribution. b, The amount of radioactivity in a region of interest can be monitored over time to generate time–activity curves, which can be fitted to determine the contribution of a certain tracer to the overall signal. Data from ref. . c,d, The PET tracers N-[¹¹C]methylpiperidinylpropionate (PMP) and [¹¹C] flumazenil (FMZ) were injected into human patients with a 20-min offset. Time–activity curves for different brain regions (c) were obtained from the imaging data and subsequently fitted to generate parametric images (d) that inform on specific pharmacokinetic parameters (k₁, k₃). Ctx, cortex; DV, distribution volume. Three different brain levels are shown, corresponding to the indicated distances between the image planes and the anterior commissure–posterior commissure plane. Panels adapted with permission from: c,d, ref. , Sage.

Fig. 3 ∣. Integration of SPECT and PET imaging.

a, Single photon emitters (orange) can be discriminated from positron-emitting radionuclides (yellow) using SPECT scanners functionalized with a coincidence detector. b, Combined SPECT–PET imaging can be achieved by integrating detectors for each modality within the same ring. In such a system, the SPECT detectors typically operate independently from the PET detectors and in anticoincidence mode, to limit scatter effects. c, A mouse injected with ^99mTc-sestamibi and [¹⁸F]FDG. Both tracers were simultaneously imaged using the YAP-(S)PET hybrid SPECT–PET system. Images show the sagittal plane. d, Installing a collimator into the gantry of PET scanners makes them suitable for SPECT imaging. e, Replacing traditional collimator pinholes with clusters of smaller pinholes with the same combined field of view helps reduce pinhole edge penetration and scatter artefacts. f, Mouse brain images acquired by simultaneous SPECT–PET imaging. The animal was injected with¹²³I-FP-CIT and [¹⁸F]FDG, and imaged 1 h later, when the ratio of both isotopes was approximately 1:1. From left to right, the columns show coronal, sagittal and horizontal slices. The numbers indicate brain regions: 1, striatum; 2, olfactory tubercle; 3, Harderian glands; 4, olfactory bulbs; 5, cerebral cortex; 6, thalamic and midbrain areas. g, A phantom representing a defective small-animal myocardium was filled with ^99mTc-MIDI and [¹⁸F]FDG and imaged on a VECTor apparatus. nonSC, non-scatter correction; SC, scatter correction; SA, short axis; VLA, vertical long axis; SIA, single-isotope imaging; DISA, dual-isotope imaging. Panels adapted with permission from: c, ref. , IEEE; e, ref. under a Creative Commons licence CC BY 4.0; f, this research was originally published in JNM, ref. , © Society of Nuclear Medicine and Molecular Imaging; g, ref. , Society of Nuclear Medicine and Molecular Imaging.

Fig. 4 ∣. Approaches for dual-PET imaging based on prompt-γ-ray detection.

a, Description of the process of prompt-γ-ray imaging. b, Brain PET imaging via prompt-γ-ray detection in mice, involving simultaneous imaging of¹²⁴I-β-CIT (dopamine transporters) and [¹⁸F]FDG (glucose metabolism). The arrows indicate the location of the thyroid (Th), the striatum (S) and the Harderian glands (H). mPET, multiplexed PET. Panel adapted with permission from: b, ref. , Annual Reviews.

Fig. 5 ∣. Compton imaging for medical purposes.

a, Schematic depiction of the Compton imaging principle, dotted lines show the Compton cone. b, Tripleisotope Compton imaging of¹³¹I-NaOH at 364 keV, ⁸⁵SrCl₂ at 514 keV and ⁶⁵ZnCl₂ at 1,116 keV in a single mouse. Representative 2D cross-sections of the signal from each isotope individually and a 3D overlap of all three radionuclides are shown. c, Compton cameras can be modified to also detect photons with relatively low energy by making a pinhole in the scatterer detector. d, A whole γ-ray imaging camera prototype comprising a circular array of Compton cameras. In this example, the scatterer detector was 20 × 5.2 cm, and the absorber detector was 66 × 21.4 cm. e, Whole γ-ray imaging can detect several photon types. f, Maximum-projection PET (511 keV) and Compton (909 keV) images of a mouse injected with ⁸⁹Zr-oxalate 23 h earlier. The absorber and scatterer detector were 214-mm wide and 52-mm wide, respectively. Panels adapted with permission from: b, ref. under a Creative Commons Licence CC BY 4.0; d, ref. , IOP Publishing; f, ref. , IOP Publishing.

Fig. 6 ∣. The principles underlying Cerenkov luminescence and secondary Cerenkov-induced fluorescence imaging (also called Cerenkov radiation energy transfer).

a, Left: a charged β-particle (red dot) travelling faster than the speed of light in the same medium polarizes the surrounding molecules. Right: as the molecules return to their ground state, the acquired energy is emitted as blue-weighted light (blue wavy lines). b, In SCIFI, the charged particles produced by radioactive decay generate CL, which in turn excites fluorophores. c, Combining a Cerenkov source with a fluorochrome can result in three types of signal: blue CL, light emitted from the excited fluorochrome (SCIFI) and the PET signal. The latter can be used as an internal standard for improving the quantitation of the SCIFI signal. Panels adapted with permission from: a, b, c (top), ref. , Springer Nature Limited; c (bottom), ref. , Springer Nature Limited.