Project Summary
This project aims to develop a microscopic nuclear imaging instrument that can provide the 3D distribution of
clinical radiotracers within a thick, living multi-cellular organism. The proposed instrument will improve our
understanding of how clinical radiotracers are distributed at a cellular level and thereby help us to better interpret
clinical nuclear imaging. It will also allow us to test and select new radiopharmaceuticals using patient-derived
organoids for safe and effective therapy.
Microautoradiography (MAR) is the current gold-standard method for microscopic nuclear imaging. However, it
can provide a high resolution only for thin (~ 5 µm) sliced specimens using low-energy radionuclides; the
resolution is only moderate with the high-energy radionuclides used for clinical imaging, even for a thin sliced
specimen, let alone single intact cells or cell spheroids. The resolution of MAR and other existing methods rapidly
decreases with the sample thickness, because the positrons obliquely incident onto the scintillator spread more
widely as the distance between the source and the scintillator increases. The scattering of positrons in the
sample, which increases with the sample thickness, will further blur the image.
We will develop a new instrument that can address these challenges by using radioluminescence microscopy
(RLM) as a base platform. RLM is a state-of-the-art technique recording the scintillation tracks of individual
positrons. The enabling idea is to record the 3D orientations of scintillation tracks at the crystal surface and trace
them backward to the source (i.e., the radionuclide that emitted the positron) in the sample volume. To record
the 3D orientations of scintillation tracks, which decay within tens of microseconds, we will use snapshot
projection optical tomography (SPOT). SPOT was previously demonstrated with fluorescent or absorbing
specimens, which is extended here to record the 3D orientations of scintillation tracks, for the first time to our
knowledge. For a thin specimen, we can trace back the scintillation tracks using filtered backprojection. For a
thick specimen, we will investigate statistical reconstruction, which can handle the scattering of positrons in the
sample and the ill-posedness of the inverse problem.
Using radiolabeled beads, [18F]FDG-filled microdroplets, and biological specimens incubated with [18F]FDG, we
will demonstrate the feasibility of 3D microscopic nuclear imaging of a thick, living multi-cellular organism using
the same radiotracers as used in clinical imaging and therapy. In particular, we aim to demonstrate subcellular
resolution (< 1 µm) for confluent living cells and cellular resolution (10–20 µm) for thick cell spheroids.
