"Our study suggests that Cerenkov luminescence is feasible and compares favourably with the other signals available for detection," said Justin Klein, first author and PhD candidate at the University of California, Davis. "Our hope is that this information will be useful for guiding future studies and informing the design of a system for imaging during surgery."

The flexibility of CLI makes it an attractive option for intra-operative imaging. It can exploit an existing range of clinically approved PET radiotracers that are tailored to detect specific pathologies. Detection of the visible photons at the tumour site also removes the need for bulky, impractical tomographic imaging systems.

Working with colleagues Gregory Mitchell and Simon Cherry, Klein carried out Monte Carlo simulations of radiation detection from a 1-mm diameter spherical tumour positioned between 0 and 5 mm deep in a 15 cm3 block of normal brain tissue. The geometry replicates the typical scenario in theatre where the obvious bulk of a tumour has been removed, but remnants remain. Simulations were repeated for several types of brain tissue, including white and grey matter and cerebellum, using optical properties in the literature.

Fluorine-18, a PET tracer commonly used for oncological PET imaging, was placed in the tumour and at realistic, lower background levels in the brain tissue. The isotope emits two annihilation gamma photons, one beta particle and 2.7 Cerenkov photons per decay event.

Two types of detector were simulated to investigate two possible clinical applications: a 1 x 1 cm non-imaging detector, analogous to an intra-operative probe, and a 15 cm x 15 cm pixelated imager. Both were positioned at the brain tissue surface.

The simulations quantified the signal reaching the detectors for each radiation type. Non-imaging detection of the tumour was quantified using a contrast-to-noise ratio (CNR) derived from the simulation results, calculated as the signal from the tumour divided by the uncertainty in the background count. Imaging detection of the tumour was assessed using the receiver operating characteristics (ROC) from a channelized Hotelling observer, an ideal numerical observer. Area-under-the-curve (AUC) statistics were derived from the ROC data.

Simulations of the non-imaging detector revealed that at depths less than 1.5 mm, Cerenkov light had a significantly higher CNR than both beta particles and gamma rays. At greater depths, however, gamma rays had a higher CNR than Cerenkov light from white matter and cerebellum, as attenuation of Cerenkov photons increased.

Using the imaging detector, Cerenkov photons were also best able to discriminate the tumour from the brain background at depths less than 2 mm. Beta particle detection only provided good discrimination at depths less than 0.5 mm. At depths greater than 2 mm, there was no significant difference between Cerenkov and gamma photons.

Though the results are positive, development of a CLI system is still in its early days and in vivo testing in humans may be up to five years away, Klein told medicalphysicsweb. "We don't want to oversell the implications of our results. They are very encouraging, but the next step will be to determine if it is actually practical."

Unlike gamma rays, Cerenkov photons can be easily absorbed and converted to charge in a photodetector. However, they deliver the least energy of the three radiation types. Klein estimates the physical signal will be at least five orders of magnitude less than ambient light. Designing a system that can measure the small signals without disrupting surgical workflow is one major challenge the group face as they work to develop CLI technology. If successful, the system will be used in pre-clinical trials.

Related articles in PMB
Quantitative assessment of Cerenkov luminescence for radioguided brain tumor resection surgery
Justin S Klein et al Phys. Med. Biol. 62 4183
Computed Cerenkov luminescence yields for radionuclides used in biology and medicine
Ruby K Gill et al Phys. Med. Biol. 60 4263
Optical properties of biological tissues: a review
Steven L Jacques Phys. Med. Biol. 58 R37

Related stories

• Guiding brain surgery with β– decays
• Fluorescent probe guides tumour excision
• Microspheres boost image contrast
• A clinical take on Cerenkov imaging