A research team headed up at Dartmouth College (Hanover NH) is developing a new method for quantitative imaging of tissue pO2 during external-beam radiotherapy (EBRT). The technique – called Cerenkov excited phosphorescence oxygen (CEPhOx) imaging – is based on phosphorescence lifetime imaging of oxygen-sensitive probes (Phys. Med. Biol. 59 5317).

The megavoltage radiation beams used to deliver radiotherapy induce Cerenkov emission in tissue. This emission acts as an internal source to excite the phosphorescent probe PtG4, which emits near infrared light with a lifetime related to the local oxygen concentration. This relationship is described by the Stern-Volmer equation, which defines pO2 as a function of the phosphorescence lifetime in a deoxygenated medium, the lifetime at pO2 and an oxygen quenching constant.

"Tumour pO2 is an important measure of the effectiveness of EBRT," explained researcher Robert Holt. "Since EBRT is often delivered in multiple fractionated doses, it would be valuable to know whether or not the treatment will be effective after the first dose, instead of taking a wait-and-see approach."

During treatment, PtG4 phosphorescence leaving the patient is measured by an intensifier-gated CCD camera synchronized with the linac radiation pulses. CEPhOx imaging exploits the multiple incident beam angles used for EBRT to enable tomographic reconstruction of phosphorescence. By including prior information regarding the beam direction, tumour size and location, and the exponential decay of phosphorescence, it's possible to determine the oxygen distribution in the tumour region.

Phantom tests

To verify their CEPhOx scheme, Holt and colleagues examined a cylindrical liquid phantom containing PtG4 at 1 µM concentration and housing a smaller cylindrical inclusion containing 5 µM PtG4. The probe concentrations reflect realistic, non-toxic values and the expected preferential probe accumulation at a tumour site. The inclusion was fully oxygenated for a first experiment and fully deoxygenated for a second.

The researchers irradiated the phantom with 6 MV, 3.25 µs radiation pulses at 20 incident angles and measured the resulting phosphorescence at delays of 3.2 to 203.25 µs (in 5 µs increments) after each pulse. For each delay time, signals from 100 radiation pulses were accumulated into one image frame.

They then separated the imaging domain into two regions: the bulk of the phantom and the inclusion; this separation enabled image reconstruction using a simple least squares method. In a clinical situation, this "image guidance" is based on the imaging performed prior to treatment planning. "This anatomical prior information is already part of the EBRT workflow, so we're not asking for more imaging," said Holt.

Plotting the phosphorescence intensity in each region versus the delay time allowed the researchers to determine phosphorescent lifetimes. For the oxygenated inclusion experiment, the recovered lifetimes were 23.9 and 26.3 µs, for the inclusion and background regions, respectively. The deoxygenated experiment showed phosphorescent lifetimes of 41.2 and 21.3 µs, for the inclusion and background, respectively.

The lifetime values were used to calculate a time-resolved phosphorescence map, which then allowed recovery of the pO2 distribution using the Stern-Volmer model. For the oxygenated experiment, pO2 values were 113 and 92.4 mmHg, while the deoxygenated experiment showed pO2 values of 15.4 and 142 mmHg, for the inclusion and background, respectively. The results demonstrated that region-averaged pO2 values were recovered successfully, and that oxygenated and deoxygenated inclusions could be differentiated.

Brain tumour imaging

Finally, the team performed a simulation study to examine the possibility of CEPhOx monitoring during EBRT of human brain cancer. They used a representative human head MRI and added an elliptical tumour target (ranging in size from 5 x 7.5 mm to 20 x 30 mm) at edge depths from 5 to 25 mm.

The researchers divided the imaging domain into tumour, scalp and normal brain tissue, and assigned the probe concentration ratio as 4:1 for tumour-to-normal brain and 5:1 for scalp-to-brain tissue. They applied the same treatment used for the phantom to the simulation and recovered the phosphorescence distribution as a function of time.

Recovered pO2 values for the scalp and bulk tissue did not show significant errors. In the tumour, however, simulated data showed increasing error in oxygenation recovery with increasing tumour depth and size. The researchers propose that CEPhOx recovery could be effective for pO2 estimation in tumours of all sizes, as long as they are located at edge depths of less than 10-20 mm (depending upon the oxygenation value). With these constraints, the oxygen recovery error was within 15 mmHg.

The authors conclude that the CEPhOx method can provide non-invasive monitoring of tumour pO2 during EBRT. They point out that this method only requires the addition of a non-toxic phosphorescent tracer and the presence of a camera in the treatment room, making it compatible with established radiotherapy workflow.

The team is now testing CEPhOx imaging on rats bearing brain tumours. "We injected the probe directly into the tumours and measured the luminescence lifetime while the tumour was irradiated by an arc delivery of small stereotactic radiation beams," co-first author Ronxiao Zhang told medicalphysicsweb. "Initial results agree with the normal levels of oxygenation for small animals. However, the signal is weak and noisy, partly because the treatment region is much smaller than that in large animals and humans. The next step is to expand to different tumour types and reconstruct the pO2 distribution in vivo."

Related articles in PMB
Cherenkov excited phosphorescence-based pO2 estimation during multi-beam radiation therapy: phantom and simulation studies
Robert W Holt et al Phys. Med. Biol. 59 5317
Superficial dosimetry imaging of Cerenkov emission in electron beam radiotherapy of phantoms
Rongxiao Zhang et al Phys. Med. Biol. 58 5477
Projection imaging of photon beams using Cerenkov-excited fluorescence
Adam K Glaser et al Phys. Med. Biol. 58 601

Related stories

• Cerenkoscopy tested on human patient
• Tracking oxygenation during radiotherapy
• Luminescence imaging tracks hypoxia
• Blood monitoring predicts therapy efficacy