In principle, proton therapy's main benefit over other types of radiotherapy, such as X-ray and electron therapy, is that the protons have a very finite range, and can therefore be focused on malignant tumours without overly affecting surrounding healthy tissue. In practice, however, it is difficult to pinpoint where this sharp cut-off occurs.

Radiographers need to know exactly how much tissue the beam passes through before reaching the tumour, and what the stopping power of that tissue is. Such stopping power can be calculated from CT scans, but these leave an uncertainty of a few percent. The result is that radiographers currently have to employ a margin of error, and direct the proton beam so that crucial healthy tissues lie to the sides rather than to the beam's end.

There have been attempts to improve the ranging accuracy. Using PET, for example, scientists have been able to detect the gamma rays generated in the decay of positron- emitting isotopes that are themselves generated when protons irradiate tissue. These "secondary" gamma rays give an indication of where the protons stop, but it is not a precise indication because the radioactive isotopes often diffuse to different locations before their gamma rays are emitted. Another idea is to look for the gamma rays that are emitted when atomic nuclei, excited by protons, fall back into their original state. These "prompt" gamma rays appear on much shorter timescales, and therefore have the potential to provide accurate range information.

Now, Joost Verburg and Joao Seco, from Harvard Medical School and Massachusetts General Hospital in Boston, MA, have demonstrated a system for proton range verification using prompt gamma rays. They placed a collimator and a gamma-ray detector, the efficiency of both of which could be established, beside a phantom tissue sample. They irradiated the phantom with a proton beam, and then measured the resultant prompt gamma-ray spectrum. From this spectrum, using computer models, they could work out suitable values for the depth of irradiated matter and its composition.

The researchers found that they could determine the absolute range of the proton beam to a standard deviation of 1.0–1.4 mm. "We anticipate that improving detection efficiency by an order of magnitude is feasible for a full-scale system, which brings the required dose for millimetre accuracy well within clinically useful levels," they write.

Improving accuracy of proton therapy would be advantageous for the treatment of all cancers. But, says Verburg, some cancers might benefit more than others. Lung tissue, for instance, is not very dense, and so an uncertainty in beam-falloff for lung cancer can be problematic. Verburg and Seco's method could also help the treatment of prostate cancer, where proton beams currently have to be carefully directed so that they do not hit the rectum.

"We believe the results of our experiments show that the method is promising for clinical use, but the range problem certainly has not been fully solved yet," says Verburg. "We are working on the design of a full-scale prototype system for clinical trials and we anticipate further development of the method for clinical use. The adaptation of treatment plans based on range measurements is also an important aspect that needs further study."

"Routine use of prompt gamma-ray detection to improve proton therapy treatments could become a reality in a few years from now," Verburg adds.

Related articles in PMB
Proton range verification through prompt gamma-ray spectroscopy
Joost M Verburg and Joao Seco Phys. Med. Biol. 59 7089
Energy- and time-resolved detection of prompt gamma-rays for proton range verification
Joost M Verburg et al Phys. Med. Biol. 58 L37
Measurement of characteristic prompt gamma rays emitted from oxygen and carbon in tissue-equivalent samples during proton beam irradiation
Jerimy C Polf et al Phys. Med. Biol. 58 5821
In vivo proton range verification: a review
Antje-Christin Knopf and Antony Lomax Phys. Med. Biol. 58 R131

Related stories

• Acoustic waves characterize proton beams
• Resolved gamma lines verify proton range
• Resolvable gammas boost range verification
• Can prompt gammas monitor in real time?