Proton therapy is currently planned using X-ray CT images for target volume delineation, as well as dose and range calculations. However, conversion of CT Hounsfield units to proton stopping power – the key parameter in proton treatment planning – is a major source of inaccuracy that results in range errors of roughly 3–5%. The use of protons to image a patient's anatomy and tumour provides a direct measure of the distribution of proton relative stopping power, reducing this uncertainty to less than 1%. In turn, this should enable more accurate tumour targeting and reduce dose deposited in surrounding healthy tissue.

The PRaVDA consortium recorded the breakthrough images at the proton therapy facility at the iThemba LABS in South Africa, using the South African National Cyclotron. "To produce these proton CT images, we built a unique medical imaging platform, which uses the same high-energy particles that are used to destroy a tumour during proton therapy," said Nigel Allinson, Project Director and Principal Investigator.

Proton images are created by measuring the energy loss as a proton beam passes through the target – in this case, the patient. To perform proton CT, it's also necessary to track the path of each individual proton, from the point where it enters the patient to the point where it leaves. The PRaVDA instrument achieves this complex procedure by using 34 layers of silicon microstrip sensors, which perform both tracking and residual energy detection.

Tests on a range of tissue surrogates demonstrated that the PRaVDA scanner delivered the lowest ever levels of uncertainty in relative stopping power measurements. "The images we have created are in fact of a humble lamb chop," explained Allinson. "But they highlight the fantastic potential for using proton CT images to aid cancer treatment in the very near future – as part of the planning process, as well as during and after treatments."

Comparing the proton CT image with a conventional X-ray CT revealed that, although the proton CT was slightly blurrier than the X-ray image, it showed exactly how protons interact with tissues – in the same way that the therapeutic protons would do. Allinson noted that image quality could be improved, for example, by using finer rotation angles or recording more proton events (subject to the need to limit diagnostic dose). "In these images, the height of slices in the CT reconstruction was 1 mm, but our sensors have about a 100 µm pitch," he added. "We can easily resolve tungsten carbide spheres 2 mm in diameter."

Allinson and colleagues are currently studying the best way to acquire images on proton pencil-beam delivery systems. Subject to funding, they then intend to install a proton CT system on a delivery system within an operational proton therapy centre.

Added imaging options

During this work, the PRaVDA team discovered that a number of other parameters can be also measured as protons pass through the target, which could be used to produce a set of complementary CT images. These include scattering and attenuation CT images, which "come free" with the tracking measurements and do not require measurement of residual energies. Another measurable parameter is straggling energy, the additional spread in exit energy due to random scattering, though Allinson notes that this is more difficult to record.

The researchers plan to exploit this finding in the next generation of their proton CT scanner. "These various parameters are related and open up of a rich area for future study," Allinson told medicalphysicsweb. "How we combine X-ray CTs and the various proton CTs in a principled manner is not known – but a whole new imaging modality of proton imaging has been opened up."

• The PRaVDA consortium, funded by a £1.8 million translation grant from the Wellcome Trust and led by the University of Lincoln, consists of five UK universities, four UK NHS Trusts and Foundation Trusts, University of Cape Town and IThemba LABS, South Africa, and Karolinska University Hospital, Sweden.

Related stories

• Proton CT could ease treatment planning
• Scientists put final pieces into place for imaging cancer with protons
• Active pixel sensors track single protons
• The potential of proton CT
• Proton detectors image beam dose