To perform quantitative PET, it's necessary to carry out accurate attenuation correction of the recorded tracer distribution – a task most commonly achieved using CT data. But in some cases, for example when using a PET/MR system, CT data are simply not available. In their prize winning PMB paper, Defrise and colleagues showed that for time-of-flight (TOF) PET scanners, this attenuation can be estimated, except for a constant, using just the TOF emission data.

The reason that this is possible, they explain, is the additional information provided by the time-of-flight measurements, which can be used to determine both the attenuation image and the emission image.

"The proof that the TOF PET data determine the attenuation sinogram up to a constant is really the key contribution, because this provides the necessary basis and confidence for applying 'CT-less' TOF PET in practice," Defrise told medicalphysicsweb. "We feel honoured and very happy to receive this prize, especially from a leading journal in this field, and one in which I published my first paper on tomography 25 years ago."

The authors developed a simple analytic method for estimating the attenuation sinogram and demonstrated the feasibility of their approach on a software phantom. Defrise notes, however, that the specific algorithm proposed in this study is likely to be superseded by iterative maximum-likelihood algorithms currently being developed.

As such, the researchers are now working with colleagues at Siemens Medical Solutions, Molecular Imaging (in Knoxville, TN) to investigate a discretized version of the same problem that accounts for the discrete sampling of the tomography data acquired by a PET scanner and also models the Poisson noise that affects these data.

"Moving towards specific applications, we also extended the method to take into account the scatter and random backgrounds and to exploit partial CT data when they are available," said Defrise.

Prize winners
The Roberts Prize is a joint award from IOP Publishing, the publishers of PMB, and the journal owners, the Institute of Physics and Engineering in Medicine (IPEM). The procedure for deciding the winner is a two-stage process. First, a shortlist of contenders is drawn up based on those papers that had the best referees' quality assessments, with a further quality check and endorsement by the Editorial Board. The shortlisted papers are then reviewed by a specially convened committee of IPEM members with fellow status and the winner is selected.

The following articles (listed in alphabetical order) were shortlisted for the 2012 Roberts Prize:

• More information on the shortlisted papers can be found here.