This article was originally published on the IPEM Website and can be found here.
It was a beautiful, sunny morning in York as around 60 delegates, speakers and exhibitors gathered at the Park Inn Hotel to discuss CT Optimisation. This area is a hot topic due to the substantial increase in the number of scans, the increase of dose from these scans and the rapid development in CT technology meaning the need for optimisation is high. This takes many forms be it image quality, dose or clinical protocol selection and all of these were discussed throughout the day by two invited speakers and ten proffered papers. There was plenty of time for discussion and some great points and ideas raised.
The first invited speaker was radiologist Damian Tolan (Leeds) to talk about what he wants from a CT and how this can be optimised in terms of dose, image quality and protocols. The talk revolved around the idea that optimisation isn’t easy and takes time and a multidisciplinary approach to achieve the best possible outcome for a patient. Talking to an audience made up almost exclusively of physicists he urged us to get in contact with our local radiologists and encourage them to work with us to reduce complexity in clinical protocols, reduce dose and maintain image quality for all patients whilst also introducing new technology such as iterative reconstruction to further improve these areas. The idea that optimisation takes time and effort was echoed by many of the day’s speakers including Mandy Moreton (Colchester) whose talk was titled “Chipping away at CT Doses – the long road to optimisation”. Mandy gave a great insight into how many factors affect the patient dose and that to truly know what a scanner is doing the manufacturer will need to be consulted and a great deal of time spent with the scanner. She showed that significant dose reduction can be achieved but it has to be carried out in a controlled manner with input from all involved parties.
Gareth Iball (Leeds) gave us examples of how different CT scanners from the same trust can give different doses and how they can be optimised against each other. The first stage involved using the same automatic exposure controls (AECs) on all scanners and the second stage was to ensure protocols used the same parameters. It was found that even after rebuilding all scanner protocols from scratch and with input from the manufacturer, doses for certain exams are still different on some scanners, a problem which is yet to be resolved. This shows that even when great care and attention is paid to what the scanner is doing, inherent differences can still be found.
Invited speaker number two was Koos Geleijns (Leiden) to give an account of optimisation of image quality. After explaining some image quality parameters that may be investigated Koos went on to describe how a human observer model had been used to assess image quality whilst changing factors that lowered dose. This was done by reconstructing datasets using different techniques rather than imaging the same patient 4 times! He then went on to describe how a mathematical model could be used instead of the human observer to predict clinical performance. This work was highly academic and I’m not convinced that radiologists are in any danger of losing their jobs any time soon.
Hugh Wilkins (Devon and Exeter) gave a very comprehensive talk on what optimisation means, why it is required from a legal perspective and some of the methods we can use to audit doses including, manual data recording and analysis, dose structured reports, of collection of RIS data. Paul Charnock (IRS ltd) later discussed the possible need for more RIS codes so audits can be carried out on examinations looking for specific things such as head for stroke or trauma heads, these may use different protocols and hence give different doses. But the limited findings he presented showed that this may not be necessary after all as no significant differences were found, but stated that it is not a simple job to get these clinical indications and more work needs to be carried out to confirm that these findings are accurate.
At the end of every session there was ample time left for discussion about optimisation, the problems that people may face and solutions that people had found to overcome some of these problems. I for one left with a sense that everyone could do more in terms of optimising the scanners they are charged with looking after. The task seems a daunting one with so many parameters that can be changed, the number of different reconstructions methods that can be employed and the fact that manufacturers are very protective over how their systems actually function. But by chipping away at the task and remembering that optimisation is a journey, not an instant result we can all get the best possible images for the lowest possible doses, from the fewest scan protocols.