Standardisation so important for imaging
Friday 26 September 2008 Madeleine Armstrong - Senior Science Reporter

One thing attendees at the Imaging in Oncology conference held in London agreed on was the need to standardise new imaging techniques.

What is the point of using cutting-edge and expensive modalities in clinical cancer trials if differences in practices mean these trials cannot be compared?

Even within the same study, practices between centres can apparently vary extensively.

Computed tomography (CT) is already widely used to measure tumour size in cancer patients, and emerging methods include fluorodeoxyglucose positron emission tomography (FDG-PET), which looks at tumour metabolism, and fluoro-L-thymidine (FLT)-PET, which investigates tumour proliferation.

Variability within these modalities could be dangerous when using imaging to decide whether drugs should be advanced or shelved. Ed Ashton, chief scientific officer of Rochester, US-based image analysis firm VirtualScopics, said this was especially important in phase I studies, where there is a low number of patients and investigators are usually only looking for one or two responders.

Spurious positive results (such as smaller tumour size or decreased metabolism) could push an unsuccessful drug forward, while false negatives could kill off promising candidates, he explained.

"Moral duty"

There is also the ethical issue of exposing patients to high doses of radiation who probably will not ultimately benefit from the data they generate.

Dr Ashton said researchers had a moral duty to honour these patients' contributions rather than being forced to discard their data because of technical reasons. Some CT scans can deliver radiation doses higher than the US annual federal limit of 50mSv, he pointed out, which could cause secondary cancers.

Variations could also come from different interpretations of images from different readers _ one study has shown a 30% misclassification rate based on RECIST criteria, with different readers analysing exactly the same dataset.

RECIST uses X-ray, CT and MRI to measure the size of tumours and classify whether a patient has responded, and estimates tumour volume using the longest diameter. RECIST is often used as an endpoint for clinical trials, but it appears that there is still room for improvement in the application of the criteria.

Different protocols at different centres could also lead to variability. For example, with FDG-PET the fluorodeoxyglucose tracer is first injected into the patient, then the image taken after a period of waiting.

This is typically an hour, but could vary from 50 minutes at some sites to over 70 minutes at others. This could affect how much tracer is left in the body and therefore the response seen.

Many exciting possibilities are being opened up by new imaging technologies, but speakers at this conference were at pains to emphasise how important it is to apply them properly, if we are to get the best out of them.