PET with 2-fluoro-2-deoxyglucose (18F-FDG), a radioactive tracer that mimics the action of glucose, is becoming a routine tool for cancer diagnosis. Because PET reveals the metabolic activity within a suspicious mass, it can show doctors which areas might be most vulnerable to a particular treatment, and can also provide a better insight into the extent of the cancer. These features have led to PET being increasingly used for planning radiation-therapy treatments.

With the CT or MRI images that are usually used for treatment planning, the anatomical contours of tumours and organs are clearly visible. These contours are marked, either manually or by the treatment-planning software, and used as the basis for designing an intensity-modulated radiation-therapy (IMRT) treatment. "Most treatment-planning packages have only rudimentary tools for quantitatively analysing images," explained Edward Graves, an assistant professor of radiation oncology at Stanford.

PET images, however, don't provide much spatial information so they need to be quantitatively analysed before the tumour can be segmented from the other tissues. Since conventional treatment-planning systems aren't cut out for this, at the moment there is no standard method for using PET images in treatment planning.

This less-than-ideal situation motivated the Stanford group to develop a dedicated software program for quantitatively analysing PET images. The researchers hope that this will make life easier for radiation oncologists and help speed up the acceptance of PET as a radiation-therapy planning technique. "Incorporation of PET into clinical radiation oncology for [treatment-planning] applications requires robust and practical methods for delineating PET-based target volumes," the team stated in Technology in Cancer Research and Treatment (TCRT 6 111).

The researchers are confident that RT_Image has the necessary features to help achieve this. It is a stand-alone image display and analysis tool that can be added to any DICOM-compatible clinical data management platform. And although it was developed with PET in mind, the software can be used to evaluate images from most modalities.

"The goal of RT_Image is to provide a platform on which a wide variety of candidate techniques can be implemented and evaluated," said Graves. Because it is free and open-source, researchers can use it to implement experimental image-processing methods and add new features, and can share these developments quickly and easily with the rest of the research community. "At this stage, it was my feeling that a widely available research tool would be more useful than a commercial product," he added.

Stanford radiation oncologists and radiologists have been using the software in the clinic since 2004. Because it is DICOM-compliant, it was easily integrated into the radiation-therapy planning workflow, and so far the clinicians have successfully analysed more than 200 PET data sets. "The potential benefits stem largely from an improved ability to interpret functional imaging examinations," explained Graves. "This can take many forms, from the aforementioned improved technique for defining radiotherapy target volumes, to optimizing patient disease staging."

• RT_Image can be downloaded here.