That additional overhead is highlighted in a recent study that retrospectively examined data from 803 patients treated with IMRT (JNCI 5 300). The research team, led by Indra Das from the University of Pennsylvania (Philadelphia, PA), examined the variation in IMRT dose prescription, treatment planning, dose recording and dose delivery among cancer patients. The analysis revealed substantial variation in prescribed and delivered doses among medical institutions, with sites employing different IMRT planning software coming up with dramatically different treatment plans. When comparing the prescribed dose to that actually delivered to the target volume, the researchers concluded that: "In IMRT, the prescribed dose rarely corresponds to the planned, or delivered, dose."

Several factors underlie these findings. For starters, there's the complexity of the IMRT system, not to mention challenging dosimetric goals such as assigning dose limits to multiple critical structures that it's expected to meet. Another significant influence is the wide range of commercial IMRT planning systems that are available, all of which are based on unique algorithms. Add to this the variation in target-volume delineation between physicians, and a picture emerges in which IMRT standardization becomes pretty tough.

Looking at Das's results, the situation appears far from ideal. The patients in the study were treated for brain, prostate and head-and-neck cancer at one of five US medical institutions, all of which used a different IMRT treatment-planning system. The researchers noted a wide variability from the prescribed dose for all disease sites and all treatment-planning systems. For example, 46% of patients received a maximum dose at least 10% higher than the prescribed dose, while 63% of patients received a dose to the tumour site of more than 10% lower than the prescribed dose.

Despite all of its advantages, IMRT is inherently an inexact art and thus some variation from the prescription is not unexpected. Its very nature demands a level of compromise - for example, it's not uncommon to accept lower tumour coverage to limit the dose to organs at risk. And, as Das and co-authors point out: "In IMRT, one usually does not get the exact dose distribution as prescribed, but a good treatment planner can get fairly close to the desired initial goal." Problems arise, however, because the variation in IMRT treatment plans between medical institutions (whether due to differences in the treatment-planning system or in the individual work regime) makes it near impossible to draw valid conclusions from results of multicentre clinical IMRT trials.

Harmonized approach
So what is the best way to standardize IMRT planning and reduce these variations between the prescribed and delivered dose? Das and co-authors reckon the answer lies in establishing national and/or international guidelines for IMRT, a task that they acknowledge will take time and effort from practising physicians and physicists. In an accompanying editorial (JNCI 5 288), John Willins and Lisa Kachnic of Boston Medical Center and Massachusetts General Hospital concur that greater harmonization of methods for IMRT dose specification and reporting is needed. "Widespread use of IMRT planning standards would not only facilitate multicentre clinical trials but would also provide clinicians with solid guidance in their everyday practice on the question of what constitutes a 'good' IMRT plan," they write.

How best to define such IMRT planning and reporting standards is another matter. Willins and Kachnic propose setting a limit as to how much radiation above the peak dose is acceptable. This limit is generally taken as 10% in clinical IMRT, but this figure is not dictated by clinical data or widely accepted guidelines. As for how to measure plan quality, Das suggests that the median dose could be used for dose reporting in IMRT. Willins and Kachnic, on the other hand, believe that neither the isocentre, median or minimum dose are dependable indicators of plan quality, and instead prefer indicators such as the percentage target volume encompassed by the prescription isodose surface. They also point out that "any standardization should be done in a way that prioritizes clinical utility and preserves flexibility for the practitioner to encourage acceptance."

It may turn out that setting the guidelines for IMRT dose specification proves as complex a task as the treatment-planning process itself. But greater synchronization of prescription and reporting standards for IMRT plans can ultimately only benefit patients. As such, it's a task that calls for dedicated attention from the radiation oncology community as a whole.