Stable Disease – A Problematic Endpoint

Lately I’ve noticed an increasing number of papers, meeting abstracts, news reports, and (worst of all) self-congratulatory press releases which more or less treat stable disease as evidence the treatment in question is showing benefit or promise in the treatment of advanced cancer. Some papers and abstracts combine stable disease with objective response into a “clinical benefit percentage”, and some press releases play up this percentage, as evidence the company’s treatment is promising, even when the majority of the cases of alleged benefit are stable disease. Now of course, if a treatment can really halt the progress of the disease for an appreciable time, it’s a real benefit. Moreover, some new treatments might only be expected to halt the progress of the disease rather than shrink the tumors. This is true of some anti-angiogenic treatments which are designed to stop the development of new tumor blood vessels and thus stop the tumors from growing. With the intense interest in anti-angiogenic therapies, it’s no surprise there’s more interest in stable disease as a clinical outcome. So what’s the problem here?

When someone’s tumors undergo major shrinkage (partial response) due to treatment or go away entirely (complete response) it’s relatively easy to tell. And since cancer only rarely goes away by itself, any appreciable percentage of objective responses in difficult to treat advanced cancer indicates promise even in an early study. When this happens, there’s no real question something is happening, though the durability of these responses and how it will compare to standard treatment are still critical questions.

Preliminary reports of stable disease are usually based on very short follow-up. Often only 8-12 weeks. Slow growing cancers can easily appear to be relatively stable over short periods, and some types of cancer not infrequently have natural periods of relative stability.

The definition of stable disease means no new tumors appear and (roughly) that there is little change in the size of the known tumors. But it does not mean that there is no change. In fact, a small amount of growth over baseline is still considered to be stable disease. This is entirely appropriate since there is some uncertainty in all measurements. But it also means that over a short period of time continued growth can still be called stable disease. As you can see, given this and the often varying natural history of cancer in many cases, it would be no surprise to find a percentage of patients who are stable by the official criteria, especially over a short period, even if the treatment doesn’t work at all.

This doesn’t mean that it’s never possible to see promise in stable disease. In a preliminary, uncontrolled study, a really high percentage of patients with ongoing stable disease after a significant period of time is a reasonable suggestion of promise (I’d say at least 6 months, with a year being much better for most cancers, but it helps to know something about the natural history of your type of cancer). I also think it’s still questionable if only a few patients are stable for an extended period. This might simply represent the extremes of the natural distribution of disease growth rates. More convincing from the point of view of the evidence (Even if less appetizing to patients who don’t relish being randomized), is confirmation of stabilized disease in a randomized study with Time To Progression compared to control as the end point. The unfortunate reality of all of this is that it takes more time to obtain anything resembling a reliable indication of efficacy based on this weaker endpoint than is the case for stronger endpoints like objective response, especially for uncontrolled studies.

One other thing to watch out for is just how the duration of stable disease is described. Very often preliminary reports give the duration of every objective response, along with whether it is still ongoing, but only specify a minimum duration of stable disease (or at best a range) without indicating how long it lasted (and whether it’s ongoing) in each patient. Unfortunately, this just isn’t enough information to tell whether anything’s happening or not.

Finally, don’t distrust research just because stable disease is reported. It’s both traditional and appropriate to report the current outcome of the patients who were treated, including both apparent stable disease and progressive (worsening) disease. It’s only when claims of benefit or promise are made or implied that you should be wary.

Further Reading

For the technical details of how stable disease, and partial and complete response are defined, see the National Cancer Institute’s RECIST Criteria.

This CancerGuide Page By Steve Dunn. © Steve Dunn
Page Created: 2002, Last Updated: May 25, 2003