A ‘STEPP’ Toward Personalized Medicine

Using a powerful new statistical tool, biostatistician Richard Gelber, PhD, and his colleagues say they can find statistical patterns within clinical trial outcomes that predict how different groups of patients will respond to drugs.

Gelber, a member of Dana-Farber’s Department of Biostatistics and Computational Biology, explained that the analytical tool helped researchers determine which patients with early breast cancer would do better with various combinations of estrogen-blocking compounds.

“It is a methodology that allows you to extrapolate from one-size-fitsall to a more reasoned, personalized approach in deciding what treatments to choose,” says Gelber.

The approach, called STEPP (Subpopulation Treatment Effect Pattern Plot) is an alternative to the widely-used technique of subgroup analysis, which evaluates drug responses for specific subsets of pati study group. Such analysis might suggest, for example, that subgroups of patients whose tumors expressed particular levels of a protein “biomarker” in their blood or urine benefited more than other patients.

However, the range of biomarker values in each of the subsets is arbitrary. For instance, all patients with biomarker levels from 1 to 10 might be in one subgroup; 11 to 20 in another, 21 to 30 in another, and so on. A STEPP analysis, by contrast, has the power to examine the pattern of responses to treatment in terms of a continually varying factor, such as increasing amounts of the biomarker, rather than merely in arbitrarily defined “buckets.”

 

Gelber says the STEPP method is more informative and yields better predictions of response based on biomarkers. “You’re less likely to be misled by wrongly interpreting the outcome data,” he says.

Gelber described a new study in which STEPP analysis proved valuable in determining which postmenopausal women with early breast cancer would benefit the most from tamoxifen, an aromatase inhibitor (letrozole, or brand name Femara), or a combination of the two given in different sequences. The STEPP was applied to the previously reported findings of a large international clinical study involving more than 8,000 patients.

The main finding of that study was that women who took letrozole did better than those who took only tamoxifen. The STEPP analysis allowed the researchers to weigh the benefits and side-effect profiles of the different treatment options as a function of the patients’ initial risk of cancer recurrence.

“This STEPP analysis clarified which patients might benefit the most from letrozole, and which obtain sufficient benefit from tamoxifen,” Gelber says.

Turning Point 2011 Dana-Farber Cancer Institute

This entry was posted in Breast Cancer, cancer research, Dana-Farber, support for cancer and tagged , , , , , .