According to the National Breast Cancer Foundation, one in eight women will be diagnosed with breast cancer in her lifetime, making it the most commonly diagnosed cancer in women worldwide. Medical research continues to advance our knowledge and improve treatment of the disease. More importantly, it has extended survival rates through all subsets and stages.
One such advancement in breast cancer treatment is molecular testing. Of all the available breast cancer molecular tests, the Oncotype DX test has the ability to predict the benefit of chemotherapy in a select group of breast cancer patients.
The TAILORx study was a landmark study that proved the value of the Oncotype DX 21-gene expression assay in determining which patients with early-stage, hormone-receptor-positive breast cancer benefitted from adding chemotherapy to endocrine therapy and which patients did not. Opened in 2006, the results of TAILORx were first published in the New England Journal of Medicine in November 2015, with follow-up results published in June 2018.
Prior to TAILORx, many women with early stage estrogen-receptor positive, human epidermal growth factor receptor type 2 (HER2)-negative breast cancer were frequently treated with both post-operative chemotherapy and endocrine therapy. However, approximately 85% of these women were suspected to be over-treated as they would have remained recurrence-free at 10 years with post-operative endocrine therapy alone.
The study enrolled 10,253 women with hormone-receptor-positive, HER2-negative, axillary node-negative breast cancer at 1,182 sites in the United States, Canada, Australia, Ireland, New Zealand and Peru. These womens’ cancers were analyzed utilizing a 21-gene molecular test (the Oncotype Dx Breast Recurrence Score) and were assigned a recurrence score (0-100). Women with a low recurrence score (0-10) received hormone therapy only and women with a high recurrence score (26 and above) received both chemotherapy and hormone therapy. Women with an intermediate score (11-25) were randomly assigned to receive either hormone therapy alone or hormone therapy with chemotherapy.
The TAILORx study proved the clinical usefulness of the Oncotype DX assay and showed no benefit from chemotherapy in 70% of women enrolled in the study. Removing this unnecessary care not only saves hundreds of millions of dollars across the U.S. healthcare system, but also avoids unnecessarily exposing patients to potentially harmful drugs.
The TAILORx study was a major landmark in setting a new standard of care for breast cancer patients. But how well does real world treatment and outcome mirror that of the clinical trial?
Using COTA’s real world observational database from electronic health records (EHRs), researchers sought to confirm the results of the TAILORx trial. These findings were presented at the 2018 San Antonio Breast Cancer Symposium. The study used eligibility criteria similar to TAILORx, and identified 1,009 patients with a RS 11-25 from 23 cancer centers who had received adjuvant endocrine therapy alone (E) or adjuvant chemoendocrine therapy (CE) as part of standard care. Estimated 5 year overall survival rates were 98.9% with E and 97.8% with CE – consistent with the findings from TAILORx.
Researchers found that using a real-world data source, endocrine therapy alone appears to yield excellent five-year survival rates among patients with a similar cancer profile as those enrolled in the TAILORx study. Unfortunately, they also illuminated a treatment bias as oncologists were more likely to utilize CE in younger patients, large tumors and higher recurrence scores, thus limiting full confirmation of TAILORx in this initial, exploratory analysis.
While the COTA results were consistent with the findings from TAILORx, the analysis also shed light on a potential drawback associated with real-world data – the presence of treatment selection bias that is eliminated in a randomized study design. Even taking this into account, there is significant value in using RWD as confirmatory evidence to complement randomized clinical trials. In addition, a number of established and emerging statistical methods can be applied to high-quality datasets with the intent of reducing bias associated with known prognostic factors.
Thanks to a growing understanding and acceptance of RWD by regulatory bodies, changes like this are being made in clinical research today, helping to expedite studies compared with traditional clinical trial methods. Rather than replacing clinical trials, these two models can work together both in clinical trial designs and postmarket research, allowing for a more complete picture of the patient group being studied.