Comparison of Prognostic Tests in Breast Cancer
Transcript:Adam M. Brufsky, MD, PhD: Mark, you mentioned that comparison. I call it the “bake-off” that occurred at San Antonio this year. Do you want to tell the audience what that was about?
Mark E. Robson, MD: So, they took the TransATAC trial and they looked at a variety of different scores including, was that your IHC4 score?
Adam M. Brufsky, MD, PhD: No, it wasn’t our IHC4. It was the one that Mitch Dowsett did.
Mark E. Robson, MD: Right. And then—I can’t remember all the different ones, but things like EndoPredict, the 21-gene recurrence score—they essentially saw how much incremental discriminatory capability you would get from tacking scores together. And you have the details, the facts, and figures much more in your mind because I’ve seen you present this. But the bottom line is, everything added a little bit to everything else. The question was: which one added more in what circumstance, especially for the later recurrences that were obviously a big issue for us because of the extended adjuvant question?
Adam M. Brufsky, MD, PhD: So, the most fascinating part, and I want to hear you guys’ opinions about it, was when you looked at the data, they had the IHC4 and what was basically added to the IHC4. Then they had the clinical parameters for tumor grading, nodes, etc, and how much it added to it. It was interesting. The tests were 21-gene, PAM50, Breast Cancer Index (BCI)—which we’re going to talk about as we talk about extended adjuvant therapy in a minute—and EPclin, which is EndoPredict. I think there was 4.
Mark E. Robson, MD: That was it.
Adam M. Brufsky, MD, PhD: That was it. And so what they found for node-negative breast cancer, all of them seemed to do really well looking at 0 to 10 years of follow-up. So, in the 0 to 10 years of follow-up, all these tests did really well. For node-negative from 5 to 10 years, all of them also seemed to do pretty well, maybe with the exception of the 21-gene assay, which did not do as well. And where the tests seemed to be starting to fall apart a little bit, BCI did well, too, actually. It was in the node-positive. So, in the node-positive, really the ones that seemed to do best were the ROR (risk of recurrence) scores PAM50 and EndoPredict. They seemed to do the best. And, in particular, they did best in that 5-to-10 year range, where BCI and the rest of them did very well. But the only ones that seemed to really be discriminatory and find a group of patients that truly were not going to relapse for the node-positives were the PAM50 and the EndoPredict, or the EPclin. And so, it’s really interesting because you follow that up with another piece of data, where they did an analysis—they presented this at San Antonio—where they analyzed a breast cancer database with 75,000 women in it. And of those 75,000 women with node-positive, ER-positive breast cancer, 15,000 got the 21-gene assay for that. So, it’s being used a lot. That’s 15% or 20% of the population. It’s being used a lot, and we’re not quite sure if it really has discriminatory power in the node-positives. So, again, I’m really curious. This is where we are in this world. Do you have any final thoughts before we move on to the next topic, guys?
Kimberly L. Blackwell, MD: Just to follow up on that node-positive point: I think MINDACT and the TransATAC trials, because they included node-positive patients, all of these actually make me feel a little bit better about sending the 21-gene assay in node-positive. So, the reality is, we had a limited data set of, I think, 700 patients from Kathy Albain’s study in which the 21-gene assay was validated in node-positive. But that was all the data we had until MINDACT and the TransATAC data. I guess I’m using it pretty widely in my node-positive patients, 1 to 3 positive lymph nodes. And for me, the lack of precision in the extended setting for node-positive was not all that concerning. Because when I have node-positive, ER-positive patients, I am recommending extended adjuvant therapy if they’re tolerating it. That’s not as big of an issue as the node-negative, where we saw the precision for extended adjuvant. So, 2 points: I think the extended adjuvant data, node-positive, wasn’t so worrisome to me from TransATAC and across the studies. I think we’re seeing a consistent signal that these genomic predictors actually do work in node-positive. They, at least, are prognosticating in node-positive. Whether or not they predict chemotherapy remains to be seen.
Transcript Edited for Clarity
