The practice-changing TAILORx trial was brought to you by publicly funded cancer clinical trials system, pointing to its continuing relevance, said Jeff Abrams, NCI acting director for clinical research and associate director of the Cancer Therapy Evaluation Program.
Genomic Health, the company that developed the test, used specimens from NSABP to test the Oncotype DX recurrence scores, “These samples from the NSABP trials were very important in the development of Oncotype Dx, although Genomic Health also used samples from other studies to help confirm the results they achieved with the NSABP samples,” Abrams said.
“It is important to cite the people who made this study possible. First, and foremost, the over 10,000 volunteers who participated in this study. We owe them immense gratitude,” Abrams said. “ECOG- ACRIN ably led this study on behalf of all the adult Group including the Alliance, NRG, and SWOG, plus the Canadian Clinical Trials Group. Without participation by all, we would not have been able to complete such a large study.
“It speaks to the fact that we really do have a national network that can conduct studies like this.”
Abrams spoke with Paul Goldberg, editor and publisher of The Cancer Letter.
Paul Goldberg: Can we do a little history? It all begins with the 2000 NIH consensus conference on adjuvant breast cancer; right?
Jeff Abrams: Well, the NIH Consensus Conference took place in 2000, and it was based on some NSABP trials and other trials that showed a small survival benefit for women who had node-negative, hormone-receptor positive breast cancer with no underarm lymph nodes who got hormonal therapy plus chemotherapy compared to hormone therapy alone. And, based on these findings, the recommendation was that chemo should be considered in all these women, even though the benefit from chemo was small. We knew we were probably overtreating many of the women to help a few, but we were not able to select who would benefit.
I guess, there was no way to stratify risk; right?
JA: There were some ways to select the patients based on tumor size. If the tumor was very small, then the patients didn’t get chemotherapy. Basically, if the tumor was a centimeter or more, and they had positive receptors, then chemo was indicated, even though the survival benefit was small overall.
The question became: can we figure out who are the patients who really need the chemo based upon a high risk of recurrence? And who are the patients who are getting only the side effects of chemo, as their risk is low and chemo isn’t likely to help?
What about the technology? That wasn’t available until much later.
JA: It’s an interesting story. Steve Shak had worked for Genentech, where he had helped in the development of Herceptin back in the 1990’s, and that’s when I first met him.
He had moved to a new company that was just starting up, Genomic Health, and because we had known each other, he came to visit me at NCI, and asked if I knew of any possible patient samples which could help them confirm the genetic test they were developing. Their 21-gene recurrence test, called Oncotype DX, was being developed to determine the risk of recurrence in patients being considered for adjuvant chemotherapy.
NCI had supported two large NSABP studies, the B14 study with tamoxifen, and the B20 study which used tamoxifen and chemotherapy. They were the perfect studies to test their Oncotype DX test on.
Steve Shak met up with Soon Paik, who was the pathologist at NSABP, who had access to all these samples, and Steve worked very closely with Soon Paik to test this recurrence score, doing their particular genetic test, on all these samples from the two studies.
These samples from the NSABP trials were very important in the development of Oncotype Dx, although Genomic Health also used samples from other studies to help confirm the results they achieved with the
NSABP samples.
It goes back to the question of how useful are cooperative groups. Coke versus Pepsi, all this criticism. We wouldn’t be here if not for those annotated samples.
JA: Exactly. And, it was a great thing that the Groups had those samples stored, on a high percentage of the patients, which made this retrospective analysis more believable.
And then, you go onto a prospective study. But, before we get to that, as I look at the results, I’m seeing the Genomic Health test, now, being hard-wired into the system. Can any other test be used?
JA: There are other genetic tests on the market. This study began in 2006. It accrued patients from 2006 to 2010. And then we had to wait for the outcome results, to follow all these patients for another eight years before we had sufficient follow-up to be certain that we could omit chemotherapy in some patients. And, it is now very reassuring to the have nine-year follow-up.
In this particular disease, if we only had five-year follow up, I think a lot of people would be saying, “Oh, you have to wait longer to see if the chemotherapy didn’t have an effect over more time.” But now that we have nine-year results, I think we feel very secure in the recommendation to omit chemo.
But, the Genomic Health test is really the only one to use?
JA: There are other tests on the market that can select patients at risk of recurrence.
Right, that’s why I was asking.
JA: MammaPrint, from Agendia, was tested in a major international study called MINDACT. The study was conducted by BIG, the Breast International Group, primarily outside the U.S.. And that study actually reported out, two years ago at ASCO. The way that test works is it puts patients into two baskets: low-risk, and high-risk. So, it doesn’t have this recurrence score idea, but rather puts patients into the two categories.
And, then, there’s yet another test called Prosigna, by NanoString. And it also tries to predict risk. So all these tests try to predict recurrence in the same group of node negative, hormone receptor positive patients taking hormonal therapy.
What about the cost of this trial? How much did it cost?
JA: NCI doesn’t fund its trials on a trial-by-trial basis. We give grants to our trial Groups, and they do a number of trials with the funds, but a rough calculation is possible. We estimate the trial cost about $35 to $40 million over the years.
That money was to reimburse the sites around the country for their work. There were a lot of sites who participated in this trial. Then, the funds also go to pay the operations and statistical costs of ECOG-ACRIN, and then there were costs for the Oncotype Dx test. We were fortunate at NCI to receive Breast Cancer Stamp Act funds to help pay for the test, as well as Genomic Health donated some in-kind cost reductions for the tests.
How much did the Breast Cancer Stamp provide?
JA: About $5 million.
So, that’s another five, so could be around $45-ish million at the most.
JA: Yes, that is a reasonable estimate.
Can a trial like this be done today? Because you’re talking about randomizing over 6,000 women.
JA: I think it could be, if we had an important question to ask. However, we’ve gotten much better at picking out molecular subsets. Due to the advances in diagnostics, most of our trials nowadays are much smaller, focusing on the subset we think will benefit most from the treatment. So, we aren’t doing nearly as many very large trials as we used to.
So, you don’t think it would be necessary? You might be able to zero in on a specific set of patients.
JA: I think there may still be some large trials. For instance, we’re doing a trial called BWEL in breast cancer, which is still pretty large. Not 6,000, but several thousand patients, and that’s a trial looking at exercise and weight loss in women who’ve had breast cancer to see how these factors can change outcomes. So, when you don’t have something that’s targeting a specific molecular effect, you have to do a larger trial, but for our treatment trials, where we’re using drugs, we try in most cases to limit the trial to include only those patients likely to benefit. Such trials are usually under a thousand, even under 500, patients nowadays.
What are the questions that are left unanswered right now about this trial?
JA: If your score is above 25, I think the study provides evidence that one should discuss chemotherapy with these women. The study showed that women who received hormones plus chemotherapy in this group had a higher recurrence rate than the women with scores under 25. Based on the earlier studies that indicate a benefit for adding chemotherapy to hormones, the data from this study suggest that any woman with a recurrence score above 25 should at least consider chemo with their oncologist.
The area of uncertainty in this trial is for women 50 and younger who have a recurrence score between 16 and 25. The data indicate that for women with a recurrence score of 11 to 15, chemotherapy isn’t going to improve their outcome, so no need to get it.
But, from 16 to 25, the data show some benefit for chemotherapy in women under 50 years old. From 16-20, the benefit is still small but becomes more obvious from 21-25.
So, I think RS 16 to 20, it’s worth a discussion, and each patient and their doctor will have to decide; from 21 to 25 in the younger women, the benefit appears more clear and these women should be considered to be more similar to other women with scores above 25.
Now, there is one other very interesting point in these younger women, and that may be that what chemotherapy actually does in these younger women: it may be that chemotherapy primarily works to suppress their ovarian function, put them into menopause. Thus, it may not be the typical cytotoxic effect of chemotherapy on the tumor that is occurring but rather the chemotherapy is suppressing estrogen production by the ovaries. We’ve learned from other studies that ovarian suppression combined with hormonal therapy is necessary to get the best effect in younger women.
So, what remains unclear from this study is: do they need chemotherapy or could they do just as well with other non-chemo drugs that suppress ovarian function?
Are you planning to address this question?
JA: Well, we do know from an international study that our Groups helped to conduct called SOF that, in younger women, if you suppress the ovaries with a drug, plus give an aromatase inhibitor to those women, their outcome is better than women who just receive hormonal therapy with tamoxifen.
So, we know that, in younger women before menopause, ovarian suppression is critical. What remains uncertain is: is chemotherapy doing anything beyond just ovarian suppression? How to answer that question would require further study.
What about the improvement in hormonal treatments and the chemotherapies used over this time, or what’s available now versus what was available when the trial was started? Is that a factor?
JA: It is a factor, although it’s interesting. Back in the 1990s, a lot of women got an anthracycline-containing regimen which has cardiac toxicity concerns long-term. But in this study already, the field had begun to change, and the majority of women in this study did not get an anthracycline, so that’s consistent with modern chemotherapy.
It was mostly a taxane and cyclophosphamide that was used in this study. As far as the hormonal therapy, most of the women in this study got aromatase inhibitors, not tamoxifen, except if you were under 50, where aromatase inhibitors don’t work so you have to get tamoxifen.
So, I think even though it started a long time ago, the therapy used in this study is consistent with the type of treatments women are getting now.
If you were to look back and think of the overtreatment size—this is 20/20 hindsight of course—but how would you estimate that?
JA: With 260,000 women a year being diagnosed with invasive breast cancer, about half are in this node-negative, hormone-receptive-positive group. So, that’s 130,000 women. Back in 2000 at that consensus conference, we were suggesting that chemotherapy be considered in most of them.
Now, it’s sort of totally reversed.
Using Oncotype or similar tests, chemotherapy is now considered for only 20-30 percent of patients.
This test and others have really turned our recommendation around and chemotherapy is really reserved for those with the high recurrence scores at any age or for those women that I just talked about who are under 50 and have the score in the 16 to 25 range.
We estimate that’s only about 20-30% of the group that might benefit from chemotherapy so 70 percent or more can be spared chemo.
What’s the take-home from all of this? Is there anything we have learned that surprises you here?
JA: As we learned over many years. cancer is, at base, a genetic disease, and if you can understand the genes, it really goes a long way to helping you decide what’s the appropriate treatment. With all the targeted therapies today, oncology has really changed; we are now targeting the therapy to the genetic changes.
That was the theme behind the MATCH study that NCI and ECOG-ACRIN are conducting. This approach has helped women with breast cancer too, dating back to the discovery of of Herceptin for the HER2+ breast tumors, and more recently the advent of the CDK4/6 inhibitors that improve outcomes for women with hormone sensitive metastatic disease.
It’s all based on truly understanding the underlying genetics of the tumor.
It’s not where we all were in 2000, which isn’t a long time ago, really.
JA: Now we have next generation DNA sequencing technology that can be applied broadly which has resulted in a major change in how we approach patients with cancer. Fortunately, recent approval of this testing for advanced cancer by Medicare should help make this technology more affordable for all patients with cancer.
Is there anything we’ve missed?
JA: It is important to cite the people who made this study possible. First and foremost, the over 10,000 volunteers who participated in this study. We owe them immense gratitude. ECOG-ACRIN ably led this study on behalf of all the adult Group including the Alliance, NRG, and SWOG plus the Canadian Clinical Trials Group. Without participation by all, we would not have been able to complete such a large study.
It speaks to the fact that we really do have a national network that can conduct studies like this.
This is the type of study that a pharma company would not do because it’s not intended to bring a new treatment into medicine; it’s really intended to decide how best to treat patients, and actually ended up omitting chemotherapy for many women.
It shows that the government still has a role to play for this kind of important trial.
Actually, to go down that path, if you were to take the number of women who should not be treated, which is roughly around 100,000 a year, right, and multiply out the cost of chemotherapy and not even looking… think of how many billions of dollars could be saved.
JA: I think there are going to be health economists who will do these calculations more accurately, but the one thing that is obvious is the test itself, and the other tests that I mentioned to you, all run around $4,000, whereas a course of chemotherapy can vary from as low as $25,000 if there’s no complications, and up to $100,000 or more if you have to use a lot of supportive care or hospitalization. So, when you look at the $4,000 versus $25,000 or more, there’s a major difference there.
Right, and when you multiply it all out by the number of women who are not going to get that $25,000 to $100,000 expense, in addition to morbidity and side effects from chemotherapy, you’re talking, really, many, many, many, many, many billions of dollars saved every year for the system.
JA: Firstly, it’s a great outcome for women who don’t have to undergo the toxicity. But second, it’s a good outcome for the healthcare system.
Part of the savings that we just discussed in terms of the difference between the cost of chemo and the cost of the test have already been realized. Over the the intervening years while we were waiting for the results from TAILORx, the Oncotype Dx was on the market and was already being used by physicians to help them decide whether or not to recommend chemo.
A large part of the savings from a recurrence score of zero to 18 has already been realized because many oncologists and patients were already acting on the information. What is likely is that we will see further savings, because there was a zone of uncertainty between 18 and 31, where people were still debating, based on the older, retrospective results, whether they should or shouldn’t give chemotherapy.
And, with the new results from TAILORx, there’ll be more certainty that from 18 to 25 women can be spared, and only above 25 should chemotherapy be considered.
Thank you so much.
