ALTTO demonstrated that adding lapatinib (Tykerb/Tyverb) to standard trastuzumab (Herceptin) therapy after surgery does not improve survival in women with HER2-positive early breast cancer.
The results were presented at the annual meeting of the American Society of Clinical Oncology (ASCO) and, as reported by Medscape Medical News, caught many in the world of breast cancer research by surprise.
"This is a serious disappointment, not just for investigators, but for the entire field," announced study discussant George Sledge, MD, from the Stanford University School of Medicine in California, at the meeting.
The results have "tremendous" and "profound" implications for future breast cancer drug research, noted ASCO president Clifford Hudis, MD, who is chief of the breast cancer medicine service at the Memorial Sloan Kettering Cancer Center in New York City.
But as the dust continues to swirl, there remains a deafening silence from the US Food and Drug Administration (FDA) — which could upend research on the basis of the ALTTO results.
It was expected that the combination of lapatinib and trastuzumab would improve survival because the neoadjuvant NeoALTTO trial hinted at this. However, that trial simply showed that the combination could shrink tumors before surgery.
Many researchers assumed that this shrinkage, known as a pathologic complete response (pCR), was an early signal that the therapy would prolong lives. They expected that ALTTO would confirm the NeoALTTO signal, and that the combination would improve survival.
The failure of ALTTO to prove this point has cast doubt on an exciting frontier in cancer drug development.
The FDA was so convinced of the prognostic value of pCR that it issued a draft guidance in 2012, outlining how drugs that are shown to induce pCR can get accelerated approval for use in the clinic.
Drugs currently move at a snail's pace through the approval pipeline, so accelerated approval is an exciting possibility. In general, a drug must be tested for many years before it is considered to improve survival, but with accelerated approval, several months of study can show that a drug can induce pCR. In theory, accelerated approval can shorten costly trials, save millions of dollars, and take years off the wait for a new drug.
"The FDA and many others have argued that the improvement in pCR is a surrogate — a leading indicator — for improvement in survival," Dr. Hudis told Medscape Medical News. "If a regimen leads to higher pCR rates, then it should, when applied in the adjuvant setting, lead to improved survival. If this is true as a principle, then we can shave years and millions off our adjuvant therapy trial designs and investment."
The FDA draft guidance on accelerated approval was the short-cut that got pertuzumab (Perjeta, Genentech), another breast cancer drug, to the market in 2012. That approval was granted on the basis of pCR results from the NeoSPHERE (Lancet Oncol. 2012;13:25-32) and TRYPHAENA (Ann Oncol. 2013;24:2278-2284) trials. Whether those pCR results translate into improved survival for pertuzumab will not be known until results from the ongoing APHINITY trial are released in a few years.
But in the wake of the NeoALTTO and ALTTO trials, the APHINITY results seem less clear, and the ability of pCR to predict survival has become a debate.
Since the death knell for pCR was sounded at the ASCO meeting, some dissenting voices have emerged.
"pCR is absolutely a good predictor of outcome," said Laura Esserman, MD, a breast cancer surgeon and director of the Carol Franc Buck Breast Care Center at the University of California, San Francisco. "Actions speak louder than words, and there's no withdrawal of the FDA's draft guidance," she pointed out in an interview.
An FDA spokesperson told Medscape Medical News that, although it cannot provide a timeframe, it is working to finalize the guidance and cannot comment further on the subject.
As coprincipal investigator of the I-SPY 2 trial, Dr. Esserman has many reasons to defend pCR.
The goal of I-SPY is accelerated drug development, and pCR is its primary end point. In fact, in 2012, Dr. Esserman and her I-SPY colleagues reported results showing that pCR is predictive of survival (J Clin Oncol. 2012;30:3242-3249).
But Kathy Miller, MD, associate professor of medical oncology at the Indiana University School of Medicine in Indianapolis, who helms the Medscape Miller on Oncology blog, is not convinced by the I-SPY results. She said that although pCR can sometimes predict survival, it does not always.
"pCR has always been a very imperfect surrogate," she told Medscape Medical News. "It is not a good surrogate for survival at all in patients with estrogen-receptor-positive disease, and it is a better though still very much imperfect surrogate for those with triple-negative disease or HER2-positive/estrogen-receptor-negative disease." The I-SPY results were driven largely by triple-negative and HER2-positive patients, she explained.
Because of this, pCR is an unreliable end point to drive drug approval, said Dr. Miller. "It may get some drugs that are ultimately very beneficial to our patients sooner; if it does, that is a very good thing. But if we're going to take that path, we need to do so with open eyes and recognize that what we've seen in ALTTO, a fully powered adjuvant trial that didn't confirm the benefit we thought we were going to see on the basis of pCR, will almost certainly be replicated. We all need to acknowledge that."
But the I-SPY investigators take issue with the interpretation that ALTTO failed to confirm the NeoALTTO pCR findings. Within days of the ALTTO news breaking at ASCO, Dr. Esserman and her coprincipal investigator, Don Berry, PhD, from the University of Texas M.D. Anderson Cancer Center in Houston, issued a statement voicing their disagreement with the conclusion that pCR is not useful in predicting survival.
Their arguments are many and varied, and perhaps contradictory. Statistically, both trials are "amazingly consistent," said Dr. Berry, who is a statistician. "The arguments coming out of ASCO about the relevance of neoadjuvant end points are based on ALTTO not reproducing NeoALTTO. These arguments are based on a misinterpretation of statistical significance or lack thereof. Namely, NeoALTTO was 'statistically positive' for pCR, while ALTTO was 'statistically negative' for event-free survival. Statistical significance does not imply truth, and lack of statistical significance does not imply falsity," he explained.
Adding to the statistical consistency argument is the debate about inconsistencies in the 2 trial populations, said Dr. Esserman. This makes it difficult to extrapolate the findings of one trial to the other, she noted. In addition, there are many possible explanations for the unexpected ALTTO results, such as differences in estrogen-receptor status, node status, and duration and sequence of therapy.
"I don't think the ALTTO story should be the one to kill pCR in terms of accelerated development," she said. "This makes me feel even more strongly that this is an important model to test. I honestly cannot understand why we would abandon this process. Why would we want to go back to a method that guarantees we have a 15-year turnaround for drug approval? There is no reason for us not to continue down this path. My personal belief is that we should be driving to compress the timeline, that we owe it to our patients and to science, that there is absolutely no downside to moving forward, and that there is no indication at this time, certainly from this trial, that we should abandon everything."
"I'm not surprised that the I-SPY group is trying to ignore — or excuse — the ALTTO results and the potential implications it brings for relying on pCR as a surrogate," countered Dr. Miller.
But the dream of accelerated approval, and all its potential, might be taking on a dangerous life of its own, she suggested.
We don't know yet "whether the potential benefit outweighs the potential downside," she explained. "But I do think we're only now beginning to recognize and discuss the potential downside. Unfortunately, these are discussions we should have had before. It seems likely to me that if the timing had been different and the FDA guidance had already been in effect, NeoALTTO would have led to accelerated approval of lapatinib, and then the ALTTO results would have come out and the approval would have to be removed. I imagine you'd get some folks from the FDA who would argue that shows the process works. But in the meantime, think about how many patients would have been exposed to an agent that has substantial costs, substantial toxicity, and, at least according to the best standard we have, does not improve survival," Dr. Miller said.
One thing both Dr. Miller and Dr. Esserman agree on is that the ALTTO fallout is accumulating.
"I do think the ALTTO results created general confusion," said Dr. Esserman.
"I'm not sure there is a majority opinion, and I don't think there was a majority at ASCO either," said Dr. Miller.
But Dr. Esserman is confident that the future is bright for I-SPY and it's pursuit of the pCR end point.
"There's not going to be a big surprise in the final FDA guidance," she predicted. "I-SPY investigators are still totally on board, the FDA is totally on board. I've been in touch with them, and there isn't anything here that I believe has changed their stance. This path forward will dramatically change the time it takes to get access to drugs for our patients. We absolutely have to run this experiment. We cannot abandon it."
référence Landray:2014hh inexistante Pubmed