Neoadjuvant therapy in breast cancer is a hot topic these days. Giving drugs in the neoadjuvant setting has several advantages:
An early surrogate marker that is being increasingly used in breast cancer is pathologic complete response (pCR). A surrogate for a true endpoint is an endpoint that can be used in lieu of the true endpoint to assess treatment benefits.
So, what makes a good surrogate marker?
The measurement must be robust, standardized, and reproducible. The relationship with the gold standard endpoint must be better than just simple correlation--it must be valid at both the patient and trial level, and requires continued follow-up to establish the relationship to the gold standard of event-free survival (EFS) or overall survival (OS).
In breast cancer, it has been hypothesized that an increase in pCR with the addition of a new agent to standard therapy in the neoadjuvant setting will predict a survival benefit in the early breast cancer setting. As such, the FDA developed a plan for "neoadjuvant drug approval" with the goal of establishing a pathway for accelerated drug approval in breast cancer, using the neoadjuvant setting while establishing appropriate endpoints and determining when to use pCR as a surrogate endpoint.1
Does this work? Is pCR truly a surrogate endpoint for survival?
Cortazar and colleagues performed a meta-analysis of almost 12,000 patients receiving neoadjuvant chemotherapy for breast cancer. On a patient level, individual patients who had a pCR did have superior outcomes, and patients with HER2-positive and triple-negative breast cancer (TNBC) had the highest pCR rates. However, at the trial level, there was no relationship between the magnitude of improvement in the pCR rate and EFS/OS. Potential explanations for this finding are that the included trials had low pCR rates and only one small trial included targeted therapy. The optimal trial design for neoadjuvant drug development must include adequately powered randomized controlled trials, which are designed to test for superiority, isolate the drug effect, and standardize pathologic assessment.2
Other factors that may affect the prognostic value of pCR on EFS/OS are the administration of postoperative therapy, toxicities limiting delivery of the novel agent or regimen, antagonism between standard and novel therapy if added to standard therapy, the inclusion of good prognosis patients, and differences between the tumor in the breast and micrometastatic disease (i.e., cancer cells and cancer-stromal interactions).3
The two studies frequently discussed in this context are, NeoALTTO and ALTTO , which evaluated dual anti-HER2 therapy with trastuzumab and lapatinib in patients with HER2-positive breast cancer in the neoadjuvant and adjuvant settings, respectively. These trials evaluated whether the addition of lapatinib to standard chemotherapy and trastuzumab in patients with HER2-positive breast cancer improved pCR rates in the neoadjuvant setting and disease-free survival (DFS), or OS in the adjuvant setting. More importantly, the trials raised the question of whether an increase in pCR in the neoadjuvant setting predicted DFS or OS in the adjuvant setting. The results: NeoALTTO showed an increase in pCR, whereas ALTTO did not show a survival benefit.4, 5
Does this mean that pCR is not a good surrogate endpoint? It may not be that simple. Some key points to keep in mind: unlike ALTTO, NeoALTTO was not powered for survival endpoints. NeoALTTO included patients with more advanced stage disease who only received taxane-based and HER2 therapy neoadjuvantly, and anthracycline-based therapy adjuvantly. Furthermore, these trials included predated standards for pathologic assessment and included absence of disease in the breast as a pCR definition as opposed to the more current definition of pCR, which includes absence of disease in the breast and axilla. ALTTO enrolled a very “good risk” population which is perhaps why this study showed a lower than expected event rate.1, 4, 5
It is clear that the validation of pCR as a predictive endpoint is still needed, and the “neoadjuvant hypothesis” does hold potential for faster and more focused development of new agents for breast cancer. It is important to note that there are other benefits of more effective neoadjuvant therapy in addition to predicting EFS and OS:
We need innovative trial approaches to move us forward which are where the I-SPY2 and I-SPY3 trials come in. I-SPY2 is a randomized phase II screening trial that aims to identify new agents and/or biomarker combinations that will improve pCR. If agents or combinations have an 85% probability of success in a phase 3 trial, they will “graduate” from I-SPY2 and be tested in I-SPY3, a randomized phase III confirmatory trial. So far, the combination of veliparib (a PARP inhibitor), and carboplatin has graduated I-SPY2 in the TNBC signature and there is a confirmatory phase III trial in progress in the neoadjuvant setting. Neratinib graduated from I-SPY2 in the HER2-positive, hormone receptor-negative signature, and there is a confirmatory phase III trial in development.1, 6, 7, 8
Despite all that we know, a lot of questions remain. Might treatment convert a "good responder" patient to a pCR, but have little impact on a "poor responder" patient? If so, can we do better at identifying patients who are unlikely to respond to a standard neoadjuvant therapy and thus, are more likely to respond to a novel agent or regimen? We need to better understand which patients could benefit most from more effective neoadjuvant therapy and thus, better understand both patient and tumor biology.
However, at this point, it looks like the verdict--and the jury--is still out.