In his article, Dr. Seidman gives a succinct and comprehensive review of the landmark events in the systemic therapy of breast cancer over the past 2 decades. The topics covered are germane to clinical practice. The next 20 years is presented with optimism, and there is reason for this. Systemic therapy, including targeted/biologic therapies, is exploding. The "beginning of the end of chemotherapy" is an era eagerly awaited by many patients hoping it will occur soon enough to impact them. Optimism, however, must be tempered. There are reasons to be cautionary, and herein the review can be improved. Dr. Seidman raised several important issues that need to be addressed as we move ahead. What is missing, in my opinion, are his insights on what we need to be vigilant about to realize the promise of the next 2 decades. What are the questions we must critically answer? What pitfalls must we avoid? What has the past taught us about drug development? How do we rationally develop targeted agents? Although these themes are difficult to craft and can be colored by one's perspective, his candid opinion(s) would have given the article a quality beyond the eloquent review of significant trials in breast cancer. Thus, I will focus more on these and less on trial results already substantively discussed by Dr. Seidman.

Targeted Therapy: Lessons Learned

From promising preclinical observations to pivotal trials in metastatic disease and most recently in adjuvant therapy, the development of trastuzumab(Drug information on trastuzumab) (Herceptin) is a spectacular story. Development was rational, rapid, and practice-changing. What can we learn from this process? Some things were done well: (1) A target was identified, (2) the best way to assay that target was defined, (3) appropriate patients were selected based on presence of the target, and (4) pharmacokinetics was incorporated to make scheduling more convenient.

We could improve on/learn from some aspects: (1) Is continuing trastuzumab after disease progression valuable? (2) Did we need that many trials to see the benefit of trastuzumab in early-stage disease? Could we have come to a conclusion with fewer? (3) By allowing only patients 6 months from their adjuvant therapy to get trastuzumab, we may have lost an opportunity to determine the benefit of delayed initiation of adjuvant trastuzumab. In a different population, MA.17, by allowing crossover to letrozole(Drug information on letrozole) (Femara) regardless of when tamoxifen was completed, is providing evidence for benefit of delayed initiation of aromatase inhibitor therapy.[1]

In contrast to trastuzumab, there is less enthusiasm for bevacizumab(Drug information on bevacizumab) (Avastin) and uptake has not been as rapid. E2100 is no less spectacular than the pivotal trastuzumab trial[2] and toxicity is manageable, so why the lack of enthusiasm? The initial phase III trial of bevacizumab[3] was perceived as negative despite the doubling of response rate. Negative trials create less excitement, but this is not the sole reason for the slower uptake. E2100 accrued well, so there was enough faith in bevacizumab to put patients on the trial.

There are some legitimate reasons behind the slower uptake: (1) Lack of a marker to select the most appropriate patients for this therapy; (2) absence of a reliable test to determine if angiogenesis is being modulated by the therapy; and (3) cost. The cost issue is more profound when taken globally. There are patients world-wide that can benefit from this treatment but the cost is prohibitive. It is disturbing to envision a dichotomy where patients have worse outcomes simply because of economics. Breast cancer patients worldwide share the same aspirations regardless of their ability to pay—all hope for a cure, for fewer recurrences, and for less toxicity. The translational model must flow not only from bench to bedside but also to the global breast cancer community.

Metastatic disease is lethal; we continue to develop better therapies in the adjuvant setting where we have a better chance for cure. As we focus on improving cure, it is easy to forget a point of diminishing returns. Some patients' benefit from a given therapy is marginal at best. But can we afford to compromise the cure of a few to spare many from the toxicities of therapy? Thankfully, models such as Adjuvant! Online and molecular assays like the Oncotype DX have been developed to better predict benefit from a given adjuvant therapy. The Breast Intergroup's collaborative effort in the TAILORx study is laudable. Foretelling that it will validate the Oncotype DX assay, my enthusiasm is tempered by the prospect of a "poor man's recurrence score assay" and whether the Oncotype DX assay will perpetually be out of reach for patients in less affluent countries.

Clinical Trials: Can We Do Better?

Dr. Seidman's analogy of a "book of many chapters" for metastatic breast cancer is quite fitting. Although we would like to ultimately impact survival, it may not be the best endpoint to look at in the in the early chapters of the book. Pinning down the most suitable endpoint is complicated by the scenario of dissimilar endpoints for different therapies. For instance, response rate may be a reasonable early endpoint for chemotherapy but not for biologic therapy. It can get even more complicated when combining both therapies. A key attraction of TCH (docetaxel/cisplatin/trastuzumab) was that it was predicted to have the best synergy based on a preclinical model.[4] So the results of BCIRG 006 and BCIRG 007 were, to some degree, a letdown. More importantly, there was a serendipitous victim of these results—the hope of preclinical model(s) helping with rational design of therapies to accelerate novel drug development.