Over a 30-year period in the 20th century, human flight evolved from the propeller to the jet engine and then managed to send us to the moon and back. The changes over the past 30 years in our understanding of the biology of breastcancer and its application to treatment are no less startling. Since 1975, we have witnessed an astounding evolution in our strategies to prevent,[1] diagnose,[2] and manage[3] a disease that affects the lives of so many in the United States[4] and around the world.[5] These efforts have generated many headlines and an occasional stumble. Nonetheless, they have had a dramatic impact on the lives of millions of people, and it is hoped that the rate of improvement will further accelerate in years to come. Historical Background
Randomized trials that were started in the 1970s disproved long-established dogmas emphasizing aggressive local therapy as the key for controlling breastcancer and preventing recurrence. As a consequence, most women with earlystage breast cancer are now able to preserve their breasts[6,7] and often undergo removal of just a few sentinel nodes from the axilla.[8] Further, a generation of studies starting with seminal trials published in the mid-1970s confirmed that adjuvant systemic therapy significantly reduces the risk of breast cancer recurrence, thereby confirming that invasive breast cancer must be viewed as a systemic disease from its inception.[9,10] A key tool underlying these accomplishments was the concept of the controlled clinical trial first introduced by Sir Richard Doll almost 60 years ago.[11,12] While many of the initialadjuvant breast cancer trials were too small to detect the small improvements that might be clinically significant, the systematic review exercises held by the Early Breast Cancer Trialists Collaborative Group every 5 years since 1985 have proven invaluable to confirm the survival benefit offered by adjuvant chemotherapy for the average patient with early-stage disease (especially younger women) and by tamoxifen(Drug information on tamoxifen) for any patient with estrogen receptor alpha (ER)-positive disease regardless of age.[3] In fact, National Surgical Adjuvant Breast and Bowel Project investigators have suggested that some form of adjuvant therapy should be considered for even women with invasive tumors of 1 cm or less and negative axillary lymph nodes,[13] although such recommendations have been considered too broad by many[14] and do not account for individual variability. Evolving Guidelines
In this issue of ONCOLOGY, Mina and Sledge chronicle the development of adjuvant systemic therapy for breast cancer since the 1985 National Institutes of Health Consensus Development Conference on early-stage breast cancer[15] and provide a comprehensive summary of critical advances in the last 20 years. Many of these developments have been distilled into expert panel consensus recommendations and clinical practice guidelines like those released every 2 years following the St. Gallen meetings and at least yearly by the National Comprehensive Cancer Network (NCCN). As a result, most clinicians have been exposed to the principles of evidencebased medicine,[16,17] and cancer specialists have become quite familiar with the concept of relative and absolute risk reduction. Medical oncologists often employ validated online instruments[18,19] for real-time discussions with patients regarding the actual average improvement in 10- year risk of cancer-related recurrence and mortality offered by various chemotherapy and endocrine regimens. Treatment guidelines and software tools have proven invaluable in clinicalpractice, but their use highlights some of the challenges oncologists face while translating population and clinical trial evidence into recommendations for individual patients. In most cases, we have information about the average risk and average benefit for a group of patients with similar riskstratification features who were given the same or similar therapy, when in fact the actual variability within such a group can be substantial. But it is now clear that breast cancer is actually a collection of diseases that share a common name but have an extremely diverse biologic behavior. This is unquestionably a disease for which onesize treatment will not fit all patients. Practitioners should therefore enthusiastically welcome the most recent update of the algorithms for adjuvant therapy selection issued by the 2005 St. Gallen consensus panel and the NCCN.[20,21] These panels have now placed primary emphasis on measures of potential response to therapy rather than risk stratification. The impact of these seemingly subtle changes goes beyond a simple switch in their order of consideration. Rather, these new guidelines reflect our increasing ability to move beyond prognostic factors to focus on individual predictive factors, ie, those unique features that could potentially determine an individual's likelihood of benefiting (or not) from specific therapies. Only then do the traditional markers-such as tumor size and lymph node status-that aid in prognostic but not predictive assessment come into play. Recent Advances
Seven years ago, the American Society of Clinical Oncology annual meeting was dominated by presentations of adjuvant high-dose chemotherapy trials that emphasized dose-intensification strategies for high-risk patients without much attention to individual biologic features. Fast forward a few years. Factors long thought to carry a modest prognostic value like expression of ER and the human epidermal growth factor receptor 2 (HER2) are now fully validated as strong predictive markers of response to specific therapies thatcould potentially provide a substantial survival benefit for the right patients.[ 3,22-24] Perhaps equally important, patients whose tumors lack expression of these specific markers and who are unlikely to be helped by these therapies can be spared from exposure to a costly "placebo." These types of studies are now accelerating in the molecular era as geneexpression profile assays appear to identify several major breast cancer subtypes with prognostic implications.[ 25] At least one commercial assay using paraffin(Drug information on paraffin)-embedded tissue has undergone strict assay standardization, and was shown in retrospective validation studies to have strong prognostic[ 26] and predictive[27] value for patients with lymph node-negative, ER-positive disease. The same assay also predicted the likelihood of pathologic complete response in patients treated with preoperative chemotherapy for locally advanced disease,[28] an area that merits further study. Looking Ahead
In 2006, the promise of selecting therapy based on tumor characteristics is now an established part of clinical practice. Until now, the reliability, accuracy, and reporting standards for many assays used in clinical practice have often not been given much needed attention. In many cases, immunohistochemistry assays have not been used as stand-alone tests. Rather, they have often been used to complement conventional pathology evaluation and confirm a specific diagnosis or to stratify patients among various prognostic groups. As new assays are introduced in clinical practice to serve as the sole determinant of therapy selection, pathologists and labs performing these assays will bear a much higher degree of responsibility for patient outcome. Equally important, clinicians using these results must understand the nuances of their use, going beyond a simple interpretation of a positive or negative test result. They must understand how the test was done, be familiar with potential variables that may affect results, and know what ques-tions to ask the pathologist. Concerns about the accuracy and reproducibility of assays of hormone receptor[29,30] and HER2[31,32] expression have been well documented, and we now understand that both presence and intensity of expression are critical features for determining benefit to biologically based therapy with tamoxifen, aromatase inhibitors, and trastuzumab(Drug information on trastuzumab) (Herceptin). In addition, much of our focus thus far has been on specific tumor characteristics, but there is now increasing recognition that individual host factors can influence response and toxicity to individual therapies.[33] In other words, proper assay performance by labs and pathologists and interpretation of results by oncologists is crucial, and the need to address platform standardization and reproducibility issues is expected to be further compounded as we introduce more complex gene array[34] and pharmacogenetic assays[35] into routine breast cancer clinical practice. Addressing these technical issues[ 36,37] and disseminating this information to all end users[38] is a big challenge, but these are indeed exciting times for us and for our patients. We look forward to the next 30 years!

—Antonio C. Wolff, MD, FACP
—Nancy E. Davidson, MD