|
||||||||
|
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
Underneath John’s Beatle-esque haircut is a brain that spews megahypothesis after megahypoth- esis about breast cancer research from A to Z. My latest conversation with Professor Robertson is included in this program, and I have spent the last five days reading and rereading a 2001 paper* by his Nottingham group that he discusses during the interview. As with many of the articles in Mark Lippman’s Breast Cancer Research and Treatment (see an upcoming issue for an interview with Dr Lippman), this paper is loaded with fascinating but very complicated data points. Truth be told, I don’t fully understand John’s interpretation of this study, but it makes intuitive sense, which is always dangerous. John’s bottom line favors starting an agent with greater antitumor activity (an aromatase inhibitor) rather than one with a longer safety track record (tamoxifen). Like Soon Paik’s mining of NSABP trials B-14 and B-20 to document the benefit of the Oncotype DX™ assay, the Nottingham paper focuses on patients treated some time ago — in this case during the pre-adjuvant endocrine therapy days, when we could measure ER more accurately than we do today but did not offer hormone therapy until relapse. Of great interest were the disease-free survival (DFS) curves of patients with ER-positive tumors that did not respond to tamoxifen when treated for metastatic disease. These patients — with a median DFS of 21 months — relapsed much earlier than those who were endocrine-responsive (Figures 1, 2). The patients also had significantly shorter survivals. This pattern of early relapse and rapid downhill course is reminiscent of HER2-positive disease, and it is quite possible that many of these patients actually had HER2-positive tumors, although that information is not available.
It is fascinating that the findings from this relatively obscure paper have numerous critical research and practice implications:
A sequencing study randomly assigns patients up front to either five years of an AI or two or three years of tamoxifen followed by an AI, whereas a switching trial randomly assigns patients who have completed two to three years of tamoxifen to either continue tamoxifen or switch to an AI. Presumably, if one waits two to three years to start an AI, the response rate is likely to be higher because many patients with unfavorable, “endocrine unresponsive” tumors will have relapsed before that time. The major random- ized prospective sequencing trial (BIG 1-98) has not yet reported data on this question, but the Nottingham study suggests that hypothetical models making indirect comparison between trials such as ATAC and the switching trials are flawed because these studies focus on different patient populations.
The Nottingham data set demonstrates that in patients with ER-positive tumors who later progressed on endocrine therapy, half of the relapses occurred within 21 months. The hazard curves for recurrence in ATAC (Figure 5) clearly demonstrate a significant difference during the first two years, suggesting that anastrozole may be effective in treating some of these tamoxifen-nonresponsive tumors. I look forward to asking oncologists, including the “TECHIES” on the ASCO AI technology assessment panel, what they think about the Nottingham data set and John’s astute interpretation of it. I am particularly interested in learning whether they believe it is possible to select postmenopausal patients who should start treatment with tamoxifen rather than an aromatase inhibitor.
A small yet vocal minority of investigators believes that patients with lower-risk, node-negative, receptor-positive tumors are candidates for a sequencing strategy of two to three years of tamoxifen followed by an AI. A major part of the rationale for this line of thinking is the long-term safety data we have for tamoxifen versus the AIs. However, when one focuses on antitumor effect, following John’s logic, those few node-negative patients who do relapse are more likely to do so in the first two years of therapy and will not be salvaged by a delayed AI approach. In the next few weeks, our CME group will send out the edited proceedings of a Think Tank roundtable that we hosted recently in Miami with 12 renowned breast cancer investigators. The group was lively, to say the least, and when we discussed optimal long-term endocrine therapy for postmenopausal women with ER-positive tumors, I thought we would need to borrow Jerry Springer’s bodyguards to separate the combatants. John sat quietly through the verbal melee, and when his turn came, he quoted the 2001 Nottingham paper, and when he described the short DFS of the tamoxifen-unresponsive patients, you could see the cranial light bulbs go on around the table. Cliff Hudis interjected, “This looks like ER-negative disease.” Exactly, except in this case, there may actually be an endocrine therapy (AIs) to help many of these patients. — Neil Love, MD * Cheung KL, Nicholson RI, Blamey RW, Robertson JFR. Selection of primary breast cancer patients for adjuvant endocrine therapy — Is oestrogen receptor alone adequate? Breast Cancer Res Treat 2001;65(2):155-62. Abstract
|
|
 |
Terms and Conditions of Use and General Disclaimer Copyright © 2006 Research To Practice. All Rights Reserved |
 |
Immediately after my first
interview with Professor John
Robertson in a dusty New York
meeting room some years ago, I
took a long walk in Central Park to
ponder the man’s words.
