Journal of Clinical Oncology, Vol 19, Issue 4
(February), 2001: 980-991
© 2001 American Society for Clinical Oncology
Computer Program to Assist in Making Decisions About Adjuvant Therapy for Women With Early Breast Cancer
By Peter M. Ravdin,
Laura A. Siminoff,
Greg J. Davis,
Mary Beth Mercer,
Joan Hewlett,
Nancy Gerson,
Helen L. Parker
From the University of Texas Health Sciences Center, San Antonio, TX, and Case Western Reserve, Cleveland, OH.
Address reprint requests to Peter M. Ravdin, MD, PhD, Division of Oncology, University of Texas Health Sciences Center at San Antonio, 7703 Floyd Curl Dr, San Antonio, TX 78284; email: pravdin@ swog.org.
 |
ABSTRACT
|
|---|
PURPOSE: The goal of the computer program Adjuvant! is to allow health professionals and their patients with early breast cancer to make more informed decisions about adjuvant therapy.
METHODS: Actuarial analysis was used to project outcomes of patients with and without adjuvant therapy based on estimates of prognosis largely derived from Surveillance, Epidemiology, and End-Results data and estimates of the efficacy of adjuvant therapy based on the 1998 overviews of randomized trials of adjuvant therapy. These estimates can be refined using the Prognostic Factor Impact Calculator, which uses a Bayesian method to make adjustments based on relative risks conferred and prevalence of positive test results.
RESULTS: From the entries of patient information (age, menopausal status, comorbidity estimate) and tumor staging and characteristics (tumor size, number of positive axillary nodes, estrogen receptor status), baseline prognostic estimates are made. Estimates for the efficacy of endocrine therapy (5 years of tamoxifen) and of polychemotherapy (cyclophosphamide/methotrexate/fluorouracil–like regimens, or anthracycline-based therapy, or therapy based on both an anthracycline and a taxane) can then be used to project outcomes presented in both numerical and graphical formats. Outcomes for overall survival and disease-free survival and the improvement seen in clinical trials, are reasonably modeled by Adjuvant!, although an ideal validation for all patient subsets with all treatment options is not possible. Additional speculative estimates of years of remaining life expectancy and long-term survival curves can also be produced. Help files supply general information about breast cancer. The program’s Internet links supply national treatment guidelines, cooperative group trial options, and other related information.
CONCLUSION: The computer program Adjuvant! can play practical and educational roles in clinical settings.
|
INTRODUCTION
|
|---|
THE PRACTICE OF medicine has been moving toward more evidence-based and quantitative formats. This trend has been occurring during an era in which there are increasing data concerning the efficacy and safety of therapies from clinical trials. A growing number of tools have been developed that produce numerical estimates of the probability of outcomes, such as the Gale model, which produces estimates of the risk of developing breast cancer in individual women in the general population.1 Somewhat paradoxically, these numerical tools have the potential to humanize medicine; for example, if the results are shared with the patient in a comprehensible format, the patient can become an informed partner in making decisions about different therapeutic options.
These tools have great potential in making treatment decisions in the adjuvant therapy for individual patients with early breast cancer. Breast cancer adjuvant therapy is an important area for advancing techniques of evidence-based treatment decision making for several reasons. First, there are often several reasonable therapeutic options available. Specifically, in this clinical situation, the decision can often be between options of no additional therapy, chemotherapy, hormonal therapy, or chemoendocrine therapy. There are also options concerning the choice of the specific type of chemotherapy. Second, the degree of benefit in terms of disease-free survival (DFS) and overall survival (OS) can be small and uncertain enough to make the decision as to whether to receive a given option something for which individual patients might express very different preferences. Studies have shown that many patients are not given quantitative information about their prognoses with and without adjuvant therapy and often make inaccurate estimates.2,3 Without such information as a point of reference, these patients are not adequately informed partners in deciding whether and what type of adjuvant therapy might be most appropriate. Furthermore, the oncologists often are uncertain about what quantitative prognostic estimates to apply to any given patient.4
To address this problem, we developed a decision aid—a simple-to-use computer program entitled Adjuvant!. This program is designed to produce prognostic estimates of outcome with and without therapy, based on the estimates of individual patient prognosis and the efficacy of different adjuvant therapy options. The computer program presents this information on the computer screen for the physician as well as printed pages for use with patients. This article describes the program in detail. The use and impact of this tool in real-time treatment decision-making encounters with stages I to III breast cancer patients is currently being investigated.
|
METHODS
|
|---|
The program Adjuvant! version 2.2 runs in the Windows 95, 98, and NT environments. It requires a 32-bit 486 or higher processor. It is self-installing and requires 1.5 megabytes of hard drive space. Output requires a printer, and we recommend a color printer. Guidelines, cooperative group clinical trial information, literature search engines, and websites for patient information materials can be accessed from Adjuvant! if there is an Internet connection.
The fundamental purpose of Adjuvant! is to supply estimates of the net benefit to be expected for systemic adjuvant therapy for individual breast cancer patients. These estimates of benefit are obtained by estimating the reduction of risk of negative outcomes for individual patients. This is done by estimating a patient’s risk of negative outcome (death or relapse) and then multiplying that by the proportion of negative events that a given adjuvant therapy is known to prevent. Estimates of prognosis are based mainly from the Surveillance, Epidemiology, and End-Results (SEER) registry estimates of outcome for breast cancer patients in the general population in the United States. Estimates of the efficacy of therapy are based mainly on the proportional risk reductions (PRR) that can be obtained from the 1998 Overview summaries of the effectiveness of adjuvant therapy based on data from nearly all randomized clinical trials.5,6
A simplified view of how this is done is easy to understand. For example, consider an individual patient who, because of tumor size and number of nodes, has a 60% risk of dying of breast cancer at the 10-year follow-up. This patient can be given an adjuvant therapy that can achieve a PRR of 25%. This should reduce her breast cancer mortality by 60% times 25%, which equals 15%. A subtle but important point is that the Overview PRR are for effects of the therapy in the annual risk of negative outcome. For a patient with a 60% risk of death at 10 years, the average annual risk of breast cancer–related death is 8.8%. Therefore, the expected survival would be 91.2% of the preceding year. This would result in survivals of 91.2%, 83.3%, 76.0%, 69.3%, 63.2%, 57.7%, 52.7%, 48.0%, 43.8%, and 40%, for years 1 to 10, respectively. However, after an adjuvant therapy affording a PRR of 25%, the annual risk of breast cancer–related death would be 6.6% (8.8% reduced by 25%). This would result in survivals of 93.4%, 87.3%, 81.6%, 76.2%, 71.2%, 66.5%, 62.2%, 58.1%, 54.3%, and 50.7% for years 1 to 10, respectively. The expected benefit would not be 15% but only 10.7%. A second consideration that can reduce further the impact of adjuvant therapy is competing non–breast cancer–expected natural mortality. Obviously, if a high percentage of patients were expected to die of non–breast cancer–related causes, then there would be fewer patients that might obtain the net benefit of adjuvant therapy.
Adjuvant! takes these effects into account by using an actuarial life table technique for calculating survival.7 At the time of diagnosis (year 0), it starts with a survival probability of 100%. Then the survival probability 1 year later is calculated as 100% minus the percentage of patients that die over a year’s time of natural causes (for women of the patient’s age taken from United States mortality statistics) and also minus the percentage of patients expected to die over a year’s time of breast cancer. This process is calculated iteratively year by year with the probability of survival being calculated in each succeeding year. Of course, year-by-year the percentage risk of natural mortality gradually increases (as the patient grows older), and the year-by-year risk of dying of breast cancer (after a more complex pattern) eventually begins to decrease. In addition, estimates of the effectiveness of adjuvant therapy can be used to change the year-by-year estimates for breast cancer–specific mortality. Together this methodology can allow one to estimate the relative contributions of natural and breast cancer–specific mortality to the overall mortality of given patients in scenarios where they receive adjuvant therapy and where they do not.
Estimates of Prognosis
The estimate of risk of breast cancer–related death at a 10-year follow-up is supplied by Adjuvant! based on the tumor size, the number of involved nodes, and to a minor extent, estrogen receptor (ER) status. These estimates are based on an analysis of data from the SEER tumor registry database, which follows approximately 10% of all breast cancer cases in the United States, as well as information from a number of sources including the Overview meta-analyses.5,6 The primary data used in this analysis were individual demographic, staging, and outcome information from the SEER-9 Public Registries Files from August 1998, as provided on the April 1999 SEER*Stat2 Version 2.0 CD ROM. Patients included in the initial analyses were women who had invasive, unilateral, noninflammatory disease, had undergone definitive surgery (with radiation if lumpectomy), and had axillary staging with at least six nodes sampled. For inclusion in the analysis, patients must have had a known tumor size, number of nodes sampled, and number of nodes positive for tumor.
There are some limitations of the SEER registry-derived data. SEER registry-defined information has vital status during follow-up but does not have information about the type of adjuvant therapy received, relapse status, or reliable cause of death information. Thus, only survival analyses can be done from SEER data. Estimates of breast cancer–related mortality made using SEER data must be derived indirectly from total survival after adjustment for expected age-adjusted natural mortality. This analysis was done using the SEER*Stat2 software.
As a first part of this analysis, the effect of age on breast cancer–related mortality at 5 years was examined. The results for this analysis for patients with small node-negative tumors (tumor size 0.1 to 1.0 cm), average node-negative tumors (tumor size 2.1 to 5.0 cm), and intermediate risk node-positive tumors (one to three positive nodes and tumor size 2.1 to 5.0 cm) are shown in Fig 1. This analysis shows that young patients (in their 20s and 30s) seem to have a worse prognosis overall, and that patients in their 70s and 80s have a comparatively good prognosis. The method of deriving breast cancer mortality by adjustment of total mortality for expected natural mortality gives nonsensical negative breast cancer–specific mortalities for older patients with small node-negative tumors. A reasonable explanation for this effect is that older patients (in their late 60s, 70s, and 80s) who have the type of medical access that allows the discovery of such early tumors and who get definitive treatment are healthier and overall have less natural mortality than might be expected for their age. To eliminate this source of bias, and considering that women younger than 35 seem to have stage by stage a somewhat worse prognosis than older women, the analyses done for assigning stage-related prognoses were restricted to women in the SEER registry from 35 to 59 years of age.
The SEER data were then used to calculate the breast cancer–related mortality for patients at the 5-year follow-up and to extrapolate breast cancer mortality at the 10-year follow-up. Such extrapolations were necessary because presently the available SEER data have only complete tumor size and number of nodes for the last 9 years of data collection and ER status information for the last 7 years. The results of the analysis for 5 years of follow-up can be seen in Table 1. An unexpected aspect of this analysis is that ER status seems to be a strong prognostic factor, which is inconsistent with the literature, which suggests that ER status is not strongly prognostic. This discrepancy can be explained in part by the short follow-up. The annual hazard rates for patients when considering ER can be seen in Fig 2. Clearly, the time dependence of the hazard is markedly different for ER-negative and ER-positive tumors. Patients with ER-negative tumors annual risk of mortality peaks in the first 5 years, and patients with ER-positive tumors have a peak hazard that occurs later.
Using SEER data, annual hazard estimates derived from the 1998 Overviews5,6 and other sources,8,9 it was estimated that for ER-positive cases, the average annual hazard rates in the first 5 years were half that of the average rates for years 6 through 10; for ER-negative cases, the average annual hazard rates in the first 5 years were twice that for years 6 through 10. Using these approximations, extrapolated estimates of the breast cancer–specific mortality rates at the 10-year follow-up could be made for ER-positive and ER-negative cases. Again, particularly for node-negative cases, ER status was a prognostic variable but to a lesser degree ( Table 2). Extrapolations of breast cancer–specific mortality at 10 years of follow-up were also made for all cases (regardless of ER status) by using the average observed annual hazard rates in years 6, 7, and 8 as that for years 9 and 10, and then using these hazard rates to extrapolate the cumulative breast cancer–specific mortality at 10 years’ follow-up. The extrapolated 10-year breast cancer–specific mortality rates derived from cases regardless of ER status for low-risk tumors lay between those of ER-positive and negative cases, but for high-risk tumors, this was not the case. This is might occur either because cases without ER status had some special feature (approximately 25% of the cases had no reported ER), or because the extrapolations for the ER-positive and -negative subsets were in error.
These uncertainties for the cases with defined ER together with the fact that estimates for cases irrespective of ER were based on more cases with longer follow-up, led to the decision to base the estimates of breast cancer–specific mortality on all cases irrespective of ER. Adjustments for slightly better and worse outcomes for ER-positive and ER-negative cases, respectively, were made by adjusting ER undefined estimates using a modest 1.3 relative risk, which large long-term studies have suggested are appropriate in untreated node-negative patients.10,11
The SEER estimates of stage-specific mortality are influenced by the fact that many of the patients received adjuvant systemic therapies. This adjustment was made by modifying the estimates (increasing) for the impact of adjuvant therapy that improved the outcome of stage 1 cases (in which many patients would not have received adjuvant chemotherapy) by a proportional 15%, and by a proportional 30% for stages II and III cancers. Estimates of breast cancer–specific mortality at 10 years of outcome for untreated patients (Table 2) were derived by this method. Such adjustments may introduce small systematic errors into the estimates. For example, for patients with stage 1 breast cancer, considering the average estimate of PRR provided by adjuvant therapy of approximately 30% ( Table 3) and the estimate that approximately half of the patients receive such therapy, the adjustment of 15% was made. The true adjustment might be somewhere between 0% (if no patients receive a systemic adjuvant therapy) and 30% (when all patients receive adjuvant therapy). Thus, a systematic relative error of 15% may have been introduced, which is modest.
View this table:
[in this window]
[in a new window]
|
Table 3. Risk of Breast Cancer–Specific Mortality and Estimates of Breast Cancer Recurrence (distant, local, or contralateral) at 10 Years Follow-Up
|
|
SEER does not collect data on relapse, and there are no large population-based sources of relapse data. Relapse estimates were derived from the mortality estimates by noting that breast cancer–specific relapse rates (defined as any recurrence (distant, regional, or contralateral) must be slightly higher than mortality rates because there are always some patients who have relapsed but not yet died. Also, there are some types of relapse (for example, contralateral) that are not usually associated with eventual mortality. Considering that on average, mortality occurs approximately 3 years after relapse and that the annual hazard of contralateral breast cancer is approximatley 0.65%, the annual hazard of relapse is estimated as 1.3 times the breast cancer mortality hazard plus 0.65%. This formula was used to convert estimates of 10-year cumulative hazard of mortality to the 10-year cumulative hazard of relapse. The estimates of breast cancer–specific mortality and breast cancer relapse or recurrence used by Adjuvant! are presented in Table 3.
Derivation of Estimates of the Efficacy of Adjuvant Therapy
The estimates of PRR used by Adjuvant! either are taken directly from the 1998 Overviews5,6 of adjuvant therapy of early breast cancer or were derived indirectly from these estimates. The PRR for the 1998 Overview estimates were for recurrence and death.
The definition of recurrence used in the Overviews5,6 was first reappearance of breast cancer at any site (local, contralateral, or distant), and deaths due to breast cancer (if breast cancer had not been documented). Deaths because of unknown causes were included as breast cancer–related deaths. Deaths due to non–breast cancer–related events were not included as recurrence events, but patients were censored at that time. A careful definition of recurrence is important because such a wide variety of definitions are used in different trials, sometimes including non–breast cancer–related deaths and second primary cancers.
The definition of death used in the Overviews was death due to any cause. This definition is not ideal inasmuch as it mixes both breast cancer–related and nonrelated causes of death. This is not the definition used by Adjuvant!. Adjuvant! presents the user with PRR for breast cancer–specific mortality. The justification for this is as follows: Expected non–breast cancer death rates at 10 years’ follow-up for women in good health are approximately 1%, 2%, 6%, 12%, and 24% in their 30s, 40s, 50s, 60s, and 70s, respectively.12 Most estimates of efficacy of adjuvant therapy are derived by the Overviews5,6 from the outcomes of women younger than 70 and scenarios in which the rate of breast cancer–related death is much greater than that of non–breast cancer–related death. Thus, these estimates can be used as a reflection of the PRR of breast cancer–related mortality without modification, as performed by Adjuvant!. However, PRR given by the Overview for women 70 and older might be expected to be low because the majority of deaths would be expected to be non–breast cancer related. Analyses reported in the 1998 Overviews5,6 showed adjuvant therapies do not have a significant effect on non–breast cancer–related mortality. Perhaps the lack of benefit on overall mortality (in terms of PRR) of chemotherapy in women 70 and older is partially because a substantial percentage of these deaths would be expected to be non–breast cancer–related and might obscure an effect on breast cancer–related mortality. Adjuvant! uses the PRR of average women 50 to 69 years of age for women 70 and older, but informs the user of the controversial nature of this choice if an age of 70 or older is selected for the patient.
The 1998 Overviews5,6 report direct comparisons as made in randomized trials but do not make indirect comparisons that might be inferred. Thus, PRR are given in the 1998 Overviews5,6 for effectiveness of adding tamoxifen or polychemotherapy, but they are not given for comparisons of no therapy versus combined therapy with both chemotherapy and tamoxifen. Indirectly, such estimates can be inferred because the 1998 Overviews suggest that the benefit of adjuvant chemotherapy or tamoxifen occurs independently of whether the other modality is used. Thus, the PRR for chemoendocrine therapy are derived as the product of the PRR for polychemotherapy and tamoxifen. The actual equation used is:
equation
For example, in premenopausal women with ER-positive tumors, the PRR for mortality given by the 1998 Overviews are 20% and 28% for polychemotherapy and tamoxifen, respectively. The indirectly inferred PRR for chemoendocrine therapy is 100 - (80%) x (72%) = 42%.
The 1998 Overview5 of polychemotherapy trials reports results that are largely dominated by trials with cyclophosphamide/methotrexate/fluorouracil (CMF)–like regimens. The Overview5 also reports that anthracycline-based regimens deliver approximately 11% more PRR for recurrence and mortality compared with non–anthracycline-based regimens. The Overview does not present a separate analysis of the PRR afforded by anthracycline-based regimens versus no polychemotherapy, so this must again be indirectly inferred. This is performed using the above equation. This same problem is addressed when including the impact of adjuvant taxanes. Here the results are dependent on the short-term follow-up of one trial, which was not included in the 1998 Overviews (as the user is reminded by Adjuvant! if they choose adjuvant therapy that includes a taxane). Because the estimates of PRR at less than 5 years are generally much more optimistic than seen at 10 years of follow-up, the impact on PRR for mortality and recurrence is estimated to be 11%, not the 22% reported in the early analysis of that trial.
Adjuvant! provides estimates of PRR for recurrence and breast cancer–related mortality. These estimates are selected based on menopausal status, ER status, and the type of chemotherapy selected (CMF-like, including an anthracycline or both anthracycline and taxane). The estimates are listed in Table 4. The user of Adjuvant! is not constrained to use the estimates provided but can directly enter estimates of efficacy for individual patients or entire sets of estimates that can be saved for later use.
|
RESULTS
|
|---|
The program that we have developed, Adjuvant!, begins with a statement about its purpose, limitations, and a suggestion that it be used by health professionals and not directly by patients themselves, because of the importance of accurate information entry from sources such as pathology reports. After acknowledging these statements, a second screen opens stating that the program makes estimates for women with invasive breast cancers that are unicentric, unilateral, and for whom special considerations must be made using the "help section" if the tumor is a special histology subtype (pure tubular, pure papillary, or pure mucinous) or an inflammatory breast cancer. It states that the program makes estimates of outcome for patients after definitive tumor resection and axillary node dissection but before any systemic adjuvant therapy (neoadjuvant therapy). Furthermore, patients must not have known residual or metastatic disease. For patients receiving breast-conserving surgery, there is an assumption that radiation therapy is planned. After these requirements are acknowledged, the program opens to the main screen.
The main screen ( Fig 3) has four major components: (1) a section that allows patient information to be entered and provides an estimate of the risk at 10 years’ follow-up of breast cancer–related death or relapse unconfounded by therapy or competing causes of morbidity or mortality; (2) a section that provides efficacy information (in terms of PRR) for different adjuvant therapy options; (3) a section that shows the resulting projections of outcome in numerical and graphical format; and (4) a tool bar that allows the user to save patient information, print results, and access the program’s help files.
Entering Patient Information
The major parameters used for adjuvant therapy decision making must be entered into the program’s Patient Information section. The Age is the chronologic age (in years) of the patient. This information is used by the program to calculate the expected natural mortality. Also, it is used to produce the default estimate of menopausal status. Menopausal Status is (by default) assigned as postmenopausal for women  50 years old but can be overridden. This information effects estimates of the effectiveness of adjuvant chemotherapy. Comorbidity is an estimate of the general health of the individual for whom the estimates are being made—the default is Minor Problems. The possible choices are Perfect Health, Minor Problems, Average for Age, Major Problems (+10), Major Problems (+20), and Major Problems (+30). There are a broad range of estimates because the chronologic age clearly does not define natural mortality rates if comorbidity is not considered. The age-specific mortality rates for average women in the United States population12 (in Adjuvant! as Average for Age) are an overestimate for most women. Much of the mortality at any given age is driven by preexisting health problems. Therefore, most women have natural mortality rates that are predictably better than those for the general population. Adjustment for lack of comorbidity is because the actuaries have recognized (but not precisely defined) a population of select patients who have no comorbidity and who have lower short-term mortality rates.13,14 In older women, this may be only one third of the age-adjusted average initially but rises to the average within approximately 10 to 15 years. This select population’s natural mortality estimates are provided by selecting Perfect Health. Adjuvant! selects the default Minor Problems, which is the average of Perfect Health and Average for Age.
ER Status is used by the program as a crucial parameter in determining the efficacy of systemic adjuvant therapy options and in a minor role, in determining prognosis. Although Undefined is the default, it is anticipated that essentially all patients will have a defined ER status. ER status should be entered as either positive or negative by any criteria used by the laboratory making the measurement. PGR (progesterone receptor) status does not influence the entry if the patient is ER-positive. There is some uncertainty as to whether PGR status should influence what is entered if the patient is in the unusual situation of being ER-negative but PGR-positive (only approximately 5% of breast cancer patients are in this category). The efficacy of adjuvant therapy in the subset of patients who are ER-negative but PGR-positive is not well defined, but based on the efficacy of endocrine therapy in patients with metastatic disease with this receptor status, it is probable that these patients behave like ER-positive patients. Therefore, the program suggests that such patients be entered as ER-positive.
Tumor size is the greatest diameter of the tumor measured in centimeters. Frequently, there are several special issues that come up. One is whether, in a tumor with both intraductal and invasive components, the tumor size used in estimates should be that of the entire tumor or that of the invasive component. The data from the SEER registry are based on the size of the invasive component. Because the prognostic estimates made by Adjuvant! are based primarily on SEER data, the size of the invasive component of the tumor should be entered. Positive Nodes are the total number of positive axillary nodes reported from the axillary dissection (usually levels I and II). Number of positive nodes, together with tumor size, are the major factors used to make estimates of patient prognosis. If the patient has had a sentinel node biopsy, it is assumed that such a biopsy was done by an experienced surgical team with a low false-negative rate.15 If, under these circumstances, the node was negative, then enter 0. If the node was positive, then an estimate of number of nodes cannot be made unless there was an axillary node dissection. An issue in the evaluation of nodal status is the prognostic significance of aggregates of cancer cells within the lymph nodes missed on initial sectioning but identified later on resectioning or with special immunochemical stains. It seems that discovery of nodal involvement on standard histopathologic review does imply that the patient has a worse prognosis,16,17 although this is less clear for such nodal metastases identified by immunohistochemical18–20 or molecular biologic techniques.21 If the patient has undergone neoadjuvant therapy, then the axilla nodal status may have been significantly downstaged so that Adjuvant! cannot produce an estimate of prognosis and should not be used to make prognostic estimates.
The category 10-Year Risk is provided by Adjuvant! based on the tumor size, the number of involved nodes, and ER status. It is largely derived from data from the SEER tumor registry database22 as detailed in Methods. These estimates can be accepted or modified based on additional prognostic information. This can be done by directly entering a prognostic estimate from a literature source into the 10-Year Risk Box or using a calculator (the icon to the left of the 10-Year Risk).
The Prognostic Factor Impact Calculator (PFIC) uses a Bayesian approach to adjust a prior prognostic estimate with information from a patient’s prognostic factor test if the prevalence of a positive test and the relative risk conferred is known. Using this calculator, the independent relative risk (of tumor size and number of nodes) conferred, the prevalence of a positive test result (suggesting an unfavorable prognosis), and the baseline prior-risk estimate are entered. The calculator then produces 10-year prognostic estimates of outcome for patients with a favorable or an unfavorable prognostic test. This is done using the following formulas:
equation
equation
equation
An example of the use of PFIC can be seen in Fig 4.
View larger version (144K):
[in this window]
[in a new window]
|
Fig 4. Example of the PFIC refining the 10-year mortality estimate of 10% for the impact of a prognostic factor that confers a relative risk of 2 and with a prevalence of high-risk results of 20% (similar to Her2 amplification by fluorescence in situ hybridization by some methodologies).
|
|
Efficacy of Adjuvant Therapy
These estimates come directly from the 1998 Overviews,5,6 were indirectly derived from them or, in the case of the estimates of efficacy of adjuvant chemotherapy based in taxanes, derived from additional information in the literature. The details of these derivations can be found in "Methods." The Adjuvant! program gives the user the efficacy of various adjuvant regimens: 1) endocrine therapy (which refers to 5 years of adjuvant tamoxifen); 2) chemotherapy (polychemotherapy), either CMF-like, anthracycline-based, or anthracycline- and taxane-based. These are given by subsets defined by ER and menopausal status (the Overview uses age < or to 50 rather than menopausal status). One need not accept these values but can toggle these values or create entire sets of customized estimates of estimates that can be saved for later use.
Resulting Graphs
The section showing the Resulting Graphs allows the viewing of outcomes for survival in terms of OS at 10 years, estimates of remaining life expectancy, and long-term survival curves. It also allows outcomes to be viewed for DFS at 10 years. The OS and DFS at 10 years are presented as bar graphs and are the primary outputs of the program. Projections of outcomes beyond 10 years, although interesting, are speculative.
The bar graphs are used to show OS at 10 years’ follow-up. They show what percentage of patients are alive at 10 years, what percentage die of breast cancer, what percentage die of non–breast cancer causes of death, and an estimate of the increased percentage of patients alive at 10 years because of specific adjuvant therapy chosen. Viewing OS in this format allows a perspective on what role breast cancer mortality has within the next 10 years of the patient’s life. For older node-negative patients with small tumors, it is often quite small in comparison with other non–breast cancer mortality. While viewing these bar graphs, one can toggle between different adjuvant treatment options to examine the impact of the benefit of therapy for a variety of therapeutic options. These graphs include text showing net benefit to tenths of a percent. Although the projections are not this accurate, this feature allows the physician to see how small the outcome differences are without having rounding errors obscure them.
The bar graphs showing DFS show what percentage of patients are alive without breast cancer at 10 years, what percentage are expected to relapse with breast cancer, what percentage die of other non–breast cancer causes of death. These estimates are shown for scenarios where adjuvant therapy is either used or not, allowing one to view the additional percentage of patients who remain disease-free at 10 years because of adjuvant therapy. Viewing DFS in this format allows a perspective on the risk of relapse in the next 10 years. While viewing these bar graphs, it is possible to toggle between different adjuvant treatment options and examine the impact on DFS.
Another format for viewing the impact of adjuvant therapy is to view the impact on expected average remaining-life expectancy. Bar graphs showing life expectancy estimates provide the patient’s average life expectancy in years assuming the cancer had never happened and then allows a comparison between this figure and life expectancy for the patient in two ways: first if she is not treated by adjuvant therapy and then if she is treated with the selected form of adjuvant therapy. These graphs are speculative because they depend on unverifiable estimates of long-term mortality rates and the assumption that there will be no major improvements in breast cancer treatment and that the treatment of other health problems may have major effects on life expectancy. It seems likely that such changes will occur incrementally. By providing the information on life expectancy in this way, we allow the user to put into perspective the impact of the breast cancer on the patient’s remaining life expectancy and what the net impact of adjuvant therapy might be. These estimates are made to tenths of a year for the same reasons. The survival curves show the projected survival to 30 years postdiagnosis in scenarios in which (1) the cancer had never happened, (2) the cancer was not treated by adjuvant therapy, and finally, (3) the cancer was treated with the selected form of therapy. These curves are also speculative, because of the many uncertainties beyond 10 years’ follow-up.
Output as Hard Copy
The print option allows a combination of any of these graphs to be printed. The bar graphs of DFS and OS are printed in a format that has been specifically designed for easy patient interpretation. The format and graphic presentation of the printed format of Adjuvant! was evaluated during its development. Issues of concern were (1) wording, (2) layout, including font size and color, (3) use of bar or pie graphs, and (4) the amount and order of information to be presented. We first tested our Decision Guide with 24 breast cancer patients who were returning for either adjuvant treatment or follow-up care. They ranged in age from 47 to 74 years, were a racially diverse group, and had a wide range of education levels. After making changes iteratively, the format was then tested with 25 patients who were in the process of considering their adjuvant therapy options. Examples of the type of refinements made were replacing the term "adjuvant" therapy with "additional" therapy because the term adjuvant was confusing to most patients. Likewise, the term "relapse" was replaced by the phrase, "the cancer coming back." Originally, we labeled the graphs in terms of percentages, however, some patients felt uncomfortable with the term "percent." We replaced the percentages with a descriptive phrase: number of women out of 100. This was universally understood by patients. The format of the graphics was also altered to make them easily readable. A bar chart format was selected because patients rated this as most easily understood ( Fig 5).
View larger version (38K):
[in this window]
[in a new window]
|
Fig 5. Example of Adjuvant! print out. Print out also includes information on age, tumor size, nodes involved, and ER status.
|
|
Tool Bar Resources
There are a number of useful functions that can be performed using the tool bar at the top of the main screen. These include the ability to save individual patient analyses in files for later reference and to control and tailor the sheets printed out for individual patients. It is possible also to create custom sets of efficacy estimates for specific adjuvant therapy plans or because of disagreement with the sets supplied by Adjuvant! (up to four such sets can be created) and to save these sets of efficacy estimates for later use. There are also extensive help files. The files include sections explaining how to use the program, how it works, and how the baseline prognostic and efficacy estimates are determined. Included in the help files is a discussion of prognostic factors and how they might be used (the possibility of using Her2, measures of proliferation, and histologic grading are discussed). There is also a section that allows direct Internet links to treatment guidelines of the National Cancer Institute (NCI)/Physician Data Query System and the National Comprehensive Cancer Network, American Society of Clinical Oncology (ASCO) prognostic factor guidelines, Cooperative Group clinical trials, search engines for scanning the medical literature, and sources of information for patients.
|
DISCUSSION
|
|---|
The development of the tool Adjuvant! represents an evolutionary process.23-25 Constructing and using such a tool requires questioning the nature of the Overview estimates of the efficacy of adjuvant therapy and how to use them, how in a practical sense prognostic information is to be used, and ultimately, how best to present estimates of DFS and OS to patients. Such a program has use both in a practical and educational sense because it allows health care professionals and their patients to understand better the probable impact of adjuvant therapy and how this is derived from estimates of prognosis and the efficacy of adjuvant therapy.
The estimates of the efficacy of adjuvant therapy as derived by Adjuvant! show that adjuvant therapy for early breast cancer has a modest but important impact. To present the results in quantitative terms is important because it allows the patient to participate in the decision-making process. The most common scenario in adjuvant therapy decision making today is that of a patient at the age of 65 (the approximate median age of breast cancer patients) who is ER-positive and who has a node negative, 1- to 2-cm tumor. Adjuvant! can be used to estimate that such a patient without therapy has approximately a 10% chance of dying of breast cancer, a 10% chance of dying of other causes, and an 80% chance of surviving another 10 years. Endocrine therapy (5 years of tamoxifen) improves her chances approximately 2.4%, and combined chemoendocrine therapy with both anthracycline-based chemotherapy regimen followed by tamoxifen approximately 3.6%. Should such a woman get the combined program for the 1% survival advantage? Studies of the opinions of breast cancer patients who have had chemotherapy suggest that many women would select to receive chemotherapy for such a small survival advantage, although many would not.3,26 The same analysis done by Adjuvant! for breast cancer–free survival at 10 years suggests a 70% breast cancer–free survival that would be improved to 78% by tamoxifen and 10% by the addition of chemotherapy. These general results match the modest effects seen in clinical trials. For example, in National Surgical Adjuvant Breast and Bowel Project trial B-16, which compared adjuvant tamoxifen to four cycles of doxorubicin and cyclophosphamide followed by tamoxifen in node-positive endocrine therapy in responsive patients older than 50, modest but significant advantages for the combined approach were seen. The typical patient in this study was in her early 60s, had one to three positive nodes, and a T2 primary tumor. When these parameters are entered into Adjuvant!, the program projects a 59% overall survival at 10 years for patients receiving tamoxifen with an improvement of 5% for the patients receiving the combined therapy. This is close to the report in the trial (57% and 62% OS at 10 years for the tamoxifen only and combined groups).
The ideal validation of results presented by Adjuvant! is problematic because there are no population-based databases that have reliable information on the crucial elements needed: detailed individual staging information, breast cancer–specific mortality and recurrence information, and whether and what type of systemic adjuvant therapy was administered. The use of medical literature is also problematic because series are small, and in the reports of clinical trials, there may be biases in the type of patients entered onto trials, particularly for patients with small, low-risk tumors. For node-negative patients (particularly T1N0M0, stage 1 patients), the estimates of baseline outcome and the impact of adjuvant therapy are especially important to validate, because for these patients the estimates are most likely to influence their treatment decisions.
Estimates of breast cancer–specific mortality derived from the SEER data and used by Adjuvant! reasonably fit published estimates for small node-negative tumors. For patients with node-negative T1a and T1b tumors, Adjuvant! uses an estimate of 10-year breast cancer–specific mortality of 4%. Estimates for patients receiving no systemic therapy from a Finnish registry are 4% (n = 80) and from Memorial Sloan-Kettering Cancer Center are 7% (n = 171). For patients with T1c tumors, Adjuvant! uses an estimate of 10%, whereas Finnish27 and Memorial Sloan-Kettering28 estimates are 7% (n = 130) and 18% (n = 303), respectively. Estimates of breast cancer–specific relapse are much more problematic because of the widely varying definitions of this term. In most reports, this term can be inferred from disease-free survival estimates, but disease-free survival estimates include events other than breast cancer recurrence, usually counting as events deaths due to any cause and even any second primary cancers. An estimate of 15% for stage 1 tumors from a natural history database29 of recurrence risk at 10 years is reasonably close to that given by Adjuvant! (18%).
Another approach to this problem is to use information from adjuvant clinical trials that evaluated the impact of systemic adjuvant therapy in node-negative patients. For this approach, individual patient data were not used, but rather published tables of characteristics of patients participating in the trials were used to make estimates of the size of subsets of patients on the basis of age, tumor size, number of nodes, and ER status. The program Adjuvant! then was used to make estimates of outcome of OS and DFS in these subsets. A weighted average (using the size of the subsets) was then made and could be compared with the results of the trial.
The results of this type of analysis are shown in the Table 5. For example, in NSABP B-14, a trial examining the effectiveness of adjuvant tamoxifen in node-negative ER-positive patients (58% of which had stage 1 disease), the outcomes and degree of benefit for tamoxifen as projected by Adjuvant! corresponds within a few percent to that observed in the trial. The fit for other recent node-negative trials reported with 10 years’ follow-up is also reasonably close. Although this method of validation is not ideal, it does give some confidence that on average the projections approximate what occurs in the clinic.
View this table:
[in this window]
[in a new window]
|
Table 5. Benefit as Reported From Node Negative Trials With 10 Years’ Follow-up and as Indirectly Projected by Adjuvant!
|
|
A number of simplifying approximations were made to produce the estimates made by Adjuvant!. An ideal model (1) would have more complex time-dependent terms for the risk of relapse and mortality without therapy, (2) might use terms with a difference time-dependence for patients who received adjuvant therapy (reflecting some events being delayed rather than prevented), and (3) might use more complex time and tumor-dependent estimates for the effectiveness of adjuvant therapy. Some of the more complex terms (although interesting) currently are not available and may complicate the analysis, which is already arcane but is still understandable when given in simple terms. A criticism of Adjuvant! is that it does not provide estimates with 95% confidence intervals. With the large number of uncertainties (in the data entered, in the prognostic estimates, and estimates of efficacy), confidence intervals could not be calculated formally. Some perception of the sensitivity of the outcome estimates can be gained by the program operator by entering information across the range of possible estimates.
There are several aspects of the use of the program that can be instructive to the clinician. One of these is the PFIC. The use of this tool gives insight into the dubious way prognostic factor information is used today. The vague terms with which prognostic factor information is typically reported now (usually favorable or unfavorable), without prevalence or relative risk conferred information does not allow an intelligent use of this information. Prognostic tests that confer modest relative risks (for example, 1.3) can never cause more than 1.3-fold change the estimate of risk. Even more powerful tests may have limited use. For example, PFIC can be used to show that for the patient with a 10% baseline risk, a test that confers a relative risk of 2.0 and a prevalence of 50% will identify high- and low-risk subsets with a 6.7% and 13.3% risk, only modestly different than the baseline estimate. Indeed, the PFIC can be used to demonstrate that for a test with a 50% positive prevalence, no matter how great the relative risk conferred, the resulting estimate of risk of negative outcome in the high-risk subset cannot be more that twice the baseline risk of the group as a whole.
The inclusion of additional prognostic information may be important, but the program does not include a specific list of additional prognostic tests to be done (recommendations for this can be obtained from the Web connections within the program to the ASCO guidelines). Thus, there are often multiple methodologies for even a single factor that is used for refining prognostic estimates, with differing prevalence of positive results, conferring differing relative risks, and with differing reliabilities. Nonetheless, the inclusion of histologic grade or some measure of proliferative rate has wide acceptance30 and the help files provide information about how to use this information either directly or when using the PFIC.
There are insights into the strengths and weaknesses of the Overviews5,6 that can be gained by working closely with their results to construct models to project outcome. One of these is the particular way mortality is analyzed in the Overview. It is all-cause mortality. The analysis of PRR for all-cause mortality is important because it shows that the balance of effects of adjuvant therapy result in an improvement in overall survival. It is not ideal because when non–breast cancer mortality is included in the analysis, it can confound the results (reducing the apparent breast cancer–specific mortality PRR) particularly in the older patients where non–breast cancer mortality is relatively high, and in low-risk patients where the ratio of non–breast cancer– to breast cancer–related mortality is high. From the modeling standpoint, having the Overview analyze PRR for breast cancer–specific mortality would be useful. The Overviews’ use of PRR for total mortality (rather than breast cancer–specific mortality) is one of the reasons (along with a relative delay between relapse and death, and curable recurrences such as those that occur locally after lumpectomy and radiation) why PRR for mortality are smaller that those for recurrence in the Overviews. The discrepancy between the PRR for relapse and mortality is smallest for premenopausal ER-negative patients (Table 4), just the subset for whom confounding effects of non–breast cancer mortality would be the least, and for whom the impact of delayed deaths occurring after recurrence would be the smallest, given the early peak in annual risk of mortality that occurs in the first 5 years for these patients (Fig 2).
Ideally, patients are informed partners in the decisions about their therapy. Although information sources such as individual adjuvant therapy trials or the Overviews can provide estimates of benefit for average patients who participated in them, the tool Adjuvant! can provide estimates for individual patients. In this regard, the program has been tailored specifically to provide output that is in a format that is useful for the clinician and easily understood by the patient. Tools such as Adjuvant! by more effectively bringing patients into the decision-making process may have a significant effect on what decisions are made. In a randomized trial, we are studying the impact of this tool on patient preferences for different treatment options, on patient satisfaction with the process, and on the patient’s perceived acceptability of entry into clinical trials. Previous studies have suggested that breast cancer patients tend to overestimate their risk of negative outcome without therapy and to overestimate the impact of adjuvant therapy.3 Thus, the information provided by Adjuvant! can be both reassuring and disappointing to patients, but it may increase interest in the search for even more effective adjuvant therapies.
|
ACKNOWLEDGMENTS
|
|---|
Supported in part by grant no. RO1-HSO8516 from the National Cancer Institute.
|
REFERENCES
|
|---|
.
Costantino JP, Gail MH, Pee D, et al: Validation studies for models projecting the risk of invasive and total breast cancer incidence. J Natl Cancer Inst 91: 1541-1548, 1999[Abstract/Free Full Text]
.
Siminoff LA, Fetting JH, Abeloff MD: Doctor-patient communication about breast cancer adjuvant therapy. J Clin Oncol 7: 1192-1200, 1989[Abstract]
.
Ravdin PM, Siminoff IA, Harvey JA: Survey of breast cancer patients concerning their knowledge and expectations of adjuvant therapy. J Clin Oncol 16: 515-521, 1998[Abstract]
.
Loprinzi CL, Ravdin PM, de Laurentiis M, et al: Do American oncologists know how to use prognostic variables for patients with newly diagnosed primary breast cancer? J Clin Oncol 12: 1422-1426, 1994[Abstract]
.
Early Breast Cancer Trialists’ Collaborative Group: Polychemotherapy for early breast cancer: An overview of the randomized trials. Lancet 352: 930-942, 1998[Medline]
.
Early Breast Cancer Trialists’ Collaborative Group: Tamoxifen for early breast cancer: An overview of randomised trials. Lancet 351: 1451-1467, 1998[Medline]
.
Ingelfinger JA, Mosteller F, Thibodeau LA, et al (eds): Biostatistics in Clinical Medicine. New York, NY, McGraw-Hill Inc, 1994, pp 244-258
.
Saphner T, Tormey DC, Gray R: Annual hazard rates of recurrence for breast cancer after primary therapy. J Clin Oncol 14: 2738-2746, 1996[Abstract/Free Full Text]
.
Hilsenbeck SG, Ravdin PM, de Moor CA, et al: Time-dependence of hazard ratios for prognostic factors in primary breast cancer. Breast Cancer Res Treat 52: 227-237, 1998[Medline]
.
Arriagada R, Rutqvist LE, Skoog L, et al: Prognostic factors and natural history in lymph node-negative breast cancer patients. Breast Cancer Res Treat 21: 101-109, 1992[Medline]
.
Silvestrini R, Daidone MG, Luisi A, et al: Biologic and clinicopathologic factors as indicators of specific relapse types in node-negative breast cancer. J Clin Oncol 13: 697-704, 1995[Abstract/Free Full Text]
.
Centers for Disease Control and Prevention: Vital Statistics of the United States, Volume II. Atlanta, GA, National Center for Health Statistics, 1989, Section 6, Table 6-1, p 7
.
Medical Risks: Trends in Mortality by Age and Time Elapsed. Praeger, NY, Association of Life Insurance Medical Directors of America and the Society of Actuaries, 1990
.
Reports of Mortality and Morbidity Experience. Transactions of the Society of Actuaries. Schaumburg, IL, Society of Actuaries, 1982
.
Krag D, Weaver D, Ashikaga T, et al: The sentinel node in breast cancer: A multicenter validation study. N Engl J Med 339: 941-946, 1998[Abstract/Free Full Text]
.
de Mascarel I, Bonichon F, Coindre JM, et al: Prognostic significance of breast cancer axillary lymph node micrometastases assessed by two special techniques: Reevaluation with longer follow-up. Br J Cancer 66: 523-527, 1992[Medline]
.
Clayton F, Hopkins CL: Pathologic correlates of prognosis in lymph node-positive breast carcinomas. Cancer 71: 1780-1790, 1993[Medline]
.
Elson CE, Kufe D, Johnston WW: Immunohistochemical detection and significance of axillary lymph node micrometastases in breast carcinoma: A study of 97 cases. Anal Quant Cytol Histol 15: 171-178, 1993[Medline]
.
Byrne J, Horgan PG, England S, et al: A preliminary report on the usefulness of monoclonal antibodies to CA 15-3 and MCA in the detection of micrometastases in axillary lymph nodes draining primary breast carcinoma. Eur J Cancer 28: 658-660, 1992
.
International (Ludwig) Breast Cancer Study Group: Prognostic importance of occult axillary lymph node micrometastases from breast cancers. Lancet 335: 1565-1568, 1990[Medline]
.
Noguchi S, Aihara T, Motomura K: Detection of breast cancer micrometastases in axillary lymph nodes by means of reverse transcriptase-polymerase chain reaction: Comparison between MUC1 mRNA and keratin 19 mRNA amplification. Am J Pathol 148: 649-656, 1996[Abstract]
.
Carter CL, Allen C, Henson DE: Relation of tumor size, lymph node status, and survival in 24,740 breast cancer cases. Cancer 63: 181-187, 1989[Medline]
.
Ravdin PM: A computer-based program to assist in adjuvant therapy decisions for individual breast cancer patients. Bull Cancer (Paris) 82: 561s-564s, 1995 (suppl 5)
.
Ravdin PM: A computer program to assist in making breast cancer adjuvant therapy decisions. Semin Oncol 23: 43-50, 1996 (suppl 2)[Medline]
.
Ravdin PM: How can prognostic and predictive factors in breast cancer be used in a practical way today? Recent Results Cancer Res 152: 86-93, 1998[Medline]
.
Coates AS, Simes RJ: Patients assessment of adjuvant treatment in operable breast cancer, in Williams CJ (ed): Introducing New Treatments for Cancer: Practical, Ethical, and Legal Problems. New York, NY, J Wiley, 1992, pp 448-458
.
Joensuu H, Pylkkanen L, Toikkanen S: Late mortality from pT1N0M0 breast carcinoma. Cancer 85: 2183-2189, 1999[Medline]
.
Rosen PP, Groshen S, Saigo PE, et al: Pathological prognostic factors in stage I (T1N0M0) and stage II (T1N1M0) breast carcinoma: A study of 644 patients with median follow-up of 18 years. J Clin Oncol 7: 1239-1251, 1989[Abstract]
.
Quiet CA, Ferguson DJ, Weichselbaum RR, et al: Natural history of node-negative breast cancer: A study of 826 patients with long-term follow-up. J Clin Oncol 13: 1144-1151, 1995[Abstract]
.
American Society of Clinical Oncology: Update of recommendations for the use of tumor markers in breast and colorectal cancer. J Clin Oncol 16: 793-795, 1998[Abstract]
.
Mansour EG, Gray R, Shatila AH, et al: Survival advantage of adjuvant chemotherapy in high risk node-negative breast cancer: Ten year analysis—An Intergroup Study. J Clin Oncol 16: 3486-3492, 1998[Abstract]
Submitted May 18, 2000;
accepted September 22, 2000.
 CiteULike  Complore  Connotea  Del.icio.us  Digg  Facebook  Reddit  Technorati  Twitter What's this?
This article has been cited by other articles:
|
|
|
|
 
L. Esserman, Y. Shieh, and I. Thompson
Rethinking Screening for Breast Cancer and Prostate Cancer
JAMA,
October 21, 2009;
302(15):
1685 - 1692.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Mook, M. K. Schmidt, B. Weigelt, B. Kreike, I. Eekhout, M. J. van de Vijver, A. M. Glas, A. Floore, E. J. T. Rutgers, and L. J. van 't Veer
The 70-gene prognosis signature predicts early metastasis in breast cancer patients between 55 and 70 years of age
Ann. Onc.,
October 13, 2009;
(2009)
mdp388v1.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. M. Bueno-de-Mesquita, D. S. A. Nuyten, J. Wesseling, H. van Tinteren, S. C. Linn, and M. J. van de Vijver
The impact of inter-observer variation in pathological assessment of node-negative breast cancer on clinical risk assessment and patient selection for adjuvant systemic treatment
Ann. Onc.,
July 21, 2009;
(2009)
mdp273v1.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. J. Vickers, E. B. Elkin, P. B. Peele, M. Dickler, and L. A. Siminoff
Long-Term Health Outcomes of a Decision Aid: Data from a Randomized Trial of Adjuvant! in Women with Localized Breast Cancer
Med Decis Making,
July 1, 2009;
29(4):
461 - 467.
[Abstract]
[PDF]
|
|
|
|
|
|
|
M. C. U. Cheang, S. K. Chia, D. Voduc, D. Gao, S. Leung, J. Snider, M. Watson, S. Davies, P. S. Bernard, J. S. Parker, et al.
Ki67 Index, HER2 Status, and Prognosis of Patients With Luminal B Breast Cancer
J Natl Cancer Inst,
May 20, 2009;
101(10):
736 - 750.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Hickey, M. Peate, C.M. Saunders, and M. Friedlander
Breast cancer in young women and its impact on reproductive function
Hum. Reprod. Update,
May 1, 2009;
15(3):
323 - 339.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Schmidt, A. Victor, D. Bratzel, D. Boehm, C. Cotarelo, A. Lebrecht, W. Siggelkow, J. G. Hengstler, A. Elsasser, M. Gehrmann, et al.
Long-term outcome prediction by clinicopathological risk classification algorithms in node-negative breast cancer--comparison between Adjuvant!, St Gallen, and a novel risk algorithm used in the prospective randomized Node-Negative-Breast Cancer-3 (NNBC-3) trial
Ann. Onc.,
February 1, 2009;
20(2):
258 - 264.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
E. M. Ozanne, D. Braithwaite, K. Sepucha, D. Moore, L. Esserman, and J. Belkora
Sensitivity to Input Variability of the Adjuvant! Online Breast Cancer Prognostic Model
J. Clin. Oncol.,
January 10, 2009;
27(2):
214 - 219.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. G. Morris, H. L. McArthur, C. T. Dang, and C. A. Hudis
Selection of Appropriate Patients for Anthracycline Therapy
ASCO Educational Book,
January 1, 2009;
2009(1):
11 - 18.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. V. Pandharipande, M. G. Harisinghani, E. M. Ozanne, M. C. Specht, C. Hur, J. M. Lee, and G. S. Gazelle
Staging MR Lymphangiography of the Axilla for Early Breast Cancer: Cost-Effectiveness Analysis
Am. J. Roentgenol.,
November 1, 2008;
191(5):
1308 - 1319.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. R. Webster, S.-F. Lee, C. Ringland, A. L. Morey, A. M. Hanby, G. Morgan, K. Byth, P. A. Mote, P. J. Provan, I. O. Ellis, et al.
Poor-Prognosis Estrogen Receptor-Positive Breast Cancer Identified by Histopathologic Subclassification
Clin. Cancer Res.,
October 15, 2008;
14(20):
6625 - 6633.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Paik, G. Tang, and D. Fumagalli
An Ideal Prognostic Test for Estrogen Receptor-Positive Breast Cancer?
J. Clin. Oncol.,
September 1, 2008;
26(25):
4058 - 4059.
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. J. Wang, C. D. Fuller, J.-S. Kim, D. F. Sittig, C. R. Thomas Jr, and P. M. Ravdin
Prediction Model for Estimating the Survival Benefit of Adjuvant Radiotherapy for Gallbladder Cancer
J. Clin. Oncol.,
May 1, 2008;
26(13):
2112 - 2117.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. S. Ross, C. Hatzis, W. F. Symmans, L. Pusztai, and G. N. Hortobagyi
Commercialized Multigene Predictors of Clinical Outcome for Breast Cancer
Oncologist,
May 1, 2008;
13(5):
477 - 493.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. R. Acharya, D. S. Hsu, C. K. Anders, A. Anguiano, K. H. Salter, K. S. Walters, R. C. Redman, S. A. Tuchman, C. A. Moylan, S. Mukherjee, et al.
Gene Expression Signatures, Clinicopathological Features, and Individualized Therapy in Breast Cancer
JAMA,
April 2, 2008;
299(13):
1574 - 1587.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. Ueda, H. Tsuda, H. Asakawa, T. Shigekawa, K. Fukatsu, N. Kondo, M. Yamamoto, Y. Hama, K. Tamura, J. Ishida, et al.
Clinicopathological and Prognostic Relevance of Uptake Level using 18F-fluorodeoxyglucose Positron Emission Tomography/Computed Tomography Fusion Imaging (18F-FDG PET/CT) in Primary Breast Cancer
Jpn. J. Clin. Oncol.,
April 1, 2008;
38(4):
250 - 258.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Marchionni, R. F. Wilson, A. C. Wolff, S. Marinopoulos, G. Parmigiani, E. B. Bass, and S. N. Goodman
Systematic Review: Gene Expression Profiling Assays in Early-Stage Breast Cancer
Ann Intern Med,
March 4, 2008;
148(5):
358 - 369.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. C.U. Cheang, D. Voduc, C. Bajdik, S. Leung, S. McKinney, S. K. Chia, C. M. Perou, and T. O. Nielsen
Basal-Like Breast Cancer Defined by Five Biomarkers Has Superior Prognostic Value than Triple-Negative Phenotype
Clin. Cancer Res.,
March 1, 2008;
14(5):
1368 - 1376.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J.-A. W. Chapman, D. Meng, L. Shepherd, W. Parulekar, J. N. Ingle, H. B. Muss, M. Palmer, C. Yu, and P. E. Goss
Competing Causes of Death From a Randomized Trial of Extended Adjuvant Endocrine Therapy for Breast Cancer
J Natl Cancer Inst,
February 20, 2008;
100(4):
252 - 260.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
F. Cardoso, L. Van't Veer, E. Rutgers, S. Loi, S. Mook, and M. J. Piccart-Gebhart
Clinical Application of the 70-Gene Profile: The MINDACT Trial
J. Clin. Oncol.,
February 10, 2008;
26(5):
729 - 735.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Maraqa, M. Cummings, M. B. Peter, A. M. Shaaban, K. Horgan, A. M. Hanby, and V. Speirs
Carcinoembryonic Antigen Cell Adhesion Molecule 6 Predicts Breast Cancer Recurrence following Adjuvant Tamoxifen
Clin. Cancer Res.,
January 15, 2008;
14(2):
405 - 411.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
K. M.W. Pisters, W. K. Evans, C. G. Azzoli, M. G. Kris, C. A. Smith, C. E. Desch, M. R. Somerfield, M. C. Brouwers, G. Darling, P. M. Ellis, et al.
Cancer Care Ontario and American Society of Clinical Oncology Adjuvant Chemotherapy and Adjuvant Radiation Therapy for Stages I-IIIA Resectable Non Small-Cell Lung Cancer Guideline
J. Clin. Oncol.,
December 1, 2007;
25(34):
5506 - 5518.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
E. O. Hanrahan, A. M. Gonzalez-Angulo, S. H. Giordano, R. Rouzier, K. R. Broglio, G. N. Hortobagyi, and V. Valero
Overall Survival and Cause-Specific Mortality of Patients With Stage T1a,bN0M0 Breast Carcinoma
J. Clin. Oncol.,
November 1, 2007;
25(31):
4952 - 4960.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. Desmedt, F. Piette, S. Loi, Y. Wang, F. Lallemand, B. Haibe-Kains, G. Viale, M. Delorenzi, Y. Zhang, M. S. d'Assignies, et al.
Strong Time Dependence of the 76-Gene Prognostic Signature for Node-Negative Breast Cancer Patients in the TRANSBIG Multicenter Independent Validation Series
Clin. Cancer Res.,
June 1, 2007;
13(11):
3207 - 3214.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
H. B. Muss, L. Biganzoli, D. J. Sargent, and M. Aapro
Adjuvant Therapy in the Elderly: Making the Right Decision
J. Clin. Oncol.,
May 10, 2007;
25(14):
1870 - 1875.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. Crivellari, M. Aapro, R. Leonard, G. von Minckwitz, E. Brain, A. Goldhirsch, A. Veronesi, and H. Muss
Breast Cancer in the Elderly
J. Clin. Oncol.,
May 10, 2007;
25(14):
1882 - 1890.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Buyse, S. Loi, L. van't Veer, G. Viale, M. Delorenzi, A. M. Glas, M. Saghatchian d'Assignies, J. Bergh, R. Lidereau, P. Ellis, et al.
Validation and clinical utility of a 70-gene prognostic signature for women with node-negative breast cancer.
J Natl Cancer Inst,
September 6, 2006;
98(17):
1183 - 1192.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. R. Lairson, Y.-C. Chang, J. L. Bettencourt, S. W. Vernon, and A. Greisinger
Estimating Development Cost for a Tailored Interactive Computer Program to Enhance Colorectal Cancer Screening Compliance
J. Am. Med. Inform. Assoc.,
September 1, 2006;
13(5):
476 - 484.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
B. T. Hennessy, S. Giordano, K. Broglio, Z. Duan, J. Trent, T. A. Buchholz, G. Babiera, G. N. Hortobagyi, and V. Valero
Biphasic metaplastic sarcomatoid carcinoma of the breast
Ann. Onc.,
April 1, 2006;
17(4):
605 - 613.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J S Reis-Filho, C Westbury, and J-Y Pierga
The impact of expression profiling on prognostic and predictive testing in breast cancer.
J. Clin. Pathol.,
March 1, 2006;
59(3):
225 - 231.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R. Rouzier, L. Pusztai, S. Delaloge, A. M. Gonzalez-Angulo, F. Andre, K. R. Hess, A. U. Buzdar, J.-R. Garbay, M. Spielmann, M.-C. Mathieu, et al.
Nomograms to Predict Pathologic Complete Response and Metastasis-Free Survival After Preoperative Chemotherapy for Breast Cancer
J. Clin. Oncol.,
November 20, 2005;
23(33):
8331 - 8339.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
N. E. Davidson, M. Morrow, D. B. Kopans, and F. C. Koerner
Case 35-2005 -- A 56-Year-Old Woman with Breast Cancer and Isolated Tumor Cells in a Sentinel Lymph Node
N. Engl. J. Med.,
November 17, 2005;
353(20):
2177 - 2185.
[Full Text]
[PDF]
|
|
|
|
|
|
|
B. T. Hennessy, S. Krishnamurthy, S. Giordano, T. A. Buchholz, S. W. Kau, Z. Duan, V. Valero, and G. N. Hortobagyi
Squamous Cell Carcinoma of the Breast
J. Clin. Oncol.,
November 1, 2005;
23(31):
7827 - 7835.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Cianfrocca and W. J. Gradishar
Controversies in the Therapy of Early Stage Breast Cancer
Oncologist,
November 1, 2005;
10(10):
766 - 779.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Balducci
Squaring the Circle: Adjuvant Chemotherapy for Older Women With Breast Cancer
J. Gerontol. A Biol. Sci. Med. Sci.,
September 1, 2005;
60(9):
1135 - 1136.
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. K. Krzyzanowska and I. F. Tannock
Should Screen-detected Breast Cancers Be Managed Differently?
J Natl Cancer Inst,
August 17, 2005;
97(16):
1170 - 1171.
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. B. Peele, L. A. Siminoff, Y. Xu, and P. M. Ravdin
Decreased Use of Adjuvant Breast Cancer Therapy in a Randomized Controlled Trial of a Decision Aid with Individualized Risk Information
Med Decis Making,
May 1, 2005;
25(3):
301 - 307.
[Abstract]
[PDF]
|
|
|
|
|
|
|
I. A. Olivotto, C. D. Bajdik, P. M. Ravdin, C. H. Speers, A. J. Coldman, B. D. Norris, G. J. Davis, S. K. Chia, and K. A. Gelmon
Population-Based Validation of the Prognostic Model ADJUVANT! for Early Breast Cancer
J. Clin. Oncol.,
April 20, 2005;
23(12):
2716 - 2725.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. J. Whelan and C. Loprinzi
Physician/Patient Decision Aids for Adjuvant Therapy
J. Clin. Oncol.,
March 10, 2005;
23(8):
1627 - 1630.
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. J. van't Veer, S. Paik, and D. F. Hayes
Gene Expression Profiling of Breast Cancer: A New Tumor Marker
J. Clin. Oncol.,
March 10, 2005;
23(8):
1631 - 1635.
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. Schott and D. F. Hayes
Adjuvant Chemotherapy for Elderly Women With Hormone Receptor-Positive Breast Cancer: An Old(er) Problem
J. Clin. Oncol.,
December 1, 2004;
22(23):
4660 - 4662.
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. D. Atkins
Re: Influence of the New AJCC Breast Cancer Staging System on Sentinel Lymph Node Positivity and False-Negative Rates
J Natl Cancer Inst,
November 3, 2004;
96(21):
1639 - 1639.
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. S. Donaldson
Taking Stock of Health-Related Quality-of-Life Measurement in Oncology Practice in the United States
J Natl Cancer Inst Monographs,
October 1, 2004;
2004(33):
155 - 167.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
R. M. O'Regan
Do Tumors Detected by Mammography Screening Have a Favorable Prognosis?
JAMA,
September 1, 2004;
292(9):
1062 - 1063.
[Full Text]
[PDF]
|
|
|
|
|
|
|
N. F. Col, G. Weber, A. Stiggelbout, J. Chuo, R. D'Agostino, and P. Corso
Short-term Menopausal Hormone Therapy for Symptom Relief: An Updated Decision Model
Arch Intern Med,
August 9, 2004;
164(15):
1634 - 1640.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
S. K. Chia, C. H. Speers, C. J. Bryce, M. M. Hayes, and I. A. Olivotto
Ten-Year Outcomes in a Population-Based Cohort of Node-Negative, Lymphatic, and Vascular Invasion-Negative Early Breast Cancers Without Adjuvant Systemic Therapies
J. Clin. Oncol.,
May 1, 2004;
22(9):
1630 - 1637.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. Goldhirsch, W. C. Wood, R. D. Gelber, A. S. Coates, B. Thurlimann, and H.-J. Senn
Meeting Highlights: Updated International Expert Consensus on the Primary Therapy of Early Breast Cancer
J. Clin. Oncol.,
September 1, 2003;
21(17):
3357 - 3365.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. E. Holmes and H. B. Muss
Diagnosis and Treatment of Breast Cancer in the Elderly
CA Cancer J Clin,
July 1, 2003;
53(4):
227 - 244.
[Abstract]
[Full Text]
|
|
|
|
|
|
|
T. Whelan, C. Sawka, M. Levine, A. Gafni, L. Reyno, A. Willan, J. Julian, S. Dent, H. Abu-Zahra, E. Chouinard, et al.
Helping Patients Make Informed Choices: A Randomized Trial of a Decision Aid for Adjuvant Chemotherapy in Lymph Node-Negative Breast Cancer
J Natl Cancer Inst,
April 16, 2003;
95(8):
581 - 587.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
G. Sonpavde
Communicating the Value of Adjuvant Chemotherapy
J. Clin. Oncol.,
March 1, 2003;
21(5):
948 - 949.
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Whelan
A Trial of Two Questions
J. Clin. Oncol.,
October 15, 2002;
20(20):
4135 - 4138.
[Full Text]
[PDF]
|
|
|
|
|
|
|
C. A. Bunnell and E. P. Winer
Lumping Versus Splitting: The Splitters Take This Round
J. Clin. Oncol.,
September 1, 2002;
20(17):
3576 - 3577.
[Full Text]
[PDF]
|
|
|
|
|
|
|
N. L. Keating, E. Guadagnoli, M. B. Landrum, C. Borbas, and J. C. Weeks
Treatment Decision Making in Early-Stage Breast Cancer: Should Surgeons Match Patients' Desired Level of Involvement?
J. Clin. Oncol.,
March 15, 2002;
20(6):
1473 - 1479.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
P. A. Ganz, K. A. Desmond, B. Leedham, J. H. Rowland, B. E. Meyerowitz, and T. R. Belin
Quality of Life in Long-Term, Disease-Free Survivors of Breast Cancer: a Follow-up Study
J Natl Cancer Inst,
January 2, 2002;
94(1):
39 - 49.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
M. Levine and T. Whelan
Decision-Making Process--Communicating Risk/Benefits: Is There an Ideal Technique?
J Natl Cancer Inst Monographs,
December 1, 2001;
2001(30):
143 - 145.
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. M. Silva, G. Dominguez, J. Silva, J. M. Garcia, A. Sanchez, O. Rodriguez, M. Provencio, P. Espana, and F. Bonilla
Detection of Epithelial Messenger RNA in the Plasma of Breast Cancer Patients Is Associated with Poor Prognosis Tumor Characteristics
Clin. Cancer Res.,
September 1, 2001;
7(9):
2821 - 2825.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. M. Sleeper, S. M. Sorscher, L. L. Dietrich, C. L. Loprinzi, and S. D. Thome
Assessing Adjuvant Breast Cancer Therapy Benefit
J. Clin. Oncol.,
June 15, 2001;
19(12):
3157 - 3159.
[Full Text]
[PDF]
|
|
|
|
TOP
ABSTRACT
INTRODUCTION
, ER negative; •, ER positive; 
, ER undefined.
