|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2000 American Society for Clinical Oncology Role of Isolated Locoregional Recurrence of Breast Cancer: Results of Four Prospective StudiesByFrom the Institute of Medical Biometry and Medical InformaticsUniversity of Freiburg, Freiburg, and Department of Gynaecology, University of Heidelberg, Heidelberg, Germany. Address reprint requests to Claudia Schmoor, PhD, Institute of Medical Biometry and Medical Informatics, University of Freiburg, Stefan-Meier-Str 26, D-79104 Freiburg, Germany; email cs{at}imbi uni-freiburg.de.
PURPOSE: We evaluated the effect on future prognosis of an isolated locoregional recurrence (ILRR) after the primary diagnosis of breast cancer. Using data from four prospective studies of the German Breast Cancer Study Group, we investigated factors influencing prognosis after ILRR and defined a simple classification of patients into groups with different prognoses. PATIENTS AND METHODS: From 1983 to 1989, 2,746 patients were recruited into four studies comparing different treatments in primary breast cancer. After a median follow-up time of 8 years, 337 patients developed an ILRR as the first event. The influence of ILRRs on disease progression was examined. The effects of different prognostic factors on progression-free survival (PFS) and overall survival after ILRR were analyzed after a median follow-up time of 4.5 years. RESULTS: ILRRs increased the risk with respect to distant recurrence and death. After ILRR, 185 events occurred with respect to the PFS end point, and 171 patients died. Primary nodal status, tumor grade, estrogen receptor status of the primary tumor, and length of the disease-free interval (DFI) until the time of the ILRR had a significant prognostic impact. CONCLUSION: Determinants of prognosis after the ILRR should be taken into account for designing future risk-adapted clinical studies for these patients. Risk strata can be defined by a simple classification scheme based on primary nodal status and DFI.
THE PROGNOSTIC significance of the first isolated locoregional recurrence (ILRR) of breast cancer has been the subject of several investigations and controversial discussions. Recurrences restricted to the ipsilateral breast after lumpectomy, for example, were sometimes considered a purely local phenomenon that, after successful removal, would not alter the prognosis of these patients. In recent years, however, it has often been concluded that such a recurrence is a powerful independent predictor of distant disease.1-4 Sometimes it is argued that it is an indication of an increased risk that is already present and not a cause for distant metastases.1 Open questions remain that have not been satisfactorily answered so far, such as the influence of an ILRR on future prognosis. Another aspect is the identification of prognostic factors for a classification of patients into groups with different prognoses after the occurrence of an ILRR. The role of systemic treatment of an ILRR also remains controversial. A few randomized studies with small patient numbers have been performed so far to investigate the effect of systemic therapies of the ILRR. In a trial by Borner et al,5 in which 167 patients were randomized to either a tamoxifen or control group after surgery and radiotherapy, a benefit of tamoxifen could be shown with respect to progression-free survival (PFS) but not overall survival (OS). Fentiman et al6 investigated the effect of interferon-alfa in a randomized trial with only 32 patients and could not demonstrate any effect. However, this trial was too small to detect even an effect of moderate size. Further randomized trials are needed to investigate the role of systemic therapies for the treatment of an ILRR. In order to identify patient groups who might benefit from chemotherapy or hormonal treatment, the identification of subgroups of patients with different prognoses is an important prerequisite for study planning. Consequently, prognostic factors that influence the time to progression or death after the diagnosis of an ILRR have to be identified. A major problem of studies addressing these kinds of questions is that, even in large series of patients who have been prospectively followed after the primary diagnosis of breast cancer, relatively few develop an ILRR. In addition, these patients have to be observed after their ILRRs over a sufficiently long follow-up period. We used the data of 2,746 patients who had been enrolled onto four prospective studies of the German Breast Cancer Study Group and for whom the standard prognostic factors had been recorded at the time of the primary diagnosis. The influence on disease progression of developing an ILRR was analyzed. For 337 patients who experienced an ILRR, the effect of different prognostic factors on PFS and OS was investigated.
In 1983, the German Breast Cancer Study Group started four studies for patients with primary, histologically proven, nonmetastatic breast cancer to compare different treatment regimens. Between 1983 and 1989, 2,746 patients from 118 clinical institutions were entered onto these studies. The design of the studies, the treatment regimens, and the distribution of patients to studies and treatments is given in Table 1. Details of the studies and first results have been published previously7-10 (Schmoor et al, manuscript submitted for publication).
In study 1, after breast-conserving surgery, radiotherapy (with 1 MV of cobalt-60 or 4 to 6 MV of linear accelerator photons) was started 2 to 6 weeks after tumorectomy. Five fractions of 2 Gy per week and a total dose of 50 Gy (reference dose) in 25 fractions (5 to 6 weeks) were applied to the whole breast. An electron boost of 10 to 12 Gy (reference dose) in fractions of 2 Gy was delivered to the primary tumor bed (311 patients). If electrons were not available, the total reference dose to the breast was changed to 60 Gy (422 patients). Chemotherapy in studies 2, 3, and 4 was administered according to the modified Bonnadonna cyclophosphamide, methotrexate, and fluorouracil (CMF) regimen, which consisted of cyclophosphamide 500 mg/m2, methotrexate 40 mg/m2, and fluorouracil 600 mg/m2 given intravenously on days 1 and 8 of a 4-week treatment period. The first CMF cycle started within 36 hours after mastectomy. Hormonal therapy in study 2 consisted of a daily dose of 30 mg of tamoxifen for 2 years. Radiotherapy in study 3 was administered between the second and third cycles of CMF. The target volume included the chest wall, the parasternal and supraclavicular nodes, and the axilla. The recommended beam energy was 4 to 6 MV of photons or telecobalt. Conventional fractionation (2 Gy, five times a week) was used. The chest wall was irradiated by tangential fields (the reference point in the chest wall was 2 cm in depth) up to 50 Gy. The parasternal and supraclavicular nodes and the axilla were included in an anterior field ("hockey stick" approach; the reference point was 3 cm in depth). The total dose was 44 Gy. In the parasternal region, half of the dose was to be administered with electrons, if available. The comparative studies 1, 2, and 3 were originally intended as randomized trials but were actually planned as comprehensive cohort studies7,11 that also included eligible patients who had refused randomization because of a preference for one of the treatment arms under study. The decision to include these patients was driven mainly by two reasons. First, when planning the studies in the beginning of the 1980s, we expected problems in getting patients’ consent to randomization, especially in study 1. Second, we wanted to analyze the external validity of the results obtained in the randomized trial by comparing randomized patients with eligible patients who were treated and followed according to the same protocol as used for the randomized patients but for whom treatment assignment was performed in a nonrandomized manner. In fact, study 1 had an extremely low randomization rate of approximately 6% and was analyzed as a prospective observational study.8 A total of 473 (66%) of the patients in study 2 and 199 (61%) of the patients in study 3 were randomized. This allowed analyses of the randomized part of the trials9 as well as an analysis of the external validity based on the comprehensive cohort design.7 Study 4 was an observational study with one defined treatment strategy intended for the analysis of prognostic factors.10 The studies were performed after approval by an ethical committee. Informed consent was obtained from each patient. As potential prognostic factors, the following characteristics were determined at the time of the primary diagnosis: patient age, menopausal status, number of positive axillary lymph nodes, tumor location, tumor size, histologic tumor type, tumor grade according to Bloom and Richardson,12 and estrogen and progesterone receptor status. To determine nodal status and the number of positive lymph nodes, an en bloc axillary dissection of at least six identifiable lymph nodes was performed. The histopathologic classification was re-examined and grading was performed in two histopathologic reference centers. Methods for determining the factors have been described previously.8,9 Patients were examined at regularly scheduled follow-up visits to ensure detection of any kind of recurrence at the earliest time possible. In study 1, examinations were performed every 3 months during the first 2 years after operation, every 6 months for the following 3 years, and annually thereafter. In studies 2, 3, and 4, patients were examined every 3 months during the first 2 years after surgery, every 4 months during the following 3 years, every 6 months in years 6 and 7, and from the eighth year onward at annual intervals. The trials are at different stages of follow-up. In study 1, patients were followed until August 1995, for a median follow-up time of approximately 7 years. In studies 2, 3, and 4, patients were followed until the middle of 1997, for a median follow-up time of approximately 9 years. Not all patients followed the observation schedules planned in the study protocols. At the end of the observation period, the documentation of follow-up visits was missing for more than 2 years in 759 patients (27.6%). Therefore, information on the actual survival status of patients with incomplete follow-up data was requested from the corresponding registration offices. This additional information was obtained for nearly all patients and increased the median follow-up time with respect to OS to approximately 8 years in study 1 and to approximately 10 years in studies 2, 3, and 4. This information was used only for the calculation of OS. To calculate time to recurrence, the most recent information available from the clinical center was used. A recurrence was categorized as local (in breast in case of primary breast-conserving treatment, operation scar, or chest wall), regional (in the axillary lymph nodes or the supraclavicular region), or distant (metastases in distant sites or a contralateral or secondary tumor). The first event of failure was classified as an ILRR (appearance of local or regional recurrence at least 4 weeks before the diagnosis of distant failure or death), as a distant recurrence (with or without a simultaneous locoregional recurrence), or as death without recurrence.
In the first part of the analysis, which included all 2,746 study patients, the occurrence of an ILRR and its impact on future prognosis was analyzed. Event-specific recurrence rates for the first event of failure after the primary diagnosis were estimated using the methodology of cumulative incidence rates.13,14 The prognostic effect of an ILRR on distant disease-free survival and OS after the primary diagnosis of breast cancer was analyzed. Distant disease-free survival time was defined as the time from diagnosis of the primary tumor to distant failure or the patient’s death of any cause. OS time was defined as the time from diagnosis of the primary tumor to patient’s death of any cause. A multiple regression analysis was performed using the Cox model.15 The following standard baseline factors were included as fixed factors for adjustment: age (
In the second part of the analysis, we focused on patients who had experienced an ILRR as the first event of failure. The influence of various factors on the prognosis after the ILRR was investigated. PFS time after ILRR was defined as the period from diagnosis of the ILRR to disease progression (either locoregional or distant) or the patient’s death of any cause. OS time after ILRR was defined as the time from diagnosis of the ILRR to the patient’s death of any cause. The effects of the following factors on PFS and OS were analyzed: age ( The decision about the treatment of a recurrence was left to the individual clinical centers. Of the 337 patients who had experienced an ILRR as the first event of failure, no information about treatment was available for 122 patients. For the remaining 215 patients, the treatment was rather heterogeneous. Surgical treatment was performed in 63%, radiotherapy in 50%, and systemic treatment (chemotherapy and/or hormonal treatment) in 41%. More than one of these treatment options were administered in 46% of the patients. The effect on prognosis of ILRR treatment could not be analyzed. For univariate analyses, PFS rates and OS rates were calculated according to Kaplan-Meier.16 The relative risks between different groups, defined by prognostic factors with corresponding 95% CIs, were determined by univariate Cox regression models.15 The P values were based on the log-rank test.17,18 A simultaneous assessment of the effects of different prognostic factors was performed within a multiple regression analysis using the Cox model.15 Factors that had a P value of less than .05 in the univariate analyses with respect to PFS or with respect to OS were included. In order to identify patients with a different risk of progression after the ILRR, two different classifications were used. For the first classification, a prognostic index was developed from the multiple Cox regression analysis. On the basis of this index, the patients were divided into three groups of approximately equal size with low, medium, or high risk of progression. A second classification was based on a simpler index that used only primary nodal status and DFI as the two most important prognostic factors and defined the risk as low if both factors showed the favorable outcome, as medium if one of them was poor, and as high if both were poor. In order to compare the degree of separation of patients into risk groups as achieved by the two classifications, we calculated a summary measure of separation (SEP).10 This measure estimates for each defined risk group the risk relative to the center of the classification, takes this relative risk if it is greater than 1 or its inverse if it is less than 1, and then calculates the geometric mean of these values weighted by the number of patients in each group. In case of no separation, SEP has a minimum value of one. Model checks of the proportionality of the hazard functions14 were performed, and no indication for violation of this assumption was detected. All data storage and analyses were performed using the Statistical Analysis System.19
Incidence of ILRR After Primary Diagnosis Figure 1 shows the different types of recurrences in the 2,746 patients who entered the four studies. A total of 465 patients developed a locoregional recurrence, but in 71 patients it was diagnosed at the same time as a distant failure and in 57 patients it appeared after a distant failure had already occurred. Therefore, 337 patients remained with an ILRR.
Figure 2 shows the total failure rate in time after the primary diagnosis of breast cancer according to the specific incidences of locoregional recurrence, distant recurrence, or death as the first event of failure. The probability of developing an ILRR after 10 years was estimated as 14.8% (95% CI, 13.3% to 16.4%). The probability of developing a distant recurrence as the first event after 10 years was estimated as 27.4% (95% CI, 25.5% to 29.4%). The probability of death without recurrence after 10 years was estimated as 3.7% (95% CI, 2.9% to 4.8%).
Figure 3 shows that the ILRR rate was different depending on primary lymph node involvement. In node-negative patients, it was estimated as 12.1% (95% CI, 10.4% to 14.2%) after 10 years; in node-positive patients, it was estimated as 19.1% (95% CI, 16.7% to 21.8%) after 10 years.
The number of ILRRs observed by study and treatment arm is listed in Table 1. An analysis of the effect of the primary treatment on the occurrence of an ILRR was not the aim of this article. The effect of radiation therapy on the reduction of the ILRR rate in studies 1 and 3 was shown previously8 (Schmoor et al, manuscript submitted for publication).
Effect of ILRR on Future Prognosis
In a further analysis, the effect of an ILRR on the risk of distant recurrence or death was modeled as log-linearly changing with the time of occurrence after the primary diagnosis. The increase in risk of distant failure caused by an ILRR was estimated to decrease yearly by a factor of 0.978 (95% CI, 0.89 to 1.07) but was not significant (P = .64). With respect to OS, the increased risk after ILRR decreased significantly with increasing time since ILRR. The increase in the risk of death caused by ILRR decreased yearly by an estimated factor of 0.835 (95% CI, 0.74 to 0.94, P = .004). The result for the OS time is shown in Fig 4 together with the results obtained by the categorized approach from Table 2.
When the influence of an ILRR on distant failure and death was analyzed by one time-dependent covariate but the different effects depending on the time period to ILRR were not allowed for, the relative risk with respect to distant disease-free survival was estimated as 3.34 (95% CI, 2.79 to 4.00) and with respect to OS as 3.05 (95% CI, 2.49 to 3.74). This result is similar in magnitude to the results presented by others.1-3
Determinants of Prognosis After ILRR
We investigated the influence of various prognostic factors on PFS and OS after ILRR. Table 3 shows the patient characteristics and the univariate effects of prognostic factors on PFS and OS.
The following factors had an effect on PFS and on OS that was significant at the 5% level when considered one at a time: number of positive lymph nodes, tumor size, tumor grade, estrogen and progesterone receptor status of the primary tumor, type of locoregional recurrence, and length of DFI. These factors were included in a Cox regression analysis that investigated the effects simultaneously. This analysis was based on 302 patients with complete data (90% of the patients), and the results are listed in Table 4.
With respect to PFS, the strongest factors were the number of positive lymph nodes (P = .011) and the length of the DFI (P = .059). With respect to OS, in addition to the number of positive lymph nodes and the length of the DFI, tumor grade and estrogen receptor status had independent effects and were significant at the 5% level.
Furthermore, we investigated whether radiotherapy of the primary tumor had an effect on prognosis after ILRR. After adjusting for the most important prognostic factors, ie, primary nodal status (negative v positive) and length of DFI (
Definition of Risk Groups
Primary nodal status and length of DFI had the strongest effects on PFS. For a simpler classification, patients were divided into three groups with different prognoses as follows: low risk, primary node-negative status and DFI more than 2 years; medium risk, primary node-negative status and DFI  2 years or primary node-positive status and DFI more than 2 years; and high risk, primary node-positive status and DFI 2 years. Figure 7 shows the PFS rates of these prognostic subgroups. With this classification, the relative risk of medium versus low risk was estimated as 1.53 (95% CI, 1.03 to 2.26) and of high versus low risk as 3.45 (95% CI, 2.32 to 5.12). The SEP was estimated as 1.47, ie, the SEP using this simple classification was slightly lower than that using the above classification based on all factors in Table 4.
Table 5 shows the PFS rates and OS rates 5 years after ILRR for both classification schemes. With respect to OS, the SEP for the classification using all prognostic factors in Table 4 was 1.85; the SEP for the classification using only nodal status and DFI was 1.58. The loss of separation with the simple classification was greater for OS than for PFS because the effects of tumor grade and hormone receptor status on OS were stronger.
More patients were available for the simpler classification than for the classification based on the Cox model in Table 4 because nodal status and DFI had been documented in all patients. When a comparison of the two classification schemes was restricted to the 302 patients with complete data, results were similar to those presented above (not shown in detail).
In this study, we investigated the role of an ILRR in 2,746 patients from four prospective trials of the German Breast Cancer Study Group. The incidence of an ILRR was higher in node-positive patients than in node-negative patients. Patient entry onto the different studies was dependent upon their primary nodal status (Table 1). Of the 1,698 node-negative patients, 1,036 had not received adjuvant treatment after primary local therapy, which consisted of tumor excision plus radiotherapy or mastectomy. The remaining 662 node-negative patients had been uniformly treated with only one cycle of CMF chemotherapy immediately after mastectomy. A total of 1,048 patients had a positive nodal status and had received adjuvant treatment after mastectomy, according to the standardized treatment arms investigated in the trials. Approximately two thirds of the node-positive patients were randomly assigned to the treatment arms. Thus, as an advantage of our study, patients had been treated and followed in a standardized manner, which was formulated as an important prerequisite for an adequate analysis of the effects of prognostic factors on the course of disease.20
To analyze the influence on disease progression of developing an ILRR, we defined different time-dependent covariates for different time periods ( After the occurrence of an ILRR, the number of positive lymph nodes at the primary diagnosis and length of the DFI emerged as the strongest factors influencing PFS and OS. With respect to OS, tumor grade and estrogen receptor status of the primary tumor had independent effects as well. In several studies in the literature, a prognostic effect of the length of DFI4,5,22-31 or primary nodal involvement4,22,24,25,27,30,32,33 was observed, but not all of them analyzed the effect of different prognostic factors simultaneously in a multivariate analysis. There are also studies in which these effects could not be demonstrated. Recht et al34 provide a detailed summary in their overview. In addition to type of locoregional recurrence and length of DFI, baseline factors that had been determined at the time of the primary diagnosis were analyzed for their influence on prognosis after the ILRR. Remarkably, some of these factors, such as number of positive lymph nodes and tumor grade, which had a strong effect on prognosis after the primary diagnosis, retained their effect on progression after ILRR as well. This was not true for the age factor. Age influenced prognosis after the primary diagnosis but had no effect on prognosis after ILRR. In some studies, the effect of age at the time of ILRR was analyzed. As expected, this parameter did not have a prognostic effect in our data (results not shown in detail) because of its strong correlation with age at diagnosis. The effect of the patient’s hormone receptor status was analyzed by investigating both the estrogen and progesterone receptors. In univariate analyses, both receptors had a significant influence on distant disease-free survival and OS from the time of the primary diagnosis as well as on PFS and OS from the time of the ILRR. If both factors are analyzed simultaneously in a multivariate model, one of the effects disappears because of the rather strong correlation between the two factors (in only 23% of the patients, the receptors showed discordant results by the chosen negative/positive classification). With respect to the effect on survival after primary diagnosis, progesterone receptor status still had an effect (Table 2), and with respect to the effect on survival after ILRR, estrogen receptor status still had an effect (Table 4). This does not mean that the two receptors have different roles in the prognosis of breast cancer after primary diagnosis and after ILRR. When only one of the two receptors is included in Table 2 or Table 4, this factor has a significant effect on prognosis independent of being the estrogen or the progesterone receptor. This should be interpreted to mean that hormone receptor may be represented either by estrogen or by progesterone. We examined prognosis after the ILRR without considering the effect of the type of primary surgery. In only 69 of the 337 patients with an ILRR, the primary tumor had been treated with breast-conserving therapy. These were node-negative patients with a tumor size of less than 2 cm who had been entered onto study 1 in order to evaluate the effect of primary surgery without systemic treatment (Table 1). To investigate the effect of the primary surgery on prognosis after the ILRR, we could have compared these 69 patients with the 26 patients who had been treated in the same study by mastectomy. However, this comparison would not have been conclusive because of the small number of events (27 events with respect to PFS and 23 events with respect to OS). To investigate whether local recurrence after breast-conserving treatment had a different effect on future prognosis compared with a local recurrence after mastectomy, the patient groups had to be further reduced by excluding patients with a regional failure. Fifty-five patients with a local recurrence after primary breast-conserving therapy could be compared with only 21 comparable patients with a local recurrence after primary mastectomy. So far, only 15 of these patients have experienced an event with respect to PFS, so an analysis is not possible at the moment. One drawback of our study is that the treatment of the ILRR was not recorded with sufficient consistency to be analyzed in this report. On the other hand, our study has the advantage that the effects of prognostic factors could be analyzed in a large patient population with ILRRs being treated and followed in a standardized manner from the time of the primary diagnosis. This is in contrast to other studies21,35,36 in which the prognoses of smaller patient groups are reported which were selected on the basis of the treatment administered for the ILRR. To our knowledge, the 337 patients in our study comprise one of the largest populations analyzed with respect to prognostic factors after diagnosis of an ILRR. To plan randomized studies comparing treatments after an ILRR, we derived a simple classification of patients into subgroups with different prognoses. This was based on the most important factors in our study, the primary nodal status and the length of the DFI. For a comparison, we defined a prognostic index based on all prognostic factors that had an effect in univariate analyses. This comparison showed that the separation of patients into different prognostic groups with the simple classification scheme was only slightly improved by incorporating more factors. The advantage of simplicity outweighs the loss in separation, and the simpler index can be used to plan randomized trials in patients with ILRRs. Recently, other prognostic classifications of patients with ILRRs have been proposed in the literature. Kamby and Sengelov33 used the patient’s primary nodal status and chose, from among several blood parameters, the patient’s lactate dehydrogenase level to define three groups with different prognoses. Willner et al30 defined a very small subgroup of patients with a very good prognosis using eight prognostic factors. In general, the formation of subgroups of patients with different prognoses should be based on the following two considerations, among others. First, the factors included should be able to separate patients with respect to prognosis. Second, the factors included must be available in future patients. This means that only those factors should be used that are routinely determined in practice. The factors that fulfilled both criteria in the present situation were primary nodal status the length of the DFI. These factors had the strongest effect in our patient population and should be readily available in every patient diagnosed with an ILRR.
Study coordinators: G. Bastert and H. Scheurlen, Heidelberg; H. Bojar, Düsseldorf; J. Dunst, Halle; K. Hübner, Frankfurt; M. Olschewski, W. Sauerbrei, C. Schmoor, and M. Schumacher, Freiburg; H. Rauschecker, Rosenheim; R. Sauer, Erlangen; and Schauer, Göttingen, Germany. Participating centers: Kreiskrankenhaus, Frauenklinik, Aalen; St Josefs-Krankenhaus, Abteilung Chirurgie, Adenau; Kreiskrankenhaus, Gynäkologische Abteilung, Albstadt; Kreiskrankenhaus, Gynäkologische Abteilung, Annweiler; Städtische Krankenanstalten, Frauenklinik, Aschaffenburg; Klinikum Bamberg, Frauenklinik, Bamberg; St Franziskus-Hospital, Frauenklinik, Bielefeld; St Josefs-Hospital, Chirurgische Abteilung, Bochum; Herzogin-Elisabeth-Heim, Braunschweig; Zentralkrankenhaus Links der Weser, Chirurgische Klinik, Zentralkrankenhaus, Gynäkologische Abteilung, and Zentralkrankenhaus Bremen-Nord, Frauenklinik, Bremen; Kreiskrankenhaus, Chirurgische Abteilung, Bretten; St Rochus-Hospital, Castrop-Rauxel; Kreiskrankenhaus, Gebh-Gynäkologische Abteilung, Donaueschingen; Knappschaftskrankenhaus, Gebh-Gynäkologische Abteilung, Dortmund; Franz-Hospital, Abteilung Gynäkologie, Dülmen; St Marienhospital, Gynäkologische Abteilung, Düren-Birkesdorf; Städtisches Krankenhaus Benrath, Frauenklinik, Düsseldorf; Städtische Kliniken, Duisburg; Kreiskrankenhaus, Gebh-Gynäkologische Abteilung, Ebersberg; Kreiskrankenhaus, Chirurgische Abteilung, Eckernförde; St Antonius-Hospital, Gynäkologische Abteilung, Eschweiler; Marienhospital Essen-Altenessen, Frauenklinik, and St Josefs-Krankenhaus, Frauenklinik, Essen; Städtische Krankenanstalten, Frauenklinik, and Städtische Krankenanstalten, Strahleninstitut, Esslingen; Städtisches Marienkrankenhaus, Frauenklinik and Universitäts-Frauenklinik, Frankfurt; St Katharinen-Hospital, Gebh-Gynäkologische Abteilung, Frechen; Universitäts-Frauenklinik, Freiburg; Kreiskrankenhaus, Medizinische Abteilung, Freising; Städtisches Krankenhaus, Frauenklinik, Friedrichshafen; Zentrum für Chirurgie, Klinik für Allgemeinchirurgie, and Universitäts-Frauenklinik, Giessen; Evangelisches Krankenhaus, Abteilung Chirurgie, Göttingen-Weende; Chirurgische Universitätsklinik, Universitäts-Frauenklinik, and Krankenhaus Neu-Mariahilf, Abteilung Chirurgie, Göttingen; Kreiskrankenhaus, Frauenklinik, Groß-Umstadt; Evangelisches Amalie-Sieveking-Krankenhaus, Gynäkologische Abteilung, and Allgemeines Krankenhaus Harburg, Abteilung Gynäkologie, Hamburg; Vereins-Krankenhaus, Gynäkologische Abteilung, Hann-Münden; Kreiskrankenhaus St Marienberg, Abteilung Gynäkologie, Helmstedt; Kreiskrankenhaus, Frauenklinik, Herford; Universitäts-Frauenklinik, Homburg/Saar; Städtisches Krankenhaus, Abteilung Gynäkologie, Idar-Oberstein; Klinikum Ingolstadt, Ingolstadt; Chirurgische Universitätsklinik, Innsbruck; Städtisches Krankenhaus, Gynäkologische Abteilung, Kaiserslautern; Diakonissen-Krankenhaus, Frauenklinik, Städtisches Klinikum, and St Vincentius-Krankenhaus, Gebh-Gynäkologische Abteilung, Karlsruhe; Städtische Kliniken, Abteilung Chirurgie, Kassel; Universitäts-Frauenklinik and Chirurgische Universitätsklinik, Kiel; Krankenanstalten Konstanz, Abteilung Chirurgie, and Krankenanstalten Konstanz, Frauenklinik, Konstanz; Kreiskrankenhaus, Gebh-Gynäkologische Abteilung, Kronach; Kreiskrankenhaus, Lahr; Kreiskrankenhaus Landshut-Achdorf, Gebh-Gynäkologische Abteilung, Landshut; Borromäus-Hospital, Frauenklinik, Leer; Kreiskrankenhaus, Abteilung Chirurgie, Lich; Kreiskrankenhaus, Gynäkologische Abteilung, Lichtenfels; St Vincenz-Krankenhaus, Radiologische Klinik, Limburg; Medizinische Hochschule, Klinik für Chirurgie, Lübeck; St Hildegardis-Krankenhaus, Gynäkologische Abteilung, Mainz; St Vinzenz-Hospital, Mainz; Klinikum Mannheim, Chirurgische Klinik, and Klinikum Mannheim, Frauenklinik, Mannheim; Diakonie-Krankenhaus, Gebh-Gynäkologische Abteilung, Marburg; Kreis- und Stadtkrankenhaus, Abteilung Chirurgie, Marktredwitz; Elisabeth-Krankenhaus-Rheydt, Frauenklinik, Mönchengladbach; Evangelisches Krankenhaus, Frauenklinik, Mülheim/Ruhr; Frauenklinik der TU Rechts der Isar, Klinikum Großhadern, Frauenklinik, Klinikum Großhadern, Plastische Chirurgie, and Städtisches Krankenhaus München-Neuperlach, Chirurgische Abteilung, München; Johanna-Etienne-Krankenhaus, Abteilung Gynäkologie, Neuss; Krankenhaus Hetzelstift, Gebh.-Gynäkologische Abteilung, Neustadt an der Weinstraße; Kreiskrankenhaus, Gynäkologische Abteilung, Norden; Evangelisches Krankenhaus, Frauenklinik, Oberhausen; St Josef-Hospital Sterkrade, Gynäkologisch-Gebh. Abteilung, Oberhausen; Marienhospital, Osnabrück; Paracelsus-Krankenhaus Ruit, Gynäkologische Abteilung, Ostfildern; Städtisches Krankenhaus, Chirurgische Abteilung, Passau; Städtisches Krankenhaus, Chirurgische Abteilung, Pirmasens; Krankenhaus St Josef, Gynäkologische Abteilung, Regensburg; Klinik Dr Opitz, Regensburg; Kreiskrankenhaus, Chirurgische Abteilung, Rinteln; Kreiskrankenhaus, Abteilung Gynäkologie, Rotthalmünster; Stadtkrankenhaus, Gynäkologische Abteilung, Rüsselsheim; DRK-Krankenhaus, Gynäkologische Abteilung, Saarlouis; St Elisabeth-Klinik, Gynäkologische Abteilung, Saarlouis; Martin-Luther-Krankenhaus, Gynäkologisch-Gebh Abteilung, Schleswig; Kreiskrankenhaus, Chirurgische Abteilung, Schorndorf; Verbandskrankenhaus, Gynäkologische Abteilung, Schwelm; Krankenhaus Siegburg, Allgemeinchirurgische Klinik, and Krankenhaus Siegburg, Frauenklinik, Siegburg; Städtisches Krankenhaus, Allgemeine Chirurgie, Sindelfingen; St Lukas-Klinik, Gynäkologische Abteilung, Solingen; Diakonissenkrankenhaus Gynäkologische Abteilung, Speyer; Kreiskrankenhaus, Gebh-Gynäkologische Abteilung, St Ingbert; Elisabeth-Krankenhaus, Straubing; Krankenhaus Feuerbach, Chirurgische Klinik des Bürger-Hospitals, Stuttgart; Stadtkrankenhaus, Abteilung Chirurgie, Traunstein; Krankenhaus der Barmherzigen Brüder, and Mutterhaus der Borromäerinnen, Abteilung Strahlentherapie, Trier; St Johannes-Krankenhaus, Troisdorf-Sieglar; Universitäts-Frauenklinik, Ulm; Städtische Krankenanstalten, Gebh-Gynäkologische Klinik, Villingen-Schwenningen; Städtisches Krankenhaus, Chirurgische Klinik, and Städtisches Krankenhaus, Frauenklinik, Weiden i. d. Oberpfalz; Krankenhaus Wermelskirchen, Wermelskirchen; Kreiskrankenhaus St Elisabeth, Allgemeine Chirurgische Klinik, Wittlich; Stadtkrankenhaus, Gynäkologische Abteilung, Wolfsburg; Knappschaftskrankenhaus Bardenberg, Gebh-Gynäkologische Abteilung, Würselen; and Krankenhaus St-Josef, Wuppertal-Elberfeld, Germany.
Studies were supported by a grant from the Bundesministerium für Forschung und Technologie and long-term follow-up was supported by a grant from the Deutsche Forschungsgemeinschaft, Bonn, Germany.
1. Fisher B, Anderson S, Fisher ER, et al: Significance of ipsilateral breast tumor recurrence after lumpectomy. Lancet 338:327-331, 1991[Medline] 2. Whelan T, Clark R, Roberts R, et al: Ipsilateral breast tumor recurrence post lumpectomy is predictive of subsequent mortality: Results from a randomized trial. Int J Radiat Oncol Biol Phys 30:11-16, 1994[Medline]
3.
Veronesi U, Marubini E, Del Vecchio M, et al: Local recurrences and distant metastases after conservative breast cancer treatments: Partly independent events. J Natl Cancer Inst 87:19-27, 1995 4. Arriagada R, Rutqvist LE, Mattsson A, et al: Adequate locoregional treatment for early breast cancer may prevent secondary dissemination. J Clin Oncol 13:2869-2878, 1995[Abstract]
5.
Borner M, Bacchi M, Goldhirsch A, et al: First isolated locoregional recurrence following mastectomy for breast cancer: Results of a phase III multicenter study comparing systemic treatment with observation after excision and radiation. J Clin Oncol 12:2071-2077, 1994 6. Fentiman IS, Balkwill FR, Cuzick J, et al: A trial of human alpha interferon as an adjuvant agent in breast cancer after loco-regional recurrence. Eur J Surg Cancer 13:425-428, 1987 7. Schmoor C, Olschewski M, Schumacher M: Randomized and non-randomized patients in clinical trials: Experiences with comprehensive cohort studies. Stat Med 15:263-271, 1996[Medline] 8. Rauschecker HF, Sauerbrei W, Gatzemeier W, et al: Eight-year results of a prospective non-randomized study on therapy of small breast cancer. Eur J Cancer 34:315-328, 1998
9.
Schumacher M, Bastert G, Bojar H, et al: Randomized 2x2 trial evaluating hormonal treatment and the duration of chemotherapy in node-positive breast cancer patients. J Clin Oncol 12:2086-2093, 1994 10. Sauerbrei W, Hübner K, Schmoor C, et al: Validation of existing and development of new prognostic classification schemes in node-negative breast cancer. Breast Cancer Res Treat 42:149-163, 1997 (erratum published in Breast Cancer Res Treat 48:191-192, 1998)[Medline] 11. Olschewski M, Scheurlen H: Comprehensive cohort study: An alternative to randomized consent design in a breast preservation trial. Method Inform Med 24:131-134, 1985 12. Bloom HJG, Richardson WW: Histological grading and prognosis in breast cancer. Br J Cancer 2:359-377, 1957 13. Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. New York, NY,Wiley, 1980 14. Marubini E, Valsecchi MG: Analysing Survival Data From Clinical Trials and Observational Studies. Chichester, United Kingdom,Wiley, 1995 15. Cox DR: Regression models and life table (with discussion). J R Stat Soc B 34:187-220, 1972 16. Kaplan EL, Meier P: Nonparametric estimation from incomplete observation. J Am Stat Assoc 53:457-481, 1958 17. Peto R, Pike MC, Armitage P, et al: Design and analysis of clinical trials requiring prolonged observation of each patient: Part I. Br J Cancer 34: 585-612, 1976[Medline] 18. Peto R, Pike MC, Armitage P, et al: Design and analysis of clinical trials requiring prolonged observation of each patient: Part II. Br J Cancer 35:1-39, 1977[Medline] 19. SAS Institute Inc:SAS Language, Version 6. Cary, NC,SAS Institute, 1990 20. Simon R, Altman DG: Statistical aspects of prognostic factor studies in oncology. Br J Cancer 69:979-985, 1994[Medline] 21. Chaudary MA, Nagadowska M, Smith P, et al: Local recurrence after breast conservation treatment: Outcome following salvage mastectomy. The Breast 7:33-38, 1998 22. Beck TM, Hart NE, Woodard DA, et al: Local or regionally recurrent carcinoma of the breast: Results of therapy in 121 Patients. J Clin Oncol 1:400-405, 1983[Abstract] 23. Aberizk WJ, Silver B, Henderson IC, et al: The use of radiotherapy for treatment of isolated locoregional recurrence of breast carcinoma after mastectomy. Cancer 58:1214-1218, 1986[Medline] 24. Magno L, Bignardi M, Micheletti E, et al: Analysis of prognostic factors in patients with isolated chest wall recurrence of breast cancer. Cancer 60:240-244, 1987[Medline] 25. Brunner KW, Harder F, Greiner R, et al: Das loko-regionäre Rezidiv nach operiertem Mammakarzinom: Prognostische Faktoren und therapeutische Konsequenzen. Schweiz Med Wochenschr 118:1976-1981, 1988 [Medline] 26. Kurtz JM, Amalric R, Brandone H, et al: Local recurrence after breast-conserving surgery and radiotherapy: Frequency, time course, and prognosis. Cancer 63:1912-1917, 1989[Medline] 27. Schwaibold F, Fowble BL, Solin LJ, et al: The results of radiation therapy for isolated local regional recurrence after mastectomy. Int J Radiat Oncol Biol Phys 21:299-310, 1991[Medline] 28. Halverson KJ, Perez CA, Kuske RR, et al: Survival following locoregional recurrence of breast cancer: Univariate and multivariate analysis. Int J Radiat Oncol Biol Phys 23:285-291, 1992[Medline] 29. Haffty BG, Reiss M, Beinfield M, et al: Ipsilateral breast tumor recurrence as a predictor of distant disease: Implications for systemic therapy at the time of local relapse. J Clin Oncol 14:52-57, 1996[Abstract] 30. Willner J, Kiricuta IC, Kölbl O: Locoregional recurrence of breast cancer following mastectomy: Always a fatal event? Results of univariate and multivariate analysis. Int J Radiat Oncol Biol Phys 37:853-863, 1997[Medline] 31. Gage I, Schnitt SJ, Recht A, et al: Skin recurrences after breast-conserving therapy for early-stage breast cancer. J Clin Oncol 16:480-486, 1998[Abstract] 32. Janjan NA, McNeese MD, Buzdar AU, et al: Management of locoregional recurrent breast cancer. Cancer 58:1552-1556, 1986[Medline] 33. Kamby C, Sengelov L: Pattern of dissemination and survival following isolated locoregional recurrence of breast cancer: A prospective study with more than 10 years of follow up. Breast Cancer Res Treat 45:181-192, 1997[Medline] 34. Recht A, Hayes DF, Eberlein TJ, et al: Local-regional recurrence after mastectomy or breast conserving therapy, in Harris JR, Lippmann ME, Morrow M, et al (eds): Diseases of the Breast. Philadelphia, PA,Lippincott-Raven Publishers, 1996, pp 649-667 35. Abner AL, Recht A, Eberlein T, et al: Prognosis following salvage mastectomy for recurrence in the breast after conservative surgery and radiation therapy for early-stage breast cancer. J Clin Oncol 11:44-48, 1993[Abstract] 36. Dalberg K, Mattsson A, Sandelin K, et al: Outcome of treatment for ipsilateral tumor recurrence in early-stage breast cancer. Breast Cancer Res Treat 49:69-78, 1998[Medline] Submitted April 26, 1999; accepted December 20, 1999.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2000 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
20 or < 20 fmol/mg cytosol protein). Adjustment factors were specified in advance without looking at the data, and no variable selection was performed because a data-driven analysis was not desired. The model was stratified for nodal status and treatment arm in order to account for probable confounding by adjuvant treatment of the primary tumor. From this model, estimates of relative risks with corresponding 95% confidence intervals (95% CIs) were calculated. Factors that occurred on more than two levels were coded using dummy variables in order to estimate the relative risks between the different levels separately. P values were based on Wald tests.
