Doi:10.1016/s0140-6736(99)05036-9

Articles
Tamoxifen in treatment of intraductal breast cancer: NationalSurgical Adjuvant Breast and Bowel Project B-24 randomisedcontrolled trial Bernard Fisher, James Dignam, Norman Wolmark, D Lawrence Wickerham, Edwin R Fisher, Eleftherios Mamounas, Roy Smith, Mirsada Begovic, Nikolay V Dimitrov, Richard G Margolese, Carl G Kardinal, Maureen T Kavanah, Louis Fehrenbacher, Robert H Oishi IntroductionUntil the mid-1980s, mastectomy followed by axillary Background We have shown previously that lumpectomy with dissection was the preferred treatment for primary radiation therapy was more effective than lumpectomy alone invasive breast cancer and ductal carcinoma in situ for the treatment of ductal carcinoma in situ (DCIS). We did a (DCIS). That therapeutic approach was challenged, double-blind randomised controlled trial to find out whether however, as a result of more widespread use of better lumpectomy, radiation therapy, and tamoxifen was of more diagnostic equipment and an increased effort to educate benefit than lumpectomy and radiation therapy alone for DCIS.
women about the value of early detection of breast Methods 1804 women with DCIS, including those whose tumours. Invasive and non-invasive tumours were foundmore frequently and at earlier stages, often before they resected sample margins were involved with tumour, were became clinically evident. Before the availability of randomly assigned lumpectomy, radiation therapy (50 Gy), and mammography, fewer than 3% of newly diagnosed breast placebo (n=902), or lumpectomy, radiation therapy, and cancers were DCIS, and most presented as large tamoxifen (20 mg daily for 5 years, n=902). Median follow-up palpable masses, many with areas of microinvasion.1,2 was 74 months (range 57–93). We compared annual event Since the introduction of mammography, 20–30% of rates and cumulative probability of invasive or non-invasive mammographically detected cancers are DCIS.3 ipsilateral and contralateral tumours over 5 years.
The management of primary invasive breast cancer and Findings Women in the tamoxifen group had fewer breast- DCIS was influenced by a report which showed that the cancer events at 5 years than did those on placebo (8·2 vs outcome in women with invasive breast cancer treated by 13·4%, p=0·0009). The cumulative incidence of all invasive lumpectomy and radiation therapy was similar to that of breast-cancer events in the tamoxifen group was 4·1% at 5 women treated by radical or modified radical years: 2·1% in the ipsilateral breast, 1·8% in the contralateral mastectomy.4 Mastectomy for the treatment of invasivedisease therefore became more difficult to justify for breast, and 0·2% at regional or distant sites. The risk of invasive disease, but was frequently advocated for the ipsilateral-breast cancer was lower in the tamoxifen group management of DCIS, a non-invasive cancer. Thus, even when sample margins contained tumour and when DCIS surgery for removal of localised DCIS was commonly more radical than that for removal of localised invasive disease. Uncertainty therefore arose about the clinical therapy, and tamoxifen was effective in the prevention of management of women with small, localised DCIS detected by mammography, and prompted the NationalSurgical Adjuvant Breast and Bowel Project (NSABP) to do the B-17 study, a randomised controlled trial to investigate whether excision of localised DCIS withtumour-free sample margins (referred to as lumpectomy,although most women had no palpable mass) followed by National Surgical Adjuvant Breast and Bowel Project (NASBP), radiation therapy was more effective than lumpectomy Allegheny University of the Health Sciences 4 Allegheny Suite 602, Pittsburgh, PA 15212–5234, USA (Prof B Fisher MD, alone in prevention of an invasive tumour in the ipsilateral R Smith MD); Department of Biostatistics (J Dignam PhD) and NSABP Biostatistical Center (M Begovic MD), University of Pittsburgh, The first findings of B-17 reported a significantly better Pittsburgh; Allegheny General Hospital, Pittsburgh overall 5-year event-free survival because of lower D L Wickerham MD); Institute of Pathology, Shadyside Hospital, incidence of invasive and non-invasive ipsilateral-breast Pittsburgh (E R Fisher MD); Mount Sinai Center for Breast Health, cancers among women who underwent lumpectomy and Cleveland, OH (E Mamounas MD); Michigan State University, East radiation therapy.5 At 8 years, findings were confirmed of Lansing, MI (N V Dimitrov MD); Jewish General Hospital, Montreal, lower cumulative incidence of both types of ipsilateral- Quebec, Canada (R G Margolese MD); Alton Ochsner MedicalFoundation, New Orleans, LA (C G Kardinal breast cancer because of lumpectomy and radiation Center, Boston, MA (M T Kavanah MD); Kaiser Permanente, Northern therapy for localised mammographically detected DCIS.6 California Region, CA (L Fehrenbacher MD); and University of Hawaii, The study concluded that, because of the low recurrence rate of invasive ipsilateral-breast cancer, mastectomy was not warranted in women who had DCIS similar to that of Many women were ineligible for participation in B-17 because mammography showed diffuse DCIS and resected sample margins contained DCIS, or scattered calcifications were thought to be benign or associated with unremoved DCIS. Those women were mainly treated by mastectomy. Tamoxifen had been shown in animal studies to have anti-initiator and antipromoter properties.7,8 We have also reported that tamoxifen prevents tumour recurrences in the ipsilateral breasts and second primary tumours in the contralateral breasts of women who have undergone lumpectomy and radiation therapy for primary invasive breast cancer,9,10 which suggests that tamoxifen can interfere with development of primary invasive breast cancer from the start or with Table 1: Characteristics of patients and tumours progression of DCIS to invasive cancer. As a result, wedesigned the NSABP B-24 randomised controlled trial.
scattered microcalcifications that had been classified on The B-17 and B-24 studies were based on the idea that radiological assessment as suspicious were eligible if biopsy DCIS either progresses from non-invasive to invasive cancer or is a marker of risk rather than a progenitor forthe subsequent occurrence of an invasive tumour, or that a focus of invasive cancer existing in conjunction with After women had undergone lumpectomy and given written DCIS might remain after lumpectomy. In the B-24 study, consent, they were randomly assigned radiation therapy to the a double-blind randomised controlled trial, we tested the ipsilateral breast and placebo (n=902) or radiation therapy hypothesis that, in patients with non-invasive DCIS, followed by tamoxifen (n=902). To avoid an imbalance in treatment with lumpectomy, postoperative radiation characteristics according to treatment assignment, we stratifiedwomen by age (р49 years or >49 years), tumour type (DCIS or therapy, and tamoxifen would be more effective than DCIS plus LCIS), and method of detection (mammography, lumpectomy and radiation therapy alone in prevention of invasive and non-invasive cancers in the ipsilateral and Radiation therapy (50 Gy) was administered as stipulated in the contralateral breast. We present results from B-24 and protocol,4,5,12 and was started no later than 8 weeks after surgery.
relate them to those from B-17 and the NSABP P-1 Placebo or tamoxifen 10 mg twice daily were administered within 56 days of lumpectomy and were given continuously for 5 years.
No dose modifications were made for either agent. Patientsunderwent physical examinations every 6 mammography once a year. Tumours detected at local or regional sites were taken as events only if tissue biopsy of the lesion was Women with DCIS were eligible for inclusion if their life positive. Tumours detected at distant sites (ie, before local or expectancy was at least 10 years. Women with tumours that regional invasive cancer was noted) were taken as events if clinical, consisted of DCIS and lobular carcinoma in situ (LCIS) were radiographic, or pathological findings showed that a tumour was also eligible. Although we did not require or recommend axillary present. The presence of ipsilateral-breast or contralateral-breast dissection in B-24, if it was done, all lymph nodes had to be tumours, regional or distant metastases, second primary tumours negative for tumour on histological assessment. Time between other than a breast tumour that occurred as a first event, or death in surgery and randomisation had to be 56 days or less. Women the absence of evidence of recurrent breast cancer were used to who had previously been diagnosed with cancer, except for determine event-free survival. Our primary endpoints were the those who had had in-situ carcinoma of the cervix or squamous- occurrence of invasive or non-invasive tumours in the ipsilateral or cell or basal-cell carcinoma of the skin, were not eligible. We included women with one or more masses or clusters ofcalcifications that could be excised. Women whose DCIS or LCIS showed microscopic margin-involvement were eligible, as We calculated cause-specific hazards of failure and hazard rate were women with scattered calcifications that were thought to be ratios for the various endpoints, with exact binomial methods benign or of indeterminate nature (ie, those for whom follow-up used to test for differences in rates by treatment group. Cox’s assessment was recommended). Women with scattered proportional hazards model was used to calculate relative risks of calcifications that were classified as suspicious and who had no failure according to prognostic covariates and treatment signs of invasive cancer were also eligible. Women with no simultaneously, and to find out whether there was a differential definite clusters of calcifications or masses but who had diffuse response to therapy according to characteristics (eg, treatment- Breast cancer at regional or distant sites Second primary cancers other than endometrial cancer *Rate per 1000 patients per year.
†Rate in tamoxifen group divided by rate in placebo group.
‡Includes ipsilateral-breast cancer, contralateral-breast cancer, and local, regional, and distant disease.
§Includes ipsilateral and contralateral non-invasive tumours.
Table 2: Site, cumulative incidence, rate, and rate ratios of first events covariate interactions).13 We calculated cumulative probability of A comprehensive analysis based on additional follow-up data was events by means of cumulative incidence curves, which correctly done later. Our results reflect information reported to the account for competing risks.14 Event-free and total survival NSABP data centre as of Dec 31, 1998.
curves were calculated by Kaplan-Meier analysis. Pointwiseasymptotic 95% CI are presented for cumulative incidence andsurvival curves.
We aimed to achieve 85% power to detect a 50% lower 1804 women were randomly assigned treatment between occurrence of invasive cancer for women who received May 9, 1991, and April 13, 1994 (figure 1). 29 (1·6%) tamoxifen, with a one-sided 0·05 significance criterion. We patients (11 on placebo, 18 on tamoxifen) became anticipated that women who received tamoxifen would haveoutcomes at least as favourable as those of women who received ineligible after randomisation because primary tumours placebo, and, therefore, our original study design characterised showed characteristics other than those of non-invasive the tests as one-sided. However, all p values presented are two- intraductal carcinoma (13) or because surgery after sided. The design specifications required that a minimum of 72 diagnosis was delayed, surgical procedures were not done events (invasive cancer in the ipsilateral or contralateral breast or correctly, because of previous cancers, or because of other metastases to other sites) occurred among all patients before reasons (16). Of the 1804 randomised patients, 14 (0·8%) analysis. Three interim analyses did not result in stopping the did not begin the assigned therapy. 564 (31·3%) study early. The adjusted significance criterion for the definitive patients (269 in the placebo group and 295 in the analysis, based on a log-rank test for the comparison of time toinvasive breast cancer, was 0·0483, according to the method of tamoxifen group) who started therapy discontinued Fleming and colleagues.15 After the requisite number of events treatment because of side-effects (98 placebo, had been seen, we did a preliminary analysis for the investigators.
tamoxifen), personal reasons (146 placebo, Figure 2: Cumulative Incidence of all, invasive, and non-invasive events in ipsilateral and contralateral breastAll and invasive include cancers at regional and distant sites. p values based on comparisons of average annual rates of failure.
Figure 3: Cumulative Incidence of all and invasive events in ipsilateral and contralateral breastp values based on comparisons of average annual rates of failure.
124 tamoxifen), and unspecified reasons (25 placebo, 25 tamoxifen). We included all women in the analyses, There were 295 breast-cancer and non-breast-cancer however, for whom follow-up information was available, events among the 1798 patients with follow-up (table 2).
including those who did not meet eligibility criteria. An At 5 years of follow-up, 83·3% (95% CI 80·8–85·8) of analysis of only eligible patients produced similar results patients who received placebo were event-free compared and conclusions. Median follow-up was 74 months (range with 87·4% (85·1–89·6) of tamoxifen-treated patients Patients’ characteristics were similar in the two groups Among patients who received placebo, 130 invasive and (table 1). 65% of women reported being postmenopausal.
non-invasive breast-cancer events occurred in the 16% of women had positive resected sample margins after ipsilateral breast, contralateral breast, or presented as definitive surgery. More than 80% of tumours had metastases at regional or distant sites, compared with 84 maximum dimensions of 1 cm or less, were not palpable, in the tamoxifen group (table 2, figure 2). The estimated and were detected by mammography alone.
rate ratio for all breast cancer (0·63 [95% CI 0·47–0·83]) Covariate relative risk (95% CI) ‡
*Invasive and non-invasive.
†Rate per 1000 patients per year.
‡Relative risk for patients in given covariate stratum, relative to reference (first) stratum, adjusted for treatment.
§Includes unknown (79 and 94 in placebo and tamoxifen groups, respectively). These patients had failure rates similar to failure rate in women with positive margins.
Table 3: Relation between selected chcracteristics of patients and tumours and rates and relative risks of ipsilateral-breast tumours represented 37% fewer events in the tamoxifen group.
among women who received tamoxifen (data not shown).
There were 43% fewer invasive breast-cancer events and The relative risk of contralateral-breast cancer was 0·58 31% fewer non-invasive breast-cancer events in the (0·35–0·97), which represents 42% fewer contralateral-breast tumours among women who received tamoxifen.
150 (70%) of the 214 breast-cancer events were in the Ten patients had tumours other than in the breast. In ipsilateral breast (table 2, figure 3). A lower rate of the placebo group, tumours occurred in the liver, lung, ipsilateral-breast tumours in the tamoxifen group was thorax, ipsilateral axilla (three patients), and ipsilateral- apparent only for invasive tumours (44% reduction). The chest wall. In the tamoxifen group, tumours occurred in rate of non-invasive ipsilateral-breast tumours was not the contralateral and ipsilateral axilla (two patients, significantly lower in the tamoxifen group than in the table 2). These ten patients were eligible for the trial since placebo group (18% reduction, p=0·43). Of the 150 they had no invasive disease at the time of study entry. In patients with ipsilateral-breast tumours, 64% were treated nine women, the primary DCIS was 1·0 cm or less in size; by mastectomy; the remainder had a second lumpectomy.
five tumours had comedonecrosis and seven had negative The type of surgery after ipsilateral-breast tumour was similar in the two groups; 68% of women who The two groups did not differ in the rate of occurrence received placebo and 59% of those who received of second primary cancers other than those in the breast tamoxifen underwent mastectomy. Women who had or endometrium (table 2). 26 second cancers occurred in invasive ipsilateral-breast tumours were, however, treated women who received placebo and 25 in women treated by mastectomy more frequently than those who had a with tamoxifen. In each group, the tumours were widely non-invasive ipsilateral-breast tumour (75 vs 56%, distributed in various sites; there was no evidence of a difference in the incidence of tumours in the placebo or 36 contralateral-breast tumours (23 invasive and 13 tamoxifen groups at any site. There was a non-significant non-invasive) occurred as first events in the placebo group excess of endometrial cancer in patients who received and 18 (15 invasive and three non-invasive) occurred in tamoxifen: seven in the tamoxifen group compared with the tamoxifen group (table 2, figure 3). The estimated two in the placebo group (relative risk 3·4 [0·6–33·4]).
relative risk of contralateral-breast cancer was 0·48 The incidence of deaths that occurred before a breast- (0·26–0·87), which represents 52% fewer contralateral- cancer recurrence or before a second primary cancer was breast tumours for patients who received tamoxifen. The also similar in the two groups. 11 deaths were seen in the cumulative incidence of all contraleratal-breast tumours placebo group and ten in the tamoxifen group.
occurring at 5 years as first events was 3·4% in theplacebo group and 2·0% in the tamoxifen group. The Relation of characteristics to outcome reduction in invasive contralateral-breast tumours (23 vs Age at diagnosis was significantly associated with 15 events, 37%) was not significant (p=0·22). The occurrence of ipsilateral-breast tumour. Younger patients reduction in non-invasive contralateral-breast tumours in the two groups were at higher risk than older patients seemed larger, although the number of events was small for such an event (table 3). The annual rate of ipsilateral- (13 vs three, 78%, p=0·02). The cumulative incidence of breast tumour per 1000 women aged 49 years or younger invasive and non-invasive contralateral-breast tumours at who received placebo was 33·3 and 13·03 for those aged 5 years was low. The 5-year cumulative incidence of 50 years or older. Tamoxifen administration resulted in a invasive contralateral-breast tumours was 2·3% in the 38% reduction in ipsilateral-breast tumours in women placebo group compared with 1·8% in the tamoxifen younger than 50 years and a 22% reduction in women group; for non-invasive contralateral-breast tumours, it The presence of positive tumour-margins after surgery When contralateral-breast tumours that occurred as first was also associated with an increased rate and relative risk events or subsequent to other events were taken into account, of invasive or non-invasive ipsilateral-breast tumour.
39 were seen among women in the placebo group and 23 Similar findings occurred in patients whose DCIS were Figure 4: Cumulative Incidence of all, invasive, and non-invasive events in ipsilateral and contralateral breast in B-17 and B-24studies palpable compared with those whose disease was or initial tumour characteristics were associated with a diagnosed by mammography alone. The risk was lower for significantly increased risk of contralateral-breast cancer.
ipsilateral-breast cancer among women who receivedtamoxifen, irrespective of margin status (22% lower in women in the tamoxifen group whose sample margins 28 women in the placebo group and 26 in the tamoxifen were negative, and 44% in those with positive or unknown group died. At 5 years from study entry, survival was 97% margins). For a few women whose DCIS was clinically (96–98) for the two groups (p=0·74, data not shown). Six apparent at the time of study entry, failure rates were of the 28 women in the placebo group had invasive breast substantially higher in the two groups than for those cancer and two of these six had invasive ipsilateral-breast without clinically apparent DCIS (table 3).
tumours. Four women on tamoxifen had invasive breast Patients whose initial DCIS showed comedonecrosis, as cancer, and three of these four had invasive ipsilateral- reported by institutional pathologists, were about twice as breast tumours. One developed a new primary cancer.
likely to develop an ipsilateral-breast tumour as womenwhose DCIS showed no comedonecrosis. This observation was more strongly associated with the Information about toxic events was available for 1781 occurrence of a non-invasive than with an invasive (98·7%) randomised patients (table 4). No strokes were seen ipsilateral-breast tumour (data not shown). The rate of in the two groups. Grade 4 toxic effects not usually ipsilateral-breast tumour was, however, lower by a similar associated with tamoxifen occurred with similar rates in the degree in the tamoxifen group in women who had no two groups. There was an increase in the rate of endometrial comedonecrosis (23%) and in those who showed evidence cancer in tamoxifen-treated patients, (1·53 vs 0·45 per 1000 of comedonecrosis at entry (31%). None of the patients’ patients per year in the placebo group). No deaths fromendometrial cancer occurred in the tamoxifen group.
Women with DCIS treated by lumpectomy and radiation therapy showed additional benefit from tamoxifen. The advantage was due mainly to a decrease in the rate of invasive cancer, especially in the ipsilateral breast. That effect was also seen in the rate of invasive and non- invasive tumours in the contralateral breast and at regional or distant sites. When the events of those sites were combined, there was a significantly lower rate and cumulative incidence of all breast-cancer-related events than in the placebo group. These observations suggest that focusing on the frequency with which ipsilateral- breast tumours occur after lumpectomy for DCIS is too limited; the possible effect that treatment strategies for DCIS have on all invasive or non-invasive breast-cancer events at any site seems more important. Therefore, the few metastases that were detected at regional and distant sites in this study cannot be ignored, especially since Almost two-thirds of the patients in the two groups who developed non-invasive ipsilateral-breast tumours during *Excludes alopecia, irregular menses, hot flashes, fluid retention, vaginal discharge,nadir grades, and weight gain or loss; septic episode classified as grade 4.
the first 5 years of follow-up underwent mastectomy, an Table 4: Adverse events by treatment group accepted treatment, although a second breast-conserving procedure might have been possible. The large number of tumours that were oestrogen-receptor positive. Tamoxifen mastectomies performed lessened the opportunity for also lowered rates of invasive tumours in P-1 participants estimating the frequency of subsequent invasive cancers in with a history of atypical hyperplasia or LCIS, which are women with a history of DCIS. Studies to ascertain frequently oestrogen-receptor positive.19,20 Since DCIS is whether a second lumpectomy in conjunction with commonly associated with tumour cells that are tamoxifen is adequate therapy are appropriate.
oestrogen-receptor and progesterone-receptor positive,19–25 Our findings from this study can be considered in the tamoxifen could have led to lower rates of oestrogen- context of the B-17 findings because the studies were receptor-positive invasive cancers in in this study after similar except for inclusion in B-24 of women with more extensive DCIS. The cumulative incidence of all breast- In the P-1 study, the lower risk of invasive cancer in cancer events at 5 years in the placebo group of B-24 was women with a history of LCIS or atypical hyperplasia was similar to that in women who were treated by lumpectomy because of tamoxifen administration, whereas in and radiation therapy in B-17. The spectrum of results participants with DCIS in the B-17 study and this study from the two studies clearly depicts the advantage from the lower risk was because of postoperative breast radiation therapy, as well as the added benefit of irradiation and the difference was greater when tamoxifen tamoxifen (figure 4). In the B-17 study, the cumulative was administered with radiation therapy. Therefore, the incidence of all breast-cancer-related events in women effectiveness of tamoxifen in treating DCIS without with DCIS treated by lumpectomy alone was about 25% radiation therapy is speculative. A study currently in at 5 years. The cumulative incidence was 13% after radiation therapy in the two trials, and 8% when The nature and magnitude of tamoxifen’s side-effects tamoxifen was given in B-24. This benefit was due partly were similar to those seen in other NSABP trials of to the lower rate of contralateral-breast cancer and of tamoxifen. The increase in the rate of endometrial cancer invasive cancer at regional and distant sites in tamoxifen- after tamoxifen administration and the rate of deep-vein treated women. Thus, tamoxifen and radiation therapy thrombosis and pulmonary emobolism were minimal led to a 68% lower cumulative incidence of all breast- (<1%) and there were no strokes. Despite the small cancer events at 5 years of follow-up than in women increases in rate of certain adverse events, adverse effects treated with lumpectomy alone in B-17. Compared with from surgery, radiation therapy, and chemotherapy in women who underwent lumpectomy alone, the tamoxifen unprevented breast cancers would probably have group showed a 77% reduction in all invasive breast- exceeded those associated with tamoxifen. Nevertheless, cancer events and about 64% reduction for all non- before women receive the drug, they should be carefully screened so that only those who are free of comorbid In this study, as in B-17,16,17 positive sample margins disorders that would predispose them to such events are and comedonecrosis were associated with an increased risk of ipsilateral-breast tumours. Tamoxifen reduced the Our findings could contribute to the decision-making incidence of such tumours, irrespective of the presence or absence of comedonecrosis or of margin involvement. An mammographically detected DCIS when radiological or 8-year update of the B-17 findings has, however, shown pathological evidence suggests that all of the cancer was that, although comedonecrosis continues to be a predictor not removed after lumpectomy. Currently, mastectomy is of ipsilateral-breast tumour, margin status had less of an commonly deemed appropriate when scattered influence18 than in earlier reports.
calcifications are seen radiographically or more than one Our findings from this trial and those of the B-17 study focus of clustered calcifications persist after surgery, when are related to those of the NSABP P-1 trial,11 which radiologists are uncertain about whether findings are showed that tamoxifen administered to women at indicative of invasive cancer, when sample margins increased risk for breast cancer led to 50% fewer non- contain DCIS after lumpectomy, or when a tumour is at invasive tumours (DCIS and LCIS) and 49% fewer or close to a margin. Our findings for ipsilateral-breast invasive breast cancers than placebo. B-17 showed that tumours are encouraging because they suggest that lumpectomy-treated women with a history of DCIS were mastectomy could be avoided more frequently than at at greater risk for invasive breast cancer than were women present. Moreover, the value of tamoxifen used in in P-1 who had a history of LCIS or atypical hyperplasia.
combination with radiation therapy to lower the In B-17, the annual rate of an invasive-breast-cancer event occurrence of invasive cancer justifies the suggestion that during the first 5 years after the occurrence of DCIS was a combined therapy replaces mastectomy for the treatment little more than twice that which occurred after LCIS and of DCIS patients in whom radiological findings are nearly three times that after atypical hyperplasia. The unlikely to be related to an invasive tumour.
B-17 study and the present study showed that invasive- cancer rates in DCIS patients who received radiation Bernard Fisher, Norman Wolmark, and D Lawrence Wickerham were therapy alone were higher than those in patients with a involved in the design and implementation of the study. James Dignamwas responsible for statistical analysis. Edwin Fisher was the senior history of LCIS or atypical hyperplasia who had received pathologist. Eleftherios Mamounas was the protocol officer from 1991–97.
tamoxifen alone. Since women with a history of LCIS or Roy Smith was protocol officer and, with Mirsada Begovic, extensively atypical hyperplasia have been thought to be at sufficiently reviewed patients’ records. Nikolay Dimitrov, Richard Margolese,Carl Kordinal, Maureen Kavanah, Louis Fehrenbacher, and Robert Oishi high risk of invasive cancer to warrant receiving contributed to the implementation and conduct of the study at major tamoxifen, women with DCIS, who are at even higher recruitment sites. Bernard Fisher and James Dignam prepared the first risk, should also be considered as candidates for draft of the paper, to which all investigators contributed.
tamoxifen, even if they have been treated with radiation Clinical Centres that contributed ten or more patients (shows therapy. The P-1 findings provide a biological basis for principal investigator and program coordinator)Baptist Regional Cancer Institute, Jacksonville, FL, USA N Abramson, such a theory since they showed that tamoxifen lowered P Stokes; Baylor-Sammons Cancer Center, Dallas, TX, M D Grant, the rate of invasive tumours overall because of an effect on B Quast; Baystate Medical Center, Springfield, MA, W P Reed, T Barron; Boston Medical Center, Boston, MA, M T Kavanah, Fisher B, Bauer M, Margolese R, et al. Five-year results of a D McDonald; British Columbia Cancer Agency, Vancouver, Canada, randomized clinical trial comparing total mastectomy and segmental I A Olivotto, C Trevisan; Community Clinical Oncology Program, mastectomy with or without radiation in the treatment of breast Allegheny, Pittsburgh, PA, R P Pugh; Community Clinical Oncology cancer. N Engl J Med 1985; 312: 665–73.
Program, Alton Ochsner Medical Foundation, New Orleans, LA, Fisher B, Costantino J, Redmond C, et al. Lumpectomy compared C G Kardinal, M Bateman; Community Clinical Oncology Program, with lumpectomy and radiation therapy for the treatment of Columbia River Oncology Program, Portland, OR, K S Lanier, intraductal breast cancer. N Engl J Med 1993; 328: 1581–86.
L Birenbaum; Community Clinical Oncology Program, Evanston Fisher B, Dignam J, Wolmark N, et al. Lumpectomy and radiation Hospital/Kellogg Cancer Center, Evanston, IL, J D Khandekar, B Marks; therapy for the treatment of intraductal breast cancer: findings from Community Clinical Oncology Program, Grand Rapids, MI, S L Smiley; National Surgical Adjuvant Breast and Bowel Project B-17. J Clin Community Clinical Oncology Program, Marshfield Clinic, Marshfield, Oncol 1998; 16: 441–52.
WI, J L Hoehn, L Weigel; Community Clinical Oncology Progam, Mount Jordan VC. Effect of tamoxifen (ICI 46,474) on initiation and growth Sinai Medical Center, Miami Beach, FL, E Davila, J W Grimes; of DMBA-induced rat mammary carcinomata. Eur J Cancer 1976; 12:
Community Clinical Oncology Program, Northwest, Tacoma, WA, H I Pierce, K Hart; Community Clinical Oncology Program, South East Cancer Control Consort, Winston-Salem, NC, J N Atkins, Jordan VC, Allen KE. Evaluation of the antitumour activity of the R R Burgess; Community Clinical Oncology Program, St Louis, MO, nonsteroidal antioestrogen monohydroxytamoxifen in the DMBA- R J Baglan; CHA-Pavillon Saint-Sacrement, Quebec City, Quebec, induced rat mammary carcinoma model. Eur J Cancer 1980; 16:
Canada, J Robert, E P Marcoux; City of Hope Medical Center, Duarte, CA, L D Wagman, D Hooks; Colorado Cancer Research Program, Fisher B, Costantino J, Redmond C, et al. A randomized clinical trial Denver, CO, S Sedlacek, N J Morton; Cross Cancer Institute, Edmonton, evaluating tamoxifen in the treatment of patients with node-negative Alberta, Canada, A W Lees, C Danbrook; Fallon Clinic, Worcester, MA, breast cancer who have estrogen-receptor-positive tumors. N Engl J J P Wei; Glens Falls Hospital, Glens Falls, NY, R W Sponzo, Med 1989; 320: 479–84.
B A Sponzo; Hartford Hospital, Hartford, CT, P A DeFusco, J M Kulko; 10 Fisher B, Redmond C. New perspective on cancer of the contralateral Henry Ford Hospital, Detroit, MI, R A Chapman, T J Doyle; Jewish breast: a marker for assessing tamoxifen as a preventive agent. J Natl General Hospital, Montreal, Quebec, Canada, R G Margolese, D Poulin; Cancer Inst 1991; 83: 1278–80.
Kaiser Permanente, Northern California Region, CA, L Fehrenbacher, 11 Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for A Rodoni; Kaiser Permanente, Portland, OR (Cooperative Group prevention of breast cancer: report of the National Surgical Adjuvant Outreach Programme), M U Rarick; Kaiser Permanente, San Diego, CA, Breast and Bowel Project P-1 study. J Natl Cancer Inst 1998; 90:
J Polikoff, S Cory; Lahey Clinic Medical Center, Burlington, MA, K S Hughes; Lehigh Valley Hospital, Allentown, PA, H C Hoover Jr, 12 Fisher B, Wolmark N, Fisher ER, Deutsch M. Lumpectomy and A Geshan; London Regional Cancer Center, Ontario, Canada, axillary dissection for breast cancer: surgical, pathological, and F E Perera; Lutheran Healthcare Network, Mesa, AZ, M M Long; radiation considerations. World J Surg 1985; 9: 692–98.
M D Anderson Cancer Center, Houston, TX, F C Ames; Minority-Based 13 Prentice RL, Kalbfleisch JD, Peterson AV, Flournoy N, Farewell VT, Community Clinical Oncology Program, Virginia Commonwealth Breslow NE. The analysis of failure times in the presence of University, Richmond, VA, H D Bear, L Keener; Michigan State competing risks. Biometrics 1978; 34: 541–54.
University, East Lansing, MI, N V Dimitrov, C Robins; Montreal General 14 Gaynor JJ, Feuer EJ, Tan CC, et al. On the use of cause-specific Hospital, Montreal, Quebec, Canada, M P Thirlwell; Mount Sinai failure and conditional failure probabilities: examples from clinical Medical Center, Cleveland, OH, L Levy, S Reynolds; New England oncology data. J Am Stat Assoc 1993; 88: 400–09.
Medical Center Hospitals, Boston, MA, R A Graham, M J Scannell; OhioState University, Columbus, OH, W B Farrar; Ontario Cancer 15 Fleming TR, Harrington DP, O’Brien PC. Designs for group Foundation-Hamilton Clinic, Ontario, Canada, H Lukka; Puget Sound sequential tests. Controlled Clin Trials 1984; 5: 348–61.
Oncology Consortium, Seattle, WA, R B Clarfeld, N L Knudsen; 16 Fisher ER, Costantino J, Fisher B, Palekar AS, Redmond C, Radiological Associates of Sacramento, CA, M H Leibenhaut, C Brooks; Mamounas E. Pathologic findings from the National Surgical Rockford Clinic, Rockford, IL, W R Edwards, S M Richter; Royal Adjuvant Breast and Bowel Project (NSABP) Protocol B-17: Victoria Hospital, Montreal, Quebec, Canada, H R Shibata, intraductal carcinoma (duct carcinoma in situ). Cancer 1995; 115:
J Pepe-Aylward; Rush Presbyterian-St Luke’s Medical Center, Chicago, IL, J M Wolter, D B Madrid; St Vincent Hospital, Indianapolis, IN, 17 Fisher ER. Pathobiological considerations relating to the treatment of J A Cavins; Sutter West Cancer Research Group, Greenbrae, CA, intraductal carcinoma (ductal carcinoma in situ) of the breast. CA P D Eisenberg, L Friedman; Tom Baker Cancer Centre, Calgary, Alberta, Cancer J Clin 1997; 47: 52–64.
Canada, A H G Paterson, L Dobni; University of Cincinnati, OH, 18 Fisher ER, Dignam J, Tan Chiu E, et al. Pathologic findings from the B Aron, MA Schilds; University of Hawaii, Honolulu, HI, R H Oishi, National Surgical Adjuvant Breast Project (NSABP) eight-year update D Coleman; University of Iowa, Iowa City, IA (Cooperative Group of protocol B-17. Cancer 1999 (in press).
Outreach Programme), P Jochimsen, M Spaight; University of Kentucky, 19 Giri DD, Dundas SA, Nottingham JF, Underwood JC. Oestrogen Lexington, KY, E H Romond, M Ashki; University of Massachusetts, receptors in benign epithelial lesions and intraductal carcinomas of Worcester, MA, M E Costanza; University of Medicine/Dentistry, New the breast: an immunohistological study. Histopathology 1989; 15:
Brunswick, NJ, I L Wapnir, A G Owens; University of Michigan, Ann Arbor, MI, L H Baker, B Golden; University of North Carolina, Chapel 20 Barnes R, Masood S. Potential value of hormone receptor assay in Hill, NC (Cooperative Group Outreach Program), S A Bernard, carcinoma in situ of breast. Am J Clin Pathol 1990; 94: 533–37.
G C Huitt; University of Pittsburgh Cancer Institute, PA, V G Vogel, 21 Zafrani B, Leroyer A, Fourquet A, et al. Mammographically detected L Robertson; University of Texas, San Antonio, TX, A B Cruz Jr, ductal in situ carcinoma of the breast analyzed with a new I Presas; University of Vermont, Burlington, VT (Cooperative classification: a study of 127 cases: correlation with estrogen and progesterone receptors, P53 and c-erb B-2 proteins, and proliferative activity. Semin Diagn Pathol 1994; 11: 208–14.
We thank Tanya Spewock for editorial assistance, Mary Hof for 22 Allred DC, Berardo MD, Prosser J, O’Connell P. Biologic and genetic preparation of the paper, and Linda Gilarski for data management. The features of in-situ breast cancer. In: Silverstein M, ed. Ductal study was supported by Public Health Service grants U10-CA-12027, carcinoma in situ of the breast. Baltimore: Williams and Wilkins, U10-CA-69651, U10-CA-37377, and U10-CA-69974 from the National Cancer Institute, National Institutes of Health, and Department of Health 23 Pallis L, Wilking N, Cedermark B, Rutqvist LE, Skoog L. Receptors for estrogen and progesterone in breast carcinoma in situ. Anticancer
Res 1992; 12: 2113–15.
24 Berardo M, Hilsenbeck SG, Allred DC. Histological grading of noninvasive breast cancer and its relationship to biological features.
Rosner D, Bedwani RN, Vana J, Baker HW, Murphy GP.
Lab Invest 1996; 74: 15A (abstr 68).
Noninvasive breast carcinoma: results of a national survey by the 25 Leal CB, Schmitt FC, Bento MJ, Maia NC, Lopes CS. Ductal American College of Surgeons. Ann Surg 1980; 192:139–47.
carcinoma in situ of the breast: histologic categorization and its Lagios MD, Westdahl PR, Margolin FR, Rose MR. Duct carcinoma relationship to ploidy and immunohistochemical expression of in situ: relationship of extent of noninvasive disease to the frequency of hormone receptors, p53, and c-erb B-2 protein. Cancer 1995; 75:
occult invasion, multicentricity, lymph node metastases, and short- term treatment failures. Cancer 1982; 50:1309–14.
26 Recht A, Rutgers EJ Th, Fentiman IS, Kurtz JM, Mansel RE, Sloane Ernster VL, Barclay J, Kerlikowske K, et al. Incidence of and JP. The Fourth EORTC DCIS Consensus Meeting (Château treatment for ductal carcinoma in situ of the breast. JAMA 1996; 275:
Marquette, Heemskerk, The Netherlands, 23–24 January 1998) Conference Report. Eur J Cancer 1998; 34: 1664–69.

Source: http://missinglink.ucsf.edu/lm/EEBM/RCT.pdf