|
|
| |||||||||||||||||||||||||||||||||||||||||
Background and Methods Women with ductal carcinoma in situ have been treated both by lumpectomy and by lumpectomy followed by radiation therapy, but the benefit of combined therapy is uncertain. A group of 818 women with ductal carcinoma in situ were randomly assigned to undergo lumpectomy or lumpectomy followed by breast irradiation (50 Gy). Sufficient tissue was removed that the margins of the resected specimens were histologically tumor-free. The mean duration of follow-up was 43 months (range, 11 to 86). The principal end point of the study was event-free survival, as defined by the presence of no new ipsilateral or contralateral breast cancers, regional or distant metastases, or other cancers and by no deaths from causes other than cancer.
Results Five-year event-free survival was better in the women who received breast irradiation (84.4 percent, vs. 73.8 percent for the women treated by lumpectomy alone; P = 0.001). The improvement was due to a reduction in the occurrence of second ipsilateral breast cancers; the incidence of each of the other events was similar in the two groups. Of 391 women treated by lumpectomy alone, ipsilateral breast cancer developed in 64 (16.4 percent); it was noninvasive in 32 and invasive in the remaining 32. Of 399 women treated with lumpectomy and breast irradiation, ipsilateral breast cancer developed in 28 (7.0 percent) (noninvasive in 20 and invasive in 8). The five-year cumulative incidence of second cancers in the ipsilateral breast was reduced by irradiation from 10.4 percent to 7.5 percent for noninvasive cancers and from 10.5 percent to 2.9 percent for invasive cancers (P = 0.055 and P<0.001, respectively).
Conclusions Breast irradiation after lumpectomy is more appropriate than lumpectomy alone for women with localized ductal carcinoma in situ. .
Before the advent of mammography, approximately 3 to 5 percent of new cases of breast cancer involved ductal carcinomas in situ,2,3 most of which were large, palpable masses. Information about the incidence of recurrence after the removal of these tumors was based on studies of women with clinically detectable lesions3,4,5,6,7. As a consequence of mammography, ductal carcinoma in situ is diagnosed more frequently, and a much higher proportion of the tumors are not clinically detectable8.
Because of uncertainty about the most appropriate treatment for women with these tumors, the National Surgical Adjuvant Breast and Bowel Project began a randomized clinical trial in 1985 to test the hypothesis that in women with localized ductal carcinoma in situ that was thought to have been completely removed, lumpectomy (or more accurately, local excision, since most women did not have a palpable mass) plus breast irradiation is more effective than local excision alone in preventing a second cancer in the ipsilateral breast. The initial results from this trial are presented here.
Methods
Selection of Patients
Women with localized ductal carcinoma in situ detected by physical examination or mammography were eligible for the study. Those with a tumor consisting of both ductal carcinoma in situ and lobular carcinoma in situ were also eligible, as were women with multiple lesions in the same breast if one lesion contained ductal carcinoma in situ and the others were histologically benign. Women with tumor-positive axillary nodes on clinical examination were excluded. If axillary dissection was carried out, all the nodes had to be tumor-negative. Also, women with previous cancers were ineligible, except those who had had in situ carcinoma of the cervix or squamous- or basal-cell cancer of the skin.
Within eight weeks after initial biopsy, the women underwent a lumpectomy, with removal of the tumor and a sufficient amount of normal breast tissue that the margins of the resected specimen were histologically tumor-free. Women with a histologic diagnosis of ductal carcinoma in situ whose mammograms showed diffuse microcalcifications were eligible only if histologic examination of tissue containing the microcalcifications demonstrated no tumor. The tumor diagnoses were made by institutional pathologists.
Study Design
Between October 1, 1985, and December 31, 1990, after they had undergone lumpectomy and given informed consent, the women enrolled in the study were randomly assigned to receive either ipsilateral breast irradiation or no radiation therapy. Randomization was carried out at the study biostatistical center with a stratified scheme designed to avoid inequalities in treatment assignment9. The stratification variables were age ( 
49 or >49 years), axillary dissection (performed or not performed), tumor type (ductal carcinoma in situ or ductal carcinoma in situ plus lobular carcinoma in situ), and method of detection (mammography, clinical examination, or both). Initially, all women underwent axillary-node dissection. After June 1987, such dissection became optional because of reports indicating the infrequency of positive axillary nodes in women with this type of tumor10.
Radiation Therapy
Breast irradiation was begun when wound healing permitted, but no later than eight weeks after the operation. The patients received a dose of 50 Gy, calculated at a depth equal to two thirds of the distance between the skin overlying the breast and the base of the tangential fields at midseparation. The dose was given at a rate of 10 Gy per week (2 Gy per day for five days). Nine percent of the women received additional radiation therapy to the tumor site. Interstitial radiation and regional node irradiation were not allowed. The radiation-therapy technique was similar to that used in previous studies11,12.
Follow-up and Identification of Study End Points
Women underwent follow-up examinations every six months; mammography was performed annually. A tumor detected at a local or regional site after the initial operation was considered an event when a biopsy of the lesion was positive. A tumor at a distant site was considered an event when the clinical, radiographic, or pathological findings indicated that tumor was present. The presence of ipsilateral or contralateral breast cancer, regional or distant metastasis, or a second primary tumor other than a breast tumor occurring as a first event after the operation or the patient's death in the absence of evidence of recurrent breast cancer was used to determine event-free survival.
The location of primary tumors and postoperative ipsilateral breast tumors was determined from mammographic, operative, and pathological reports submitted to the biostatistical center. Tumors were classified as being in a specific quadrant, on the border between two quadrants, central (in a subareolar position or at the border of a quadrant and the areola), or diffuse (in more than one of the preceding sites).
Statistical Analysis
The results in the two treatment groups were compared for all randomized patients for whom information was available that could be evaluated. The percentage of women in whom no event occurred was determined with a life-table estimate computed by the actuarial method13. The two treatment groups were compared in relation to the distribution of the time before the occurrence of a first event by means of a two-sided summary chi-square (log-rank) test14. P values and relative odds with 95 percent confidence intervals were determined annually and were designated as cumulative; the values summarized the outcome to the designated time15,16. Two-sided P values less than 0.05 were considered statistically significant.
Average annual incidence rates, relative risks, and 95 percent confidence intervals were calculated for each type of first event. The chi-square test was used to compare the average annual type-specific incidence rates in the treatment groups17. Relative risks and 95 percent confidence intervals were determined by univariate proportional-hazards analysis18. Cumulative incidence rates were plotted for events for which the comparison of treatments revealed a significant difference19. To test for interactions between the stratification variables and treatment in relation to event-free survival, multivariate proportional-hazards analysis was used18. The results presented here represent the first formal testing of the primary study hypothesis as originally planned on the basis of data obtained one year after the accrual of patients into the study.
Results
Distribution of Patients
A total of 818 women were enrolled in the trial (Table 1). None of the records obtained at one institution (representing 24 patients) could be evaluated, because of problems with the quality of the data. No follow-up results were available for 4 of the 794 women who could be evaluated (0.5 percent). Sixteen of the 790 women included in the analyses (2.0 percent) did not meet the criteria for study entry. Five years after the operation, follow-up results were up to date for 97.5 percent of the women who could be evaluated. All women who could be evaluated for whom there were follow-up data, including those who did not meet the criteria for entry into the study, were included in the analyses. The mean duration of follow-up was 43 months (range, 11 to 86). There were no differences between treatment groups in the distribution of selected characteristics of the patients and tumors (Table 2).
|
|
The women treated by lumpectomy followed by ipsilateral breast irradiation had significantly better event-free survival during five years of follow-up than the women treated by lumpectomy alone (84.4 percent vs. 73.8 percent, respectively; P = 0.001) (Figure 1). The cumulative odds ratio of 1.8 (95 percent confidence interval, 1.2 to 2.5) confirmed the beneficial effect of radiation therapy. There was a 47.3 percent reduction in the average annual incidence rate (per 100 women) for all first events (6.7 for lumpectomy alone vs. 3.8 for lumpectomy plus irradiation; P = 0.001) as a result of radiation therapy (Table 3). No significant interactions between stratification variables and treatment were identified.
|
|
Sixty-four of all 84 first events (76.2 percent) in the group treated by lumpectomy alone were ipsilateral breast tumors, as compared with 28 of 51 first events (54.9 percent) in the group undergoing lumpectomy and radiation therapy (Table 3). The average annual incidence rate of ipsilateral breast cancer was reduced by 58.8 percent as a consequence of radiation therapy (lumpectomy, 5.1; lumpectomy plus radiation, 2.1; P<0.001). The rates of both noninvasive cancers (2.6 and 1.5 in the respective groups; P = 0.055) and invasive cancers (2.6 and 0.6; P<0.001) were reduced, but the percent reduction was greater in the latter (42.3 percent vs. 76.9 percent). There was no reduction in the average annual incidence rate of other types of first events (contralateral breast cancers, other second cancers, and deaths from causes other than cancer) as a result of radiation therapy.
At five years of follow-up, in the women treated by lumpectomy alone, the cumulative incidence of ipsilateral noninvasive and invasive cancers was 10.4 percent and 10.5 percent, respectively, whereas in the group treated by lumpectomy and radiation, it was 7.5 percent and 2.9 percent (Figure 2). There was no significant difference between the two groups in the cumulative incidence of first events other than ipsilateral breast tumors (6.7 percent and 7.1 percent, respectively). Three of the 92 women in whom second tumors developed in the ipsilateral breast subsequently had distant disease. Two of the three were treated by lumpectomy alone, and cancer subsequently developed in the contralateral breast. The other woman was treated by lumpectomy and radiation, and liver metastases subsequently developed.
|
Among the 64 women whose primary ductal carcinoma in situ was treated by lumpectomy only and in whom a second breast tumor developed, 28 (43.8 percent) were treated by repeat lumpectomy and 36 (56.3 percent) by mastectomy. The percentage of women treated by a second local excision was similar regardless of whether the second tumors were noninvasive or invasive (43.8 percent). Twenty-eight women who received radiation therapy to treat their primary tumors had second tumors in the ipsilateral breast. Ten of the 20 who had noninvasive tumors (50 percent) were treated with lumpectomy, and 10 had mastectomies. All eight women whose second tumors were invasive had mastectomies. Six of the 54 women in both groups whose second tumors were treated by mastectomy had a prophylactic mastectomy of the contralateral breast.
Regional and Distant First Events
Few regional and distant events occurred as first events in either treatment group (Table 3). Three women had regional metastases. In two women whose ductal carcinoma in situ was detected from microcalcifications on a mammogram and who were treated by lumpectomy and radiation therapy, ipsilateral node metastasis subsequently developed (one after 15 months and one after 7 months) without evidence of disease in the breast. The third woman, treated by lumpectomy alone, also had ductal carcinoma in situ that was detected from microcalcifications; she had nodal metastases 5 1/2 years later.
Two women had distant metastatic breast cancer. In one of them, treated by lumpectomy alone, the breast cancer was found to have metastasized to her femur, with no evidence of disease in the breast eight months later. The other had concomitant contralateral supraclavicular nodal metastases, liver metastases, and extensive invasive ipsilateral breast cancer 12 months after lumpectomy and radiation therapy. She died one month after the diagnosis of metastasis.
First Events in the Contralateral Breast
Eighteen women, 8 in the lumpectomy group and 10 in the group treated by lumpectomy and radiation, had contralateral breast cancer as a first event. The initial cancer was detected by palpation in 1 woman, nipple discharge in 1, and mammography (revealing microcalcifications) in 16. Eleven of the contralateral breast cancers were invasive, and seven were ductal carcinoma in situ. Fourteen of the 18 contralateral tumors were treated by lumpectomy, 3 by mastectomy, and 1, an inflammatory cancer, by systemic therapy followed by mastectomy.
Second Primary Tumors as First Events
In the group treated by lumpectomy alone, five second cancers exclusive of those in the ipsilateral breast occurred as first events: two in the colon and one each in the pancreas, lung, and skin of the arm. Eight second cancers occurred in women who received radiation therapy: three in the colon and one each in the tonsil, esophagus, uterus, lung, and the skin of the arm.
Death as a First Event
Death from causes other than cancer occurred as a first event in five women in the lumpectomy group and two in the group receiving radiation therapy. One of the 92 women in whom a second ipsilateral breast tumor developed as a first event died of breast cancer. There were only two other deaths from breast cancer. One occurred in the woman who had extensive distant disease and concomitant ipsilateral breast involvement as a first event after lumpectomy and radiation therapy. The other woman who died of breast cancer, also in the group receiving radiation, had supraclavicular nodal metastases as the first event.
Discussion
The results of this randomized trial of therapy in women with ductal carcinoma in situ indicate that radiation therapy after lumpectomy improved overall event-free survival as compared with lumpectomy alone. All first events occurring after the initial therapy were considered in the life-table analysis, ensuring that ipsilateral breast tumors were accounted for and that events that could have been attributed to undesirable sequelae of therapy were included. Moreover, the inclusion of all the first events addressed the problem of competing risks by taking into account women who had metastases to a regional or distant site, had second tumors other than in the ipsilateral breast, or died from other causes before a second ipsilateral breast tumor developed.
The benefit of radiation therapy was most evident when each type of first event was evaluated according to its contribution to the overall effect. As expected, there was no evidence that irradiation reduced the average annual incidence rates or the cumulative incidence of non-cancer-related deaths, second primary cancers (either in the contralateral breast or at other sites), and regional or distant metastatic disease. The benefit of radiation therapy was related to its ability to decrease the cumulative incidence of all second ipsilateral breast tumors as compared with that in the group treated only by lumpectomy. This reduction was evident for both noninvasive and invasive second cancers.
It might be considered that women with ductal carcinoma in situ should be treated by mastectomy, since women treated by lumpectomy alone had a 20.7 percent estimated risk of a second ipsilateral breast tumor. However, since radiation therapy significantly reduced the incidence of both noninvasive and invasive second tumors to 7.5 percent and 2.9 percent, respectively, mastectomy may be difficult to justify to treat localized ductal carcinoma in situ. Furthermore, since the women in whom another ductal carcinoma in situ developed after radiation therapy are likely to have the same prognosis that they had when the first lesion was identified, that contention is strengthened. The development of an invasive lesion after the removal of ductal carcinoma in situ and treatment by radiation therapy, as occurred in 2.9 percent of the women in our study, may be viewed differently. It may be argued that those invasive cancers, with their potential for dissemination, could have been prevented had the primary tumor been treated by mastectomy. If the incidence of such cancers remains low, however, and if they are detected when they are small, few of them should result in distant disease and death20.
Pathologists may have difficulty differentiating between benign breast lesions (e.g., highly atypical hyperplasia) and ductal carcinoma in situ, between ductal carcinoma in situ and lobular carcinoma in situ, and between ductal carcinoma in situ and a focus of invasive cancer, especially when the latter is associated with ductal carcinoma in situ21. In large randomized trials, women with tumors of uncertain type should be equally distributed among the treatment groups, thus eliminating that source of bias. Moreover, the findings from a large trial that includes women who have ductal carcinoma in situ with a focus of invasive tumor should be applicable to the general population of women with ductal carcinoma in situ, some of whom could have tumors with undiagnosed invasive components. A review of slides submitted to the pathology center demonstrated invasive cancers in 1.5 percent of the primary tumors surveyed and atypical hyperplasia in 4.7 percent. The discordance in the diagnosis could have been related at least in part to a sampling error -- i.e., the slides submitted may not have been representative of the material from which the original diagnosis of ductal carcinoma in situ was made. Because the discordant diagnoses were few and were distributed between the two groups of women, they did not affect the conclusions of the study.
After the local excision of ductal carcinoma in situ, second ipsilateral tumors are almost always at or close to the site of the primary tumor21. In this study, 13.3 percent of the second tumors were in a different location, a proportion somewhat higher than is usually reported. Because most information about concordance comes from the study of women with invasive cancer, there may be a difference between noninvasive and invasive tumors in that regard.
In summary, since the incidence of second ipsilateral tumors is significantly reduced by breast irradiation after lumpectomy, combined therapy is more appropriate than lumpectomy alone for women with localized intraductal carcinoma in situ.
Supported by Public Health Service grants from the National Cancer Institute and by grants from the American Cancer Society.
We are indebted to Tanya Spewock for her assistance in the preparation of the manuscript.
Source Information
Institutions contributing 10 or more patients to the study are listed in the Appendix, together with the names of the principal investigators.
From the National Surgical Adjuvant Breast and Bowel Project Headquarters, Rm. 914, Scaife Hall, 3550 Terrace St., Pittsburgh, PA 15261, where reprint requests should be addressed to Dr. Fisher.
References
The following institutions, listed with the names of the principal investigators, each contributed 10 or more patients to this study.
Baptist Medical Center, Oklahoma City: K.K. Boatman; Boston University, Boston: M.T. Kavanah; British Columbia Cancer Agency, Vancouver, Canada: I.H. Plenderleith and I. Olivotto; Community Clinical Oncology Program (CCOP), Allegheny, Pittsburgh: R. Pugh; CCOP, Alton Ochsner Medical Foundation, New Orleans: C.G. Kardinal; CCOP, Columbia River Oncology Program, Portland, Oreg.: G.L. Doty; CCOP, Marshfield Clinic, Marshfield, Wis.: J.L. Hoehn; City of Hope Medical Center, Duarte, Calif.: L.D. Wagman; Good Samaritan Hospital, Cincinnati: R. Welling; Hartford Hospital, Hartford, Conn.: P.A. DeFusco; Hotel-Dieu, Montreal: A. Robidoux; Jewish General Hospital, Montreal: R.G. Margolese; Kaiser Permanente, Portland, Oreg.: A.G. Glass; Medical College of Wisconsin, Milwaukee: A.P. Walker; Michigan State University, E. Lansing, Mich.: N.V. Dimitrov; Mount Sinai Medical Center, Cleveland: R. Bornstein and L. Levy; Pennsylvania Hospital, Philadelphia: H.J. Lerner; St. Michael's Hospital, Toronto: L. Mahoney; Tufts University-New England Medical Center, Boston: D.D. Karp; University of California, Los Angeles: A.E. Giuliano; University of Hawaii, Honolulu: R. Oishi; University of Michigan, Ann Arbor: B.L. Weber; and University of Pittsburgh, Pittsburgh: B. Fisher.
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Radiation Therapy for in Situ or Localized Breast Cancer
Verhoeven D., Van Marck E., van Oosterom A.T., Lagios M. D., Page D. L., Sinoff C. L., Fisher B., Redmond C. K., Fisher E., Veronesi U.
Extract |
Full Text
N Engl J Med 1993;
329:1577-1579, Nov 18, 1993.
Correspondence
The NSABP Trials
Levenback C., Bross I. D., Davis N., Heitjan D. F., Altman L. K., Angell M., Kassirer J. P.
Extract |
Full Text
N Engl J Med 1994;
331:809-810, Sep 22, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
