Breast-conservation therapy with lumpectomy and breast irradiation is
a widely accepted, pre
ferred treatment option for both patients with early-stage invasive breast cancer and those with ductal carcinoma in situ (DCIS). Numerous randomized trials demonstrate that, in women with invasive cancer, such breast-conserving therapy provides survival outcomes equivalent to those achievable with masectomy.[2-6] The same appears to be true for DCIS,[7-9] although no randomized comparative data are available. Due to the cost, inconvenience, and potential toxicity of breast irradiation, there is continued interest in identifying subsets of patients who may be "adequately" treated with lumpectomy alone. This article reviews the results of clinical studies assessing lumpectomy without radiation therapy for early-stage invasive cancer and DCIS. The studies reviewed have varying follow-up periods. Thus, to facilitate comparisons between the studies, the annual breast failure rate will be the primary outcome reported. If not reported in the study, the annual failure rate was estimated by dividing the overall failure rate by the average duration of follow-up.
One randomized study comparing lumpectomy alone to lumpectomy plus radiation therapythe National Surgical Adjuvant Project for Breast Cancer (NSABP) B-17 trial[10-12]has been published (Table 1). Overall, in this trial, the addition of radiation reduced the annual breast failure rate from approximately 5% to 2%.
In a retrospective subset analysis, the addition of radiation decreased the annual breast recurrence rate in all subgroups, regardless of margin status and the presence or absence of moderate/marked comedo necrosis. In the subgroup of patients with the most favorable risk profile, ie, those with clear margins and absent/minimal comedo necrosis, the absolute reduction in the annual breast failure rate was less than 1%. In the "highest-risk" group, ie, those with positive margins and prominent necrosis, the addition of radiation reduced the absolute annual breast failure rate by 7% (from 10% to 3%).
Additional randomized data are available from NSABP B-06.[7,8] This study was intended to include only patients with invasive cancer. However, on histologic review, a small subset of patients were judged to have DCIS. In NSABP B-06, as in the larger NSABP B-17 study, the addition of radiation reduced the breast failure rate from approximately 6% to 1%.
Following breast-conserving therapy for DCIS, recurrent lesions may be either invasive or noninvasive. In the NSABP B-17 trial, the addition of breast irradiation reduced the fraction of recurrences that were invasive. Among the group treated with local incision alone, 50% of the recurrences were invasive, whereas in the irradiated group, only 29% of the recurrences were invasive. Thus, the addition of radiation reduced both the incidence of a breast recurrence and the fraction of recurrences that were invasive.
Looking at the data another way, the annual rate of an invasive breast cancer recurrence was reduced from approximately 2.6% to 0.6% (a 76% relative risk reduction [P < .001]). Since patients with DCIS who develop an invasive recurrence are likely to be at a higher risk of death from distant metastases than those who have a recurrence of DCIS, this is an interesting observation.
No differences in overall survival were reported in the NSABP B-17 trial.
The data from multiple nonrandomized series using lumpectomy without radiation therapy are summarized in Table 2.[13-26] The annual breast failure rate ranges from 2% to 6%, with a weighted mean of ~4%.
In the first six studies listed in Table 2, most or all of the patients had small mammographically detected lesions and negative resection margins. In this subgroup of studies, the breast failure rate ranged from 2% to 4.6%, with a weighted mean of 3.2%. In all of these studies, the observed annual breast failure rate exceeds that generally reported following lumpectomy plus radiation therapy.[9,10]
Prognostic FactorsIn several of the studies listed in Table 2, various prognostic factors were identified that are related to the breast failure rate. Although the particular factors identified in the individual studies vary, alow-grade, noncomedo subtype appears to have the most favorable prognosis (Table 3).[14-20, 25, 26] The annual breast recurrence rate among patients with these "favorable" factors ranged from 0% to 3.8%.
Dr. Silverstein and colleagues at the Breast Center in Van Nuys, California, recently described a prognostic index based on tumor size, margin width, and histologic appearance. In their retrospective analysis, patients with the most favorable prognosis, ie, those with the lowest index score (small tumor, wide excision margin, and less aggressive pathology), appeared to have a very low breast recurrence rate following lumpectomy without radiation therapy. These patients did not benefit from breast irradiation.
In contrast, patients with a moderate index score (5 to 7) had a higher relapse rate following lumpectomy alone and did derive benefit from breast irradiation. Patients with the least favorable tumors (index score 8 or 9) also derived a modest benefit from radiation therapy, although the breast relapse rate in these patients was unacceptably high with or without radiation. These results conflict with those of the randomized NSABP B-17 trial. The number of patients in these nonrandomized reports is small. Also, duration of follow-up in both the randomized and nonrandomized studies is short. Additional follow-up data from all of these studies will be interesting.
The data from five randomized studies comparing outcomes after lumpectomy alone to those after lumpectomy plus radiation therapy are summarized in Table 4.[2-5,27-33] Only patients with negative resection margins and relatively small (£2 to 4 cm) tumors were included in these studies. In each study, the addition of radiation reduced the annual breast recurrence rate from approximately 5% to 1%. No survival differences were reported (discussed below).
Table 5 outlines the results from several nonrandomized series in which radiation therapy was omitted following lumpectomy.[29, 30, 34-42] It is likely that patients in these series were selected for treatment without irradiation because they were perceived to have more "favorable" characteristics (eg, most of the studies included only patients with small tumors). Despite this, the observed annual breast failure rate was generally greater than the 1% typically seen with lumpectomy plus radiation therapy.
The study by Schnitt et al represents the "cleanest" group of women prospectively selected for lumpectomy without irradiation. This study enrolled 86 patients with a solitary T1 infiltrating cancer, without lymphatic vascular invasion or an extensive intraductal component, who had microscopically negative margins by at least 1 cm (all but two of the patients had negative reexcisions). All 87 patients had pathologically negative axillary nodes. Approximately 75% of the lesions were detected mammographically, and the median age of the patients was 67 years (range, 27 to 84 years).
With a median follow-up of 56 months, the annual breast failure rate was 3.6%. This is a remarkably high breast recurrence rate, in light of the favorable nature of the tumors treated.
In a subset analysis, no recurrences were seen in 42 of the patients with the "most favorable" histologic findings (26 grade 1 infiltrating, 9 mucinous, and 7 tubular tumors). It is difficult to know how to interpret this type of retrospective identification of a most-favorable subgroup, since all of the enrolled patients were initially considered to have a favorable prognosis. Indeed, if all 14 recurrences are assigned to the remaining 45 "less-favorable" patients, the annual recurrence rate in this group is 6.6%.
Results in Elderly PatientsSome investigators have selected women for treatment without irradiation based on their age. Table 6 outlines the results reported in "elderly" women treated with lumpectomy without radiation therapy.[41-49] As there is no uniform definition of "elderly" in these studies, the range of patient age is included.
The lowest failure rates (0% to 2.75%) were noted in the study populations of Nemoto et al and von Rueden et al, in whom margins were usually negative and tumor sizes were less than 2 to 5 cm.[42,43] However, the studies by Lee et al and Reed and Morrison reported an ~5% annual failure rate in similarly selected patients.[47,49] Higher recurrence rates were generally reported in series that did not specify margin status. The average follow-up in most of these studies was short (less than 4 years), and the patient numbers were small.