Prospective Randomized Clinical Trial Data
In 1985, the National Surgical Adjuvant Breast Project (NSABP) began protocol B-17, a prospective randomized study to evaluate the worth of postoperative radiation therapy (RT) following lumpectomy for patients with DCIS. The initial clinical and pathologic results were published in 1993 and 1995.53,54
In 1997, the results were updated.55 For this analysis, 814 patients were available with a mean time in study of 90 months (range: 67 to 130 months). Eighty percent of patients had tumors detected by mammographic screening. Nega-tive margins, defined as tumor-filled ducts not touching ink, were required. There were a total of 151 ipsilateral breast tumor recurrences (IBTR), 70 (46.4 percent) of which were invasive. Almost all of the ipsilateral breast tumor recurrences were at or near the original lesion.
The rate of IBTR was markedly reduced by breast irradiation. The incidence of invasive recurrence was 3.9 percent in the radiated group compared with 13.4 percent in the nonirradiated group (p = 0.000005). The incidence of recurrent DCIS was also significantly reduced, from 13.4 percent in the group with no radiation to 8.2 percent in the radiated group (p = 0.007). The overall survival rate did not differ between groups: 94 percent for patients treated by lumpectomy alone, 95 percent for lumpectomy and RT.
The impact of pathologic features on IBTR was reported for a subset of 623 patients in NSABP B-17 with eight-year follow-up.56 The cumulative frequency of IBTR was 137 (22 percent) for all 623 patients. Ninety-four of 303 (31 percent) occurred in the lumpectomy-only group, and 43 of 320 (13 percent) in those receiving RT. This represented a 61 percent relative reduction in IBTR for patients receiving RT (log rank test, p < 0.0001).
Nine pathologic features were examined for prognostic significance, and only moderate-to-marked comedonecrosis was an independent predictor for IBTR in nonirradiated patients. RT reduced the eight-year risk of recurrence in the breast from 40 to 14 percent in patients with moderate- or marked-comedonecrosis. Patients with absent or slight comedonecrosis experienced a decrease in local recurrence from 23 to 13 percent with RT. Of note, in irradiated patients comedonecrosis was not a predictor of an increased risk of breast recurrence. Margin status was not found to be a significant predictor of recurrence in this study, but it is likely that the definition of a negative margin that was used (tumor-filled ducts not touching ink) and the lack of post-excision mammo-grams resulted in some patients with significant residual DCIS being included in the “negative” margin group. For the “most favorable” group in the study, those with negative margins and absent or slight comedonecrosis, the addition of RT to excision resulted in a 7% absolute reduction in local failure at eight years.56
Of the 818 patients in the B-17 trial,57 only 14 had died of breast cancer with a mean follow-up of 90 months. Three deaths occurred after IBTR, six occurred in patients who had regional failure without recurrent breast tumor, and six patients developed distant metastases without locoregional disease. These findings indicate that even the most meticulous local control in the breast will not eliminate all breast cancer mortality in patients diagnosed with DCIS.
A second prospective randomized trial of the role of RT in DCIS was initiated by the European Organization for Research and Treatment of Cancer (EORTC) in 1986 and completed accrual of patients in 1996.58 Women were eligible for this trial if they had clinically or mammographically detected DCIS measuring five cm or less in size. Mammographic lesions were present in 71 percent of the study population. After complete local excision, 503 women were randomly assigned to observation with no further treatment, and 507 were randomized to postoperative radiotherapy at a dose of 50 Gy in five weeks to the whole breast. The median duration of follow-up at the time of the initial report was 4.25 years.58,59 The four year local relapse-free rate was 84 percent in the group treated with surgery only compared with 91 percent in women treated by postoperative radiotherapy (log rank p = 0.005; hazard ratio 0.62). Comparable reductions were seen for the risk of invasive (40 percent, p = 0.004) and noninvasive (35 percent, p = 0.06) local recurrence. No differences in regional recur-rences, distant metastases, or survival were noted.
Retrospective Series Data
The results of conservative surgery and radiation for DCIS from retrospective series are presented in Table 4.60–73 The crude incidence of breast recurrence ranges from four to 18 percent. Deaths due to breast cancer have been reported in up to four percent of patients treated in studies with a median follow-up of ten years or less.
Table TABLE 4. Results of Conservative Surgery and Radiation for Clinically and Mammographically Detected DCIS
|McCormick et al.60||54||18||100||3.0|
|Haffty et al.61||60||7||100||3.6|
|Kurtz et al.62||47||4||100||5.0|
|Ray et al.63||56||9||—||5.0|
|Solin et al.64||51||10||—||5.7|
|Van Zee et al.65||65||10||—||6.2|
|Hiramatsu et al.66||76||9||100||6.2|
|Sneige et al.67||49||10||—||7.2|
|Fourquet et al.68||153||16||—||9.0|
|Collaborative Group69, 70||268||17||97||10.3|
|Amichetti et al.71||139||9||100||6.8|
|Beron et al.72||185||16||99||7.5|
|Mirza et al.73||87||13||99||11.0|
The long-term results of conservative surgery and radiation for DCIS were reported by Solin, et al.69,70 This collaborative study of ten institutions in the United States and Europe analyzed outcomes in 259 patients. Seventy-eight percent of the tumors were detected solely by mammography. The ten-year actuarial risk of breast recurrence was 16 percent, and the ten-year actuarial cause-specific survival was 97 percent.69 The 15-year actuarial breast recurrence was 19 percent, and the 15-year actuarial cause-specific survival was 96 percent. Median follow-up was 10.3 years.70
Various clinical, pathologic, and treatment-related factors have been assessed for their ability to identify patients with a substantial risk of recurrence in the treated breast for whom mastectomy may be recommended. One factor for which there appears to be agreement in terms of its association with a high risk of recurrence is the presence of residual malignant-appearing calcifications on a post-biopsy mammogram. Failure to remove these calcifications prior to radiation has resulted in a 100% recurrence rate in the few patients reported.60,67 DCIS presenting as a bloody nipple discharge was noted in earlier series to be associated with a higher risk of recurrence. However, in the collaborative study, there appeared to be no increased risk in this group of patients.70
The significance of young age (less than 40 years) is controversial. Three studies have observed an increased risk of breast recurrence (approximately 25 percent) in young women with DCIS treated with conservative surgery and radiation when compared with older women (approximately 10 percent).65,74–76 However, three additional studies have found no correlation with young age and breast recurrence rates.66,67,70 The effect of age on local failure was analyzed in NSABP B-24, a prospective randomized study of 1,804 women with DCIS. All patients received radiotherapy and were randomized to tamoxifen at 20 mg daily for five years or placebo.45 Negative margins were not required. The rate of ipsilateral breast recurrence in women age 49 or less in the placebo arm was 33.3 per 1,000 per year, compared with 13.03 per 1,000 per year for those age 50 and older. For those taking tamoxifen, recurrence rates were 20.77 per 1,000 per year for those aged 49 and under, and 10.19 per 1,000 per year for those in the older age group. This randomized trial provides convincing evidence that young age is associated with a higher rate of breast recurrence. The effect of age on the risk of breast recurrence has recently been reviewed in detail.77
A similar controversy exists with a positive family history of breast cancer. Two series60,66 have reported a higher breast recurrence rate (approximately 40 percent) in women with a positive family history when compared with those with no such history (approximately 10 percent). However, a third series found no such association.75 The impact of a positive family history of breast cancer on treatment options in women with DCIS requires further evaluation.
The contribution of various pathologic factors (histologic subtype, nuclear grade, necrosis) to the risk of breast recurrence in patients treated with conservative surgery and radiation is controversial. It was initially suggested that high-grade or comedo DCIS was associated with a higher breast recurrence rate.69,78 However, in the collaborative study, the ten-year actuarial breast recurrence rate was 18 percent for tumors with the combination of both comedo pattern and a high nuclear grade versus 15 percent for DCIS in which these factors were absent (p = 0.15).70 The median interval to recurrence for comedo DCIS was 3.1 years versus 6.5 years for the noncomedo DCIS. Therefore, series with shorter follow-up tend to underestimate the number of recurrences in low-grade or noncomedo DCIS, and recurrences in the high grade or comedo DCIS predominate. As previously discussed, NSABP B-17 found that the presence of comedonecrosis was not a predictor of breast recurrence when RT was given.54
The majority of breast recurrences in patients undergoing conservative surgery and radiation for DCIS occur in the vicinity of the primary tumor, and approximately 50 percent are invasive cancers.61,63,66–70,79 Invasive recurrences appear at later intervals than noninvasive and may occur in a separate quadrant.45,62,74 Virtually all of the patients who develop a noninvasive recurrence, and approximately 75 percent of those with an invasive recurrence, are long-term survivors after mastectomy.60,61,63,66–68,79–81
Over the last ten years, there has been a significant change in the method of detection of DCIS. Approximately 85 to 90 percent of all DCIS is now detected solely as a mammographic finding, which is most often characterized by the presence of micro-calcifications. The earlier reports of conservative surgery and radiation for DCIS do not accurately reflect outcome for mammo-graphically detected DCIS since many included clinically evident DCIS (palpable mass or bloody nipple discharge), and detailed mam-mographic and pathologic correlation was frequently lacking. Unfortunately, the results of these earlier series were used for comparisons with those of conservative surgery alone for mammographically detected DCIS and not infrequently claimed to be equal. The NSABP B-24 trial45 prospectively documented that the risk of local failure for clinically evident DCIS was approximately twice that of mammographically detected DCIS.
The results of conservative surgery and radiation for mammographically detected DCIS are presented in Table 5.56,57,66,67,69,70,75,76,78,79 The ten-year actuarial breast recurrence rate ranges from 6 to 23 percent, with a ten-year cause-specific survival rate of 96 to 100 percent. The variation in the results reported reflects differences in patient selection, the extent of surgical resection, and the degree of mammographic and pathologic correlation. There is increasing evidence that wide surgical excision67 and negative margins of resection diminish the risk of a breast recurrence in patients with mammographically detected DCIS treated with conservative surgery and radiation.45,67,73 In the collaborative study with a median follow-up of 9.3 years, the crude breast recurrence rate was 29 percent for patients with a close or positive margin compared with seven percent for negative margins.73 In NSABP B-24,45 patients with positive margins had a significantly higher rate of breast recurrence than those with negative margins, regardless of whether tamoxifen was given (RR 1.68, 95, CI 1.20-2.34).
Table TABLE 5. Results of Conservative Surgery and Radiation for Mammographically Detected DCIS
|NSABP B-1756,57||411||10.0||12.1||—||—||96.0||—||7.5 mean|
|Kuske et al.78||44||7.0||—||—||—||—||—||4.0|
|Fowble et al.75||110||1.0||—||15.0||100.0||—||100.0||5.3|
|Kestin et al.76||146||8.0||—||9.2||100.0||—||99.2||7.2|
|Hiramatsu et al.66||54||2.0||—||23.0||—||—||96.0||6.2|
|Sniege et al.67||31||0.0||—||8.0||—||—|| ||7.2|
|Silverstein et al.79||133*||7.0||—||19.0||—||—||97.0||7.8|
Two series have reported the results of conservative surgery and radiation for mammographically detected DCIS in patients who would meet Lagios' criteria for observation without radiation. These criteria include calcifications as the only manifestation of DCIS, DCIS size < 2.5 cm, negative margins, and negative post-biopsy mammo-gram.73,75 In these two studies, there have been no breast recurrences reported to date in the 37 patients (median follow-up 4.9 and 9.3 years). In comparison, Lagios reported a 17% breast recurrence rate in 78 such patients treated by excision alone with a follow-up of 10.3 years.82