The goal of conservative management of breast carcinoma with conservative surgery (CS) and radiation therapy (RT) is to preserve the cosmetic appearance of the breast without compromising the survival of the patient. Although the effect of recurrence in the breast on distant disease free survival remains in question,1, 2 it is useful to estimate the risk of local recurrence following treatment with CS and RT in order to advise patients with early stage breast carcinoma of their treatment options, and thereby select patients who may be most appropriately treated with conservative management. In recent years, several factors that appear to place women at an increased risk of local recurrence after CS and RT have been described. Among these are the status of margins of excision,3 the interval between surgical excision and start of radiation therapy,4 the presence of lymphatic vessel invasion,5 tumor grade,6 and young age.7 At the same time, factors that do not appear to be prognostic for local recurrence include the size of the tumor8 and the number of involved axillary lymph nodes.1
Lobular carcinoma in situ (LCIS) was first identified by Foote and Stewart as a noninvasive lesion arising from the terminal ducts and lobules.9 The reported incidence of this in the general population has been variable, ranging from 0.5% to 3.6%%;10, 11 this wide range is at least partly related to variations in pathologic definition. In general, LCIS is an incidental finding with no specific mammographic features and with no macroscopic lesion identified on examination of the pathologic specimen.12 Although the early reports of LCIS emphasized the increased risk of developing a subsequent invasive cancer and therefore recommended treatment with mastectomy,9 more recent series have shown that the subsequent risk of invasive cancer is bilateral,13 and LCIS is now most commonly managed by observation.14
The relation between the presence and extent of LCIS and the risk of local recurrence in patients with invasive breast carcinoma treated with CS and RT has not been previously studied in detail. In an attempt to assess the impact of LCIS on local recurrence in patients given breast-conserving treatment, we retrospectively reviewed the data base of the Joint Center for Radiation Therapy at Harvard Medical School in Boston, Massachusetts.
MATERIALS AND METHODS
- Top of page
- MATERIALS AND METHODS
Between 1968 and 1986, 1625 patients with clinical Stage I and II invasive breast carcinoma were treated with CS and RT at the Joint Center for Radiation Therapy. with complete macroscopic excision of the tumor. Among these patients, 1181 had infiltrating ductal carcinoma (IDC), infiltrating lobular carcinoma (ILC), or infiltrating carcinoma with mixed ductal and lobular features (IMC); had sufficient breast tissue adjacent to the invasive tumor to assess the presence or absence of LCIS; received at least 60 gray (Gy) to the tumor bed; and were evaluable with a minimum follow-up period of 8 years. These patients constitute the population for this study. Patient characteristics are shown in Table 1.
Table 1. Patient Characteristicsa
|LCIS −||LCIS +||LCIS −||LCIS +||LCIS −||LCIS +|
|No. of patients||984||42||22||70||38||25|
|% Axillary dissection||81% (801)||81% (34)||59% (13)||76% (53)||79% (30)||68% (17)|
|% Lymph node +||38% (308)||32% (11)||23% (3)||34% (18)||50% (15)||35% (6)|
|% LVI +||38%||42%||5%||24%||34%||17%|
|% EIC +||25%||31%||10%||6%||16%||12%|
|% Margin evaluable||32% (317)||33% (14)||45% (10)||40% (28)||47% (18)||24% (6)|
|% Margin +||40%||14%||10%||39%||50%||50%|
|Median dose (Gy)||65.35||63.91||63.00||64.65||63.67||65.30|
|Adjuvant systemic therapy||29% (285)||24% (10)||14% (3)||23% (16)||39% (15)||28% (7)|
All available histologic sections from each case were reviewed for a comprehensive list of pathologic features, as described previously.7 Tumors were classified as IDC (n=1026) and ILC (n=92) according to published histologic criteria.15 The ILC group included classical as well as variant subtypes. In 63 other cases, the tumors had some histologic features of IDC and some characteristics of ILC; these were classified as IMC.
LCIS was defined, in accordance with the guidelines of the second edition of the World Health Organization Histological Typing of Breast Tumors, as “A carcinoma involving the intralobular ductules, which are obliterated and distended by loosely aggregated cells, without stromal invasion.”15 The presence of LCIS within and/or adjacent to the invasive tumors was recorded. For IDC, LCIS was only scored in the adjacent breast tissue; for ILC and IMC, LCIS was scored both within and adjacent to the infiltrating tumor. The extent of LCIS in association with the invasive carcinoma was scored as absent, scant, moderate, or marked. During the time period of this study, margins were not routinely inked for evaluation in all cases.
An axillary dissection was performed in 948 patients (80%). Positive axillary lymph nodes were found in 361 patients (38%).
The median radiation dose to the whole breast using tangential fields was 46 Gy, with a supplemental median dose to the tumor bed of 19.35 Gy. The median dose to the tumor bed was 65.20 Gy. Details of treatment techniques have been described elsewhere.16
Patients were treated with systemic chemotherapy or adjuvant hormonal therapy at the discretion of their medical oncologist. Three hundred fifteen patients (27%) received multiagent chemotherapy, 22 (2%) received tamoxifen, and 837 (71%) received no adjuvant therapy.
Patients were generally followed with frequent physical examinations and yearly mammograms. At the time this study was conducted, it was not our policy to recommend yearly bone scans or other radiologic imaging.
Outcome was calculated as the crude percentage of failure within 8 years from the start of radiation therapy using methodology described elsewhere.17 The crude rate of local recurrence (rather than an actuarial rate) was used in order to prevent bias from being introduced by the competing risk of distant disease or death. In particular, if patients in the study group had a substantially higher risk of distant failure, the actuarial rate of local control would artificially appear to be improved due to censoring at the time of distant progression. For example, the rate of local failure for patients with positive lymph nodes treated with CS and RT may appear to be superior to that of patients with negative lymph nodes, because patients with positive lymph nodes have a higher rate of distant disease. This bias is not encountered with the use of crude recurrence rates. Although the median follow-up was 161 months, we chose to restrict the time period of the study to 8 years in order to avoid the possibility of breast failures being preferentially reported. The site of first failure was described either as local (breast) with or without other sites or as distant or regional lymph node failure. Some patients were also categorized as dead of other (non-cancer-related) causes in the absence of local failure or alive with no evidence of disease (NED). The median follow-up time for surviving patients was 161 months.
- Top of page
- MATERIALS AND METHODS
Among the 1181 patients, 137 (12%) had LCIS within the invasive tumor, within the macroscopically normal adjacent tissue, or both. At 8 years, the crude distribution of outcome by site of first failure for the population as a whole was as follows: local failure, 12%; distant or regional failure, 21%; opposite breast carcinoma, 6%; dead without evidence of disease, 4%; and alive without evidence of disease, 57%. The corresponding distributions by histologic type and the presence of any LCIS are listed in Table 2. Neither factor had any significant influence on the distribution of outcomes. The 8-year actuarial overall survival rates by histologic type were 75% for patients with IDC, 79% for patients with ILC, and 78% for IMC (all P=not significant [NS]). The corresponding rates by the presence of LCIS were 74% for patients with no LCIS and 84% for patients with any LCIS (P = NS).
Table 2. Crude Distribution of Outcome at 8 Years by Site of First Failure
| ||No. of patients||Local failure||Distant/regional failure||Opposite breast carcinoma||Dead with NED||Alive with NED|
|IDC||1026||12% (122)||21% (219)||6% (57)||4% (44)||57% (584)|
|ILC||92||13% (12)||20% (18)||5% (5)||4% (4)||58% (53)|
|IMC||63||13% (8)||24% (15)||8% (5)||5% (3)||51% (32)|
|All||1181||12% (142)||21% (252)||6% (67)||4% (51)||57% (669)|
As the presence or absence of LCIS within the tumor was scored only for the ILC and IMC subgroups, outcome was analyzed separately for LCIS within the tumor and LCIS adjacent to the tumor, and further stratified by histologic group. LCIS was present adjacent to the invasive cancer in 42 of 1026 patients with IDC (4%), 55 of 92 patients with ILC (60%), and 22 of 63 patients with IMC (35%). LCIS was found within the invasive tumor in 56 of 92 ILC cases (61%) and 17 of 63 IMC cases (27%).
The relation between the presence and extent of LCIS adjacent to the invasive tumor and the 8-year crude rates of LF and opposite breast carcinoma is summarized in Table 3. No correlation was found between the presence or extent of LCIS adjacent to the invasive tumor and the likelihood of either local failure or opposite breast carcinoma within the first 8 years after treatment. This was true for all three histologic subgroups. Similarly, there was no apparent correlation between the presence or extent of LCIS within the invasive tumor and the likelihood of either local failure or opposite breast carcinoma within the first 8 years after treatment for either the ILC or the IMC subgroup (data not shown).
Table 3. Correlation of Extent of LCIS Adjacent to Infiltrating Tumor and Risk of First Failure
|Histology||LCIS adjacent||No. of patients||8-yr crude rate of failure|
|Local failure||Opposite breast carcinoma|
|IDC||Absent/scant||1002||12% (121)||5% (55)|
|Moderate/marked||24||4% (1)||8% (2)|
|ILC||Absent/scant||59||12% (7)||8% (5)|
|IMC||Absent/scant||50||10% (5)||10% (5)|
|All||Absent/scant||1111||12% (133)||6% (65)|
|Moderate/marked||70||13% (9)||3% (2)|
We attempted to assess the interaction between LCIS and an extensive intraductal component (EIC) and their combined impact on local tumor control. Given the previously described negative results, the histologic groups were combined (of note, EIC is not normally scored for ILC or IMC but was for this analysis; only 10% [15 of 154 evaluable cases] of these were EIC positive compared with 25% [260 of 1024] of the IDC group). Patients who were EIC positive had a higher crude rate of local failure at 8 years than patients who were EIC negative (23% vs. 9%; P < 0.001). EIC positive patients who had any degree of LCIS adjacent to the invasive tumor had a 16% local failure rate (3 of 19 patients) compared with 23% (60 of 256) for EIC positive patients without LCIS in the adjacent tissue. Among EIC negative patients, those with LCIS had a 12% local failure rate (12 of 100) compared with 8% (68 of 803) for patients without LCIS.
- Top of page
- MATERIALS AND METHODS
In the current series, we attempted to assess whether the presence and extent of LCIS was associated with outcome for patients with invasive breast carcinoma treated with CS and RT. With a median follow-up time of 161 months, the 8-year distribution of first sites of failure was not significantly related to presence or extent of LCIS in the group as a whole or in any subset evaluated. Furthermore, the extent of LCIS did not appear to be a significant factor for local recurrence or the development of opposite breast carcinoma. Thus, unlike our experience with this study population that demonstrated the importance of the extent of associated DCIS, the presence and extent of LCIS did not appear to influence the rate of local or distant recurrence.
The presence of LCIS in an otherwise benign breast biopsy has been shown to be associated with an elevated risk of a subsequent invasive breast carci- noma.13 However, because the invasive cancer seems to be equally likely to appear in either breast, LCIS is generally considered to be a marker for a patient's risk of developing invasive disease rather than a precursor lesion.11 In the current series, the incidence of contralateral breast carcinoma was similar in patients with and without LCIS in the initial specimen. This may cautiously be interpreted as showing that although patients with an ipsilateral cancer are more likely to develop a contralateral tumor compared with the population as a whole, the presence of LCIS does not significantly increase this risk.
The National Surgical Adjuvant Breast Project (NSABP) has recently reviewed the patterns of recurrence in patients with LCIS treated with lumpectomy alone in Protocol B-17.18 With a mean follow-up time of 60 months, there were 17 events in the ipsilateral breast, of which 3 were invasive lobular carcinoma and 6 were LCIS. One patient had both an ipsilateral finding of invasive lobular carcinoma and a contralateral lesion consisting of LCIS; three other patients developed contralateral disease (two with invasive lobular carcinoma and one with LCIS and DCIS). No patient developed invasive ductal carcinoma. The projected incidence was 2.2% for invasive ipsilateral recurrence in the breast and 1.1% for invasive contralateral carcinoma. This was considered significantly less than in previous series.
Previous publications on the influence of LCIS and the risk of local recurrence after CS and RT have been limited (Table 4). In the series studied by Moran and Haffty, there was no association between the presence of LCIS and the risk of local recurrence.19 However, this conclusion must be interpreted with caution. There was apparently no centralized pathology review, and the interpretation on LCIS was therefore dependent on several independent pathologists, possibly leading to misclassification of cases. The series studied by Griem et al. has been presented only in abstract form.20 Although there were only three failures in the 21 patients with LCIS treated with CS and RT, this was considered to suggest a higher local recurrence rate for patients with LCIS compared with the 5% local failure rate for patients without LCIS.
Table 4. Comparison of Series of LCIS in Conservatively Treated Breast Carcinoma
|Series (by institution)||No. of patients||Follow-up (mos)||No. of patients with LCIS||Risk of local failure|
|Rush-Presbyterian20||N.S.||37||21||17% (5-yr)||5% (5-yr)|
|Current series||1181||161||137||12% (8-yr crude)||13% (8-yr crude)|
The results of the current series are consistent with those of the results reported by Moran and Haffty, indicating that the presence of LCIS does not appear to increase the risk of distant failure in patients treated with CS. Thus, the presence of LCIS by itself does not appear to change recommendations for systemic adjuvant therapy. However, the recently published results of the NSABP P-1 trial (which demonstrated that tamoxifen reduces the incidence of breast carcinoma in women at increased risk for development of the disease),21 taken together with the results of NSABP B-24 (which demonstrated the effect of adjuvant tamoxifen in patients treated for DCIS),22 may cautiously be interpreted as suggesting that the risk of contralateral breast carcinoma may be reduced for patients with LCIS by the use of adjuvant tamoxifen.
We recommend that potential candidates for breast conservation be evaluated with high quality film-screen mammograms to determine the presence and extent of microcalcifications, and with specimen radiographs in cases in which mammographic localization is required for biopsy. The margins of the specimen should be evaluated with ink. The pathology report should attempt to evaluate the presence, extent, and location of both ductal and lobular components. If LCIS is present without coexisting DCIS, treatment with radiation therapy may proceed without further surgery. If the LCIS is found in conjunction with DCIS, we recommend proceeding with the determination for reexcision based on the extent and margin status of the DCIS.
This study must be interpreted with several caveats. Of primary importance, all patients received radiation therapy to a dose of 60 Gy or more. The risk of recurrence in patients with LCIS treated with CS alone was therefore not tested here. Furthermore, the number of patients in this study with pure invasive ductal carcinoma and associated LCIS was small. A positive association therefore cannot be completely ruled out, although it appears unlikely that this would be clinically significant. In a similar fashion, most patients with ILC were found to have LCIS present; therefore, we would be unable to prove a negative effect of the absence of LCIS on the risk of local recurrence within this subgroup.
In conclusion, we have shown that LCIS in conjunction with infiltrating cancer is not a contraindication to the use of CS and RT. Women with LCIS should be treated in accordance with accepted standards for breast conservation, and the presence of the LCIS may be safely ignored.