Fax: (713) 745-1462
Selective use of sentinel lymph node surgery during prophylactic mastectomy
Article first published online: 5 SEP 2006
Copyright © 2006 American Cancer Society
Volume 107, Issue 7, pages 1440–1447, 1 October 2006
How to Cite
Boughey, J. C., Khakpour, N., Meric-Bernstam, F., Ross, M. I., Kuerer, H. M., Singletary, S. E., Babiera, G. V., Arun, B., Hunt, K. K. and Bedrosian, I. (2006), Selective use of sentinel lymph node surgery during prophylactic mastectomy. Cancer, 107: 1440–1447. doi: 10.1002/cncr.22176
- Issue published online: 18 SEP 2006
- Article first published online: 5 SEP 2006
- Manuscript Accepted: 21 JUN 2006
- Manuscript Revised: 13 JUN 2006
- Manuscript Received: 3 APR 2006
- prophylactic mastectomy;
- sentinel lymph node surgery;
- breast cancer;
Patients with invasive cancer identified at the time of prophylactic mastectomy (PM) will require axillary lymph node dissection for staging; therefore, many surgeons advocate sentinel lymph node (SLN) surgery at the time of PM. The current study investigates the invasive cancer rate in PM and evaluates factors associated with invasive cancer to guide SLN surgery use.
Patients undergoing PM at the M. D. Anderson Cancer Center between January 2000 and July 2005 were identified from a prospective database. Clinical, radiographic, and pathologic data were collected.
A total of 409 patients (436 PM cases) were identified; 382 underwent contralateral PM (CPM) and 27 underwent bilateral PM (BPM). Cancer was identified in 22 of 436 PM cases (5%). Of these, 14 patients (64%) had ductal carcinoma in situ (DCIS). Only 8 patients (1.8%) had invasive cancer, with a mean tumor size of 5 mm (range, 2–9 mm). There was no difference in the occult cancer rate between CPM and BPM. No cases of invasive cancer were identified in the 23 patients with BRCA mutations. Significantly increased risk of invasive cancer in the PM breast was seen in postmenopausal patients (3.7%; P = .007), patients age >60 years (7.5%; P = .008), and patients with history of invasive lobular carcinoma (9.7%; P = .0002) or lobular carcinoma in situ (LCIS) (7.7%; P = .008).
The frequency of cancer in PM is very low and the majority represents DCIS. Therefore, routine use of SLN surgery in all patients undergoing PM is not warranted. However, patients at higher risk for whom SLN surgery should be considered include older women and patients with a history of lobular cancer or LCIS. Cancer 2006. © 2006 American Cancer Society.
Increased interest in prophylactic mastectomy (PM) has been generated from strong public interest and awareness of breast cancer, the increased incidence of breast cancer, and the availability of genetic testing. PM in patients with deleterious BRCA1 or BRCA2 mutations, high risk family history, or personal history of breast cancer has shown efficacy in reducing the incidence of breast cancer.1–6 In addition, using decision analysis modeling Schrag et al.7 reported that PM provides substantial gains in life expectancy in young women who are BRCA1 or BRCA2 mutation carriers.
The utility of sentinel lymph node (SLN) surgery during PM remains controversial. SLN surgery is the standard of care for patients with early-stage breast cancer and its role has been extended to those patients with larger primary tumors,8, 9 those with previous breast surgery,10, 11 and those who have undergone neoadjuvant chemotherapy.12–14 The argument favoring routine SLN surgery in conjunction with PM is centered around the inability to perform SLN surgery once a mastectomy has been performed and therefore committing those patients found to have invasive cancer to axillary lymph node dissection (ALND) for lymph node staging. SLN surgery is seen as a minimally morbid procedure that would obviate the greater morbidity of ALND in lymph node-negative patients with occult cancer.
The predominant argument against the routine use of SLN surgery during PM is the risk associated with SLN surgery. The complications reported recently from several randomized trials of SLN surgery include a lymphedema rate of 7%, decreased upper arm range of motion in 3% of cases, and axillary paresthesias in 7% of patients.15–17 There is a small risk of an allergic reaction if isosulfan blue dye is used.18, 19 Additional considerations include the cost of routine SLN surgery as well as the biologic significance of cytokeratin-positive SLN without the finding of cancer in the breast that may contribute toward unnecessary patient anxiety.
The rationale for this study was to evaluate whether SLN surgery should be routinely performed in patients undergoing PM. The goals of the current study were to investigate the frequency of invasive cancer in PM specimens and to evaluate factors associated with invasive cancer to guide the use of SLN surgery at the time of PM.
MATERIALS AND METHODS
Patient Selection and Data Collection
All patients undergoing PM at the University of Texas M. D. Anderson Cancer Center (MDACC) between January 2001 and July 2005 were identified from the prospective breast surgery database. Inclusion criteria included: 1) patients undergoing bilateral prophylactic mastectomy (BPM) and 2) patients with a personal history of breast cancer undergoing contralateral prophylactic mastectomy (CPM). A total of 409 patients (436 PM cases) were included in this study, which was approved by the Institutional Review Board.
Data collection included information on patient characteristics, indication for PM, risk factors for breast carcinoma, histologic findings, and patient outcome. Data were collected retrospectively by reviewing patients' medical records.
Preoperative Patient Evaluation
All patients underwent clinical breast examination and mammography of the breast before surgery. Preoperative consultation with the plastic and reconstructive surgeons was offered to all patients. Reconstruction was performed with tissue expanders or autologous tissue flaps as appropriate. Patients undergoing CPM frequently had their surgery staged so the CPM occurred at the time of definitive reconstruction of the index breast (after chemotherapy and chest wall radiation if necessary), so that symmetry could be obtained.
Total mastectomy was performed in all cases. For patients undergoing immediate breast reconstruction a skin-sparing approach was taken, using a circular incision around the nipple-areolar complex, with a lateral extension if needed for additional exposure.
Axillary lymph node staging, if performed, involved low axillary lymph node sampling with the total mastectomy specimen or SLN surgery. Patients undergoing SLN surgery at the time of PM were mapped using either technetium-99 sulfur colloid alone or combined with isosulfan blue dye. Injection of technetium 99 sulfur colloid and/or blue dye was in the subareolar location.
All PM surgery and pathological review occurred at MDACC. Patients with index tumors in their contralateral breast who underwent their initial surgery elsewhere had their tumor histology reviewed by pathologists at our institution.
All PM specimens were processed routinely with pathologic evaluation of the nipple as well as 2 sections of each quadrant of the breast. Specimens were also sliced macroscopically and any areas of abnormality by palpation were evaluated further at the pathologists' discretion. The SLNs were evaluated by serial sectioning for hematoxylin and eosin and immunohistochemical staining with cytokeratin antibody.20
Univariate analysis was performed using the Student t-test, Fisher exact test, and chi-square analysis. The Wilcoxon rank sum test and Kruskal-Wallis test were used to compare continuous variables across clinical factors. The Fisher exact test was used to assess the association between 2 factors. All statistical tests were 2-sided with a significance level of <.05.
In all, 409 patients underwent PM, providing 436 cases. Twenty-seven patients underwent BPM and 382 patients underwent CPM with a documented cancer (either ductal carcinoma in situ [DCIS] or invasive cancer) in the index breast. Patient characteristics are summarized in Table 1. The majority of patients were female, with 1 male BRCA2 mutation carrier who underwent CPM. Fifty-seven percent had a family history of breast cancer (defined as a history of breast cancer in any family member) and 29% had at least 1 first-degree relative with breast cancer. Seventy-three percent of all patients underwent skin-sparing mastectomy.
|Variable||No. of patients (n = 409)(%)|
|Mean age at PM (range), y||50 (23–86)|
|N/A (male patient)||1 (0.2)|
|Contralateral PM||382 (93.3)|
|Bilateral PM||27 (6.6)|
|Any breast cancer||233 (57.0)|
|First-degree family history of breast cancer||120 (29.3)|
|Any ovarian cancer||22 (5.4)|
|BRCA not tested||360 (88.0)|
The majority of patients had a personal history of breast cancer and had chosen to undergo CPM because of their increased risk for developing contralateral breast cancer. The risk factors for cancer in the patients undergoing BPM are shown in Table 2. Forty-three percent of BPM patients with a previous breast biopsy had a history of LCIS. Seven of the 27 patients in this group had undergone testing for BRCA1 and BRCA2 and 6 patients (22%) had tested positive for a deleterious mutation. The average Gail lifetime risk was 31% for all BPM patients (range, 14–53%), with 78% of these patients having a Gail lifetime risk for development of breast cancer of >20%.
|Risk factor||No. of patients (%)|
|Previous breast biopsy||21 (78)|
|• Atypia on previous biopsy||14 (67)|
|• No atypia on previous biopsy||7 (33)|
|• Hx of LCIS on biopsy||9 (43)|
|• No LCIS on biopsy||12 (57)|
|No previous breast biopsy||6 (22)|
|BRCA not tested||20 (74)|
|2+ relatives with breast cancer||19 (70)|
|1 relative with breast cancer||7 (26)|
|No family history||1 (4)|
|Gail lifetime risk >20%||21 (78)|
|Gail lifetime risk ≤20%||6 (22)|
Histologic findings in the PM specimen were classified as benign (fibroadenoma, fibrocystic disease, sclerosing adenosis, and benign breast tissue without atypia), atypia (atypical lobular hyperplasia [ALH], atypical ductal hyperplasia [ADH], and lobular carcinoma in situ [LCIS]) or occult cancer. Table 3 shows the distribution of histologic findings. Occult cancer was detected in 5% of PM specimens. The majority of occult cancer (14 cases; 64%) identified was DCIS. The invasive cancers consisted of 6 cases of invasive lobular carcinoma (ILC) and 2 cases of invasive ductal carcinoma (IDC). The overall rate of occult invasive cancer was 1.8% (8 of 436 specimens). All the invasive tumors were classified as T1, ranging in size from 2 mm to 9 mm, with an average size of 5 mm.
|Histologic findings||No. of PM specimens (%)|
|Occult cancer||22 (5)|
|• DCIS||14 (3.2)|
|• IDC||2 (0.5)|
|• ILC||6 (1.4)|
When comparing the BPM and CPM patients, there was no significant difference noted in the frequency of occult cancer or frequency of invasive cancer. Invasive cancer was found in 1 of the 54 PM specimens of patients undergoing BPM (1.9%) and in 7 of the 382 CPM specimens (1.8%), as shown in Table 4.
|CPM (n = 382)||BPM (n = 54)||All cases (n = 436)|
|All cancer in PM||19 (5.0%)||3 (5.6%)||22 (5.0%)|
|DCIS in PM||12 (3.1%)||2 (3.7%)||14 (3.2%)|
|Invasive cancer in PM||7 (1.8%)||1 (1.9%)||8 (1.8%)|
Forty-nine of the 409 patients underwent testing for BRCA mutations and 23 patients (5.6%) were identified to have a deleterious mutation of BRCA1 or BRCA2. The average age of these 23 patients was 44 years (range, 25–57 years). Two of these patients (8.7%) had DCIS in the PM specimen (data not shown). No cases of invasive cancer were found in this subgroup of patients.
Because a substantial proportion of our CPM patient population received chemotherapy for their index breast tumor before PM, we evaluated whether this impacted the frequency of invasive cancer in the PM specimen. A total of 167 patients received chemotherapy for treatment of their index breast cancer (109 in the neoadjuvant setting and 58 as adjuvant therapy before the time of PM). The median time interval from the initiation of chemotherapy to the time of PM was 8 months (range, 3 months to 15 years). The finding of cancer was similar irrespective of chemotherapy use. Occult invasive cancer was found in 1.8% of patients who received preoperative chemotherapy (1.8%) compared with 1.9% of those who did not receive preoperative chemotherapy (P = 1).
Lymph Node Staging of Patients with Invasive Cancer
Four of the 8 patients with invasive cancer identified in the PM specimen underwent regional lymph node staging, of whom only 1 had confirmed lymph node-positive disease. One patient had undergone SLN surgery at the time of PM and was lymph node negative. Two patients had undergone low axillary lymph node sampling at the time of PM and in each case 6 lymph nodes were removed. In 1 patient, all the lymph nodes were negative. In the other patient, 2 of the 6 lymph nodes were positive for metastatic disease and this patient returned to the operating room for completion of ALND. The fourth patient did not undergo any lymph node staging at the time of primary breast surgery. After pathology revealed invasive cancer, she returned to the operating room for formal ALND and was determined to be pathologically lymph node negative.
Of these 8 patients, 7 are alive without evidence of disease, with a median follow-up of 3.8 years (range, 0.4–5.6 years). The single patient with lymph node involvement (AJCC Stage IIIA index breast ILC; Stage IIA ILC in the PM) was alive with intraabdominal metastases at the time of last follow-up.
Results of SLN Surgery
One hundred eight patients underwent SLN surgery at the time of PM during this study period. The decision to perform SLN surgery was at the discretion of the individual operating surgeon. The SLN was identified in 106 patients (98%). Patients in whom no SLN was identified did not undergo additional axillary lymph node surgery. As mentioned above, 1 patient who was found to have an occult invasive tumor (2 mm) had a negative SLN.
Two patients without an occult cancer in the PM breast had either metastasis or a single immunohistochemistry (IHC)-positive cell cluster identified in the SLN. The first of these 2 patients was found to have a 9-mm metastatic deposit in the SLN. This patient had an inflammatory T4, N1 tumor in the contralateral breast and the positive SLN was believed to represent a cross-metastases. She did not undergo further ALND and died of her index breast cancer 6 weeks after surgery. The second patient underwent BPM and was found to have a single keratin-positive cell cluster containing only a few cells by IHC staining. She did not undergo any additional surgery or therapy and was alive and free of disease at the time of last follow-up 5 years later.
Factors Associated with Cancer in CPM
We next used our CPM cohort to evaluate a range of clinical (Table 5) and histologic (Table 6) variables that may be useful for identifying patients at higher risk for occult invasive cancer in the PM breast. We compared the variables in the 7 patients with invasive cancer with the 375 patients without invasive cancer (including DCIS). Older patients (age >60 years) at the time of initial breast cancer diagnosis were found to be at significantly higher risk of harboring occult invasive cancer (7.7% vs. 0.9%; P = .008). Patients who were postmenopausal also had a higher frequency of invasive cancer compared with premenopausal patients (3.7% vs. 0%; P = .007). Family history was not found to be a risk factor for occult invasive cancer in this study (P = 1).
|Patient variable||No invasive cancer in PM||Invasive cancer in PM||P|
|No. of patients||375||7 (1.8%)|
|Mean age, y||47.9||58.7||.007|
|• Birth-39||83||0 (0.0%)|
|• 40–49||128||1 (0.8%)|
|• 50–59||116||2 (1.7%)||.01|
|• 60–69||35||4 (10.3%)|
|• 70+||13||0 (0.0%)|
|• <60 y||327||3 (0.9%)||.008|
|• ≥60 y||48||4 (7.7%)|
|Relatives with breast cancer|
|• 0||169||3 (1.7%)|
|• 1||111||2 (1.8%)|
|• 2||58||2 (3.3%)|
|• 3||18||0 (0%)||1|
|• 4||8||0 (0%)|
|• 5||8||0 (0%)|
|• 6+||3||0 (0%)|
|• Premenopausal||192||0 (0.0%)|
|• Postmenopausal||182||7 (3.7%)||.007|
|• N/A (male)||1||0|
|Histologic variable in index tumor specimen||No invasive cancer in PM (n = 375)||Invasive cancer in PM (n = 7)||P|
|• Ductal||255||1 (0.4%)||.0003*|
|• Lobular||56||6 (9.7)|
|• DCIS||63||0 (0%)|
|• Sarcoma||1||0 (0%)|
|• Yes||20||0 (0%)||1|
|• No||355||7 (1.9%)|
|• Yes||31||2 (6.1%)||.12|
|• No||344||5 (1.4%)|
|• Yes||48||4 (7.7%)||.008|
|• No||327||3 (0.9%)|
|• Negative||99||0 (0%)|
|• Positive||210||7 (3.2%)||.1|
|• Unknown||66||0 (0%)|
|• Negative||124||0 (0%)|
|• Positive||185||7 (3.6%)||.045|
|• Unknown||66||0 (0%)|
|• Negative||227||7 (3.0%)|
|• Positive||44||0 (0%)||.6|
|• Unknown||104||0 (0%)|
Histological predictors for invasive cancer in the PM specimen are shown in Table 6. The characteristics of the tumor in the index breast formed the basis of this analysis. The pathology of the index tumor was significantly associated with rate of invasive cancer in the PM breast; patients with lobular cancer (pure ILC or mixed ILC/IDC) had a strong probability of occult contralateral disease compared with pure IDC patients (9.7% vs. 0.4%; P = .0003). The presence of LCIS associated with the index tumor was also found to be significantly associated with invasive cancer in the PM specimen. The frequency of invasive cancer in the PM specimen of patients with LCIS was 7.7% compared with 0.9% in patients without LCIS in the index breast (P = .008). Other forms of atypia (ADH, ALH) in the index breast were not found to be predictive for invasive cancer in the PM specimen.
All 7 patients with invasive cancer in the PM had index tumors that were positive for estrogen receptor (ER) and progesterone receptor (PR) and negative for HER-2/neu. Only patients with positive PR status were found to be at slightly increased risk of invasive cancer in CPM (3.6%; P = .045). However, ER, PR, and HER-2/neu status was not available for a significant number of the patients, making the association with cancer less reliable. Tumor size, lymph node stage, and overall stage of the index cancer did not predict for invasive cancer in the PM (data not shown).
The frequency of invasive cancer based on the significant clinical and pathologic predictors identified above is summarized in Figure 1.
Correlation of Factors Associated with Cancer
Multivariate analysis of the individual factors was not feasible due to the overall small number of cases of invasive cancer in the PM. However, the degree of correlation between the pairs of factors identified above was evaluated using the Wilcoxon rank sum test, Kruskal-Wallis test, and Fisher exact test. Menopausal status was significantly associated with age as well as with the pathology of the index tumor. Seventy-four percent of premenopausal patients had IDC as their index tumor and this percentage decreased to 60% in postmenopausal patients. Conversely, postmenopausal patients had a higher percentage of ILC (17%) compared with premenopausal patients (9%). Menopausal status was also significantly associated with finding LCIS in the index breast. Eight percent of premenopausal patients had LCIS in the index breast compared with a 19% rate of LCIS in postmenopausal women. Patients with ILC had LCIS present in 57% of cases, whereas in IDC patients LCIS was observed in only 4% of cases. Overall age, menopausal status, pathology of the index tumor, and LCIS status were all found to be significantly correlated with each other (P < .05).
Prophylactic mastectomy has been utilized for many years as a risk reduction strategy in women at increased risk of breast cancer. Recently, it has received increased interest with the identification of high-risk patients and the isolation of genetic mutations that predispose to breast cancer development. Lymph node staging has previously not been part of the surgical procedure for PM. With the increasing use of SLN surgery, some surgeons have advocated its use in patients undergoing PM because SLN surgery is viewed as a relatively straightforward procedure. However, recent published trials have documented measurable complications including lymphedema, axillary paresthesias, and decreased range of arm motion.15–17 Therefore, the adverse effects of SLN surgery appear to be higher than the reported rates of occult cancer after PM.1, 5, 6, 21, 22 The utility of SLN surgery during PM therefore remains controversial.
The addition of more sensitive imaging modalities, such as magnetic resonance imaging (MRI), could be considered a way to improve patient selection for SLN in PM. However, recently reported data specifically assessing the role of MRI to evaluate for invasive breast cancer in patients undergoing PM indicated that MRI missed the majority of tumors, and therefore MRI in PM does not appear to be either valuable nor cost-effective.23 We sought to identify subgroups of patients based on clinical and histologic variables whose risk for invasive cancer would support the use of SLN staging during PM.
To our knowledge there are only few published articles on the frequency of cancer in patients undergoing PM; therefore, an accurate assessment of the frequency of invasive disease is difficult. Furthermore, although the occurrence of occult cancer is reported to range from 0.5% to 8.0%,1, 5, 6, 21, 22, 24 more relevant are findings of invasive tumor that would warrant lymph node staging. Among the handful of studies that report on the rate of invasive disease, the frequency ranges from 0.1% to 3.5%. The highest rate of 3.5% was noted in a small study of 57 patients22; the small sample size may reflect substantial statistical variability. In contrast, although Hartmann et al.1 collected a cohort of 687 patients, that study spanned from 1960 to 1993 and pathologic assessment has evolved over the years. Therefore, this finding may account for the very low occurrence (0.1%) of occult cancer in this series. The current study provides what to our knowledge is 1 of the largest recent cohorts of PM patients evaluated for occult invasive cancer.
We found that overall the rate of occult invasive cancer in PM patients is 1.8% and that the rate is the same in both BPM and CPM subgroups. Chemotherapy before PM did not appear to alter the likelihood of finding invasive cancer. We found factors associated with an increased occurrence of invasive cancer in the PM breast included age at initial breast cancer diagnosis and postmenopausal status, as well as the histology and PR status of the index breast tumor. Although these are all important risk factors, because of the small number of events, we were unable to assign a hazard value to each factor to develop a predictive model.
Increased age was associated with an increased risk of invasive cancer at the time of PM. This is not surprising, because age is a known risk factor for breast carcinoma.25 Patients with LCIS are known to have an increased risk of breast cancer in the contralateral as well as ipsilateral breast over their lifetime, and therefore our finding that patients with LCIS in 1 breast have increased risk of finding invasive disease at the time of PM is consistent with the literature reports. Our finding is further supported by studies evaluating contralateral breast cancer rates in patients with DCIS that show that the presence of atypia (ALH, ADH, or LCIS) along with DCIS puts patients at higher risk of contralateral breast cancer.26 However, in contrast to the report by Adepoju et al.,26 we found only LCIS and not other forms of atypia (ADH, ALH) to be associated with contralateral invasive cancer.
Similarly, lobular histology is generally accepted as a risk factor for contralateral breast cancer.27 ILC has been shown to be more likely in older patients, to be ER-positive and PR-positive and HER-2-negative, and to have a higher rate of contralateral breast cancer (21% compared with 11% in IDC).27 Consistent with these reported characteristics of ILC, in this study we in addition found that ILC, either pure or mixed, placed patients at a 10% risk of detecting occult invasive cancer after their CPM. This is a 5-fold higher risk than the average patient undergoing PM.
Interestingly, patients in our cohort known to have deleterious mutations in BRCA1 or BRCA2 were not at increased risk for invasive cancer in the PM. To our knowledge, no study to date has specifically addressed the occurrence of invasive cancer at the time of PM in patients with deleterious BRCA mutations. However, recent studies evaluating the efficacy of PM as a risk reduction intervention for BRCA1 and BRCA2 patients also fail to report any cases of occult invasive cancer.2–4 The mean age at the time of PM in these series ranged from 36 years to 38 years. Although our BRCA patients were older (mean age, 44 years), our findings are concordant with these earlier reports. The occurrence of occult invasive cancer in BRCA patients may rise with increasing age at time of PM. In addition, the age at which invasive cancer is detected may also depend on the specific type of deleterious mutation. Further research into these areas is needed to develop rational guidelines for SLN surgery during PM in BRCA patients.
Our data also demonstrate that routine SLN surgery in PM patients may lead to findings that may not be clinically significant. Of the 106 patients who underwent successful SLN biopsy at the time of PM, we found only 2 patients with lymph node disease, 1 of which represented cross-metastasis from a locally advanced contralateral tumor. The other patient had a single IHC-positive tumor cell cluster in the SLN and underwent no additional therapy and remains free of disease after 5 years of follow-up. These findings are similar to other reports in the literature of SLN surgery in PM patients. King et al.21 found 1 patient with IHC + SLN. This patient was determined to have a cross-metastasis from a contralateral T2N3 poorly differentiated breast cancer and underwent no additional therapy. Similarly, Dupont et al.22 found 2 out of 57 patients in their study group with IHC-positive SLNs and no cancer in the PM; both patients underwent ALND with no tumor was noted in the additional lymph nodes.
In summary, patients undergoing PM should be assessed for lymph node surgery in terms of their risk for invasive cancer. Our data suggests that older patients (age >60 years), those with a history of LCIS, or those with history of ILC should be considered for SLN surgery because the risk of invasive cancer in the PM breast in this group of patients is substantially increased. The routine use of SLN surgery in all patients undergoing PM is not recommended.
We thank Auston Wei and Gregory Ayers of the Department of Biostatistics and Applied Mathematics for their assistance
- 16Post-operative arm morbidity and quality of life. Results of the ALMANAC randomised trial comparing sentinel node biopsy with standard axillary treatment in the management of patients with early breast cancer. Breast Cancer Res Treat. 2006; 95: 279–293., , , et al.
- 17Surgical complications associated with sentinel lymph node dissection (SLND) plus axillary lymph node dissection, versus SLND alone in the American College of Surgeons Oncology Group (ACOSOG) Trial Z0011. Ann Surg Oncol. 2006; 13( 2 Suppl): 4., , , et al.
- 23Detecting occult breast cancer in prophylactic mastectomies: sentinel node biopsy vs. Pre-operative MRI. Ann Surg Oncol. 2006; 13( 2 Suppl): 9., , , et al.
- 26Impact of concurrent proliferative high-risk lesions on the risk of ipsilateral breast carcinoma recurrence and contralateral breast carcinoma development in patients with ductal carcinoma in situ treated with breast-conserving therapy. Cancer. 2006; 106: 42–50., , , et al.