Survival outcomes of patients with lobular carcinoma in situ who underwent bilateral mastectomy or partial mastectomy

European Journal of Cancer 82 (2017) 6e15

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.ejcancer.com

Original Research

Survival outcomes of patients with lobular carcinoma in situ who underwent bilateral mastectomy or partial mastectomy Ze-Ming Xie a,1, Jian Sun b,1, Zhe-Yu Hu c,d,1, Yao-Pan Wu e, Peng Liu a, Jun Tang a,*, Xiang-Sheng Xiao a, Wei-Dong Wei a, Xi Wang a, Xiao-Ming Xie a, Ming-Tian Yang a a Department of Breast Oncology, Sun Yat-Sen University Cancer Center, 651 Dong Feng East Road, Guangzhou, 510060, PR China b Department of Clinical Research, Sun Yat-Sen University Cancer Center, 651 Dong Feng East Road, Guangzhou, 510060, PR China c Department of Clinical Research, First Hospital of Changsha City, No. 311, Yinpan Road, Changsha, 410005, PR China d Department of Biostatistics and Bioinformatics, Rollins School of Public Health, Emory University, 1518 Clifton Road, Atlanta, GA, 30322, United States e Department of Radiology, Sun Yat-Sen University Cancer Center, 651 Dong Feng East Road, Guangzhou, 510060, PR China

Received 16 January 2017; received in revised form 19 April 2017; accepted 17 May 2017

KEYWORDS Lobular carcinoma in situ; Prophylactic mastectomy; Partial mastectomy; Survival

Abstract Aim: To compare the survival outcomes between patients treated with bilateral mastectomy and partial mastectomy alone as the initial surgical management for primary lobular carcinoma in situ (LCIS). Patients and methods: Patients with histologically confirmed LCIS underwent partial mastectomy alone or bilateral mastectomy were identified by the SEER*Stat database (version 8.3.2) released in 2016. The primary outcome measure was all-cause mortality and the secondary outcome measure was breast cancer-specific mortality. Results: Of the 5964 cases included in the analysis, 208 cases underwent bilateral mastectomy and 5756 cases underwent partial mastectomy alone. The 1-, 5- and 10-year estimated overall survival rates were 99.7%, 96.7% and 91.7%, respectively. Univariate and multivariate proportional hazards regression (Cox) analyses showed no significant difference between the risk of all-cause mortality in the bilateral mastectomy group compared with the partial mastectomy

* Corresponding author: Department of Breast Oncology, Sun Yat-Sen University Cancer Center, 651 Dong Feng East Road, Guangzhou, 510060, PR China. Fax: þ86 20 87343066. E-mail address: [email protected] (J. Tang). 1 Authors contributed equally to this work. http://dx.doi.org/10.1016/j.ejca.2017.05.030 0959-8049/ª 2017 Elsevier Ltd. All rights reserved.

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

7

group (HR Z 1.106, 95% confidence interval [CI] 0.350e3.500, P Z 0.86). In propensity score-matched model, bilateral mastectomy still did not show benefit to overall mortality (HR Z 2.248, 95% CI 0.451e11.200). Patients older than 60 years of age showed a higher risk of all-cause mortality (HR Z 7.593, 95% CI 5.357e10.764, P < 0.0001). No risk factors, including surgery type, were identified for breast cancer-specific survival. Conclusions: Survival outcomes of patients with LCIS who underwent partial mastectomy without radiotherapy were not inferior to patients who underwent bilateral prophylactic mastectomy. Breast cancer-specific mortality in patients with LCIS was extremely low; aggressive prophylactic surgery like bilateral prophylactic mastectomy should not be advocated for most patients with LCIS. ª 2017 Elsevier Ltd. All rights reserved.

1. Introduction Lobular carcinoma in situ (LCIS) has long been considered a precursor or risk factor for the subsequent development of invasive breast cancer. Previous studies have documented that the relative risk of developing subsequent invasive breast cancer is approximately twofold higher in LCIS patients than in individuals without LCIS, and the absolute risk is approximately 1% per year [1e6]. A series of 4853 cases of patients diagnosed with LCIS that were registered to the Surveillance, Epidemiology and End Results (SEER) program between 1973 and 1998 found that the risk of developing invasive breast cancer after LCIS was 7.1% at 10 years, with equal predisposition to both breasts [6]. The optimal initial management of LCIS varies among breast surgeons [7]. For patients with classic LCIS without concomitant invasive disease or ductal carcinoma in situ (DCIS), current treatment guideline does not recommend further surgery for treatment purpose but directs it to risk-reduction consultation [8]. Risk-reduction bilateral total mastectomy remains a consensus by the NCCN Breast Cancer Risk Reduction Panel as an option for women with LCIS even without other risk factors [9], although it is not a recommended approach for most of these patients, and it is unclear whether this prophylactic approach results in improved survival outcomes. Due to the low incidence and mortality rate of LCIS, the overall survival outcome after treatment is seldom discussed in LCIS studies. In a National Surgical Adjuvant Breast and Bowel Project (NSABP) study of 180 LCIS patients, only one patient died of recurrence of previous ipsilateral invasive carcinoma and one patient died of subsequent contralateral breast cancer, with a breast cancer-specific mortality rate of 1.1% [5]. In a meta-analysis of 389 LCIS patients, the breast cancerspecific mortality rate among women who had a local excision was 2.8%, which was not significantly different from the disease-specific mortality rate in women who were initially treated with mastectomy [10]. However, due to the small sample size and heterogeneity of the data, these studies were not powerful enough to illustrate the effect of different local management strategies on breast

cancer-specific survival after a diagnosis of primary LCIS (i.e. LCIS as the first tumour in a patient). The SEER program of the National Cancer Institute (NCI) is a population-based cancer registry covering approximately 30% of the population in the United States. This database is the largest publicly available and authoritative information source on cancer incidence and survival. Using this reliable and large-scale research data set, we were able to statistically analyse the survival outcomes for patients with LCIS. The objective of this study was to evaluate the difference between bilateral mastectomy and partial mastectomy on the overall survival of patients diagnosed with primary LCIS using the case information extracted from the SEER research database. 2. Patients and methods The SEER*Stat database, which was released by the Surveillance Research Program at the NCI in 2016, was used as the data source in the present study [11]. Women diagnosed with LCIS (ICD-O-3 Histology code 8520:2) were identified in the SEER 13 Regs Research Data þ Hurricane Katrina Impacted Louisiana Cases, Nov 2013 Sub (1992e2011 varying) incidence database. We chose the SEER 13 Regs research database because it also contains information of multiple primaries. Therefore, we are able to exclude concurrent malignancy (defined as in situ or invasive cancer diagnosed within 6 months of LCIS diagnosis) by implementing the MPSIR session of the SEER*Stat software. Because the database does not include detailed information on local surgery treatment before 1998, we only included histologically confirmed cases diagnosed between January 1998 and December 2011. Patients with a prior history of any type of in situ or invasive cancer, or patients who were only initially identified by autopsy, death certificate or cytology, were excluded from the study. Based on information regarding surgery and radiotherapy treatment, the patients were categorised into three groups: partial mastectomy only (including partial mastectomy with nipple resection, lumpectomy or excisional biopsy, re-excision of the biopsy site for gross or

8

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

microscopic residual disease, wedge resection, quadrantectomy, and tylectomy), bilateral mastectomy (with or without reconstruction) and others (including partial mastectomy with radiotherapy, simple mastectomy, etc.). Because partial mastectomy with radiotherapy and simple mastectomy were seldom adopted as a treatment option for LCIS, the sample size will be too small to perform survival outcome analysis, only patients who underwent partial mastectomy without radiotherapy or bilateral mastectomy were included in the analysis. Patients who were diagnosed with a core needle biopsy only or refused surgery (446 cases) were not included in the study because the possibility of adjacent ductal carcinoma or invasive components cannot be excluded at diagnosis (Supplementary Fig. 1). Patients were followed up until December 2011. The primary outcome measure was all-cause mortality, and the secondary outcome measure was breast cancerspecific mortality. The candidate risk factors for overall survival included surgery type, age, race, tumour size, oestrogen receptor (ER) and progesterone receptor (PR) status, differential grade and laterality. Numeric variables were summarised as the mean (standard deviation) or median (interquartile range), where appropriate. Categorical variables were reported as counts (percentage). An analysis of variance was used to compare continuous variables (age) with symmetric distributions between the treatment types. Chi-square tests or Fisher’s exact tests (n < 5) were used to compare categorical variables between the treatment groups. The KaplanMeier method was used to plot the survival distributions, and the logrank test was used to assess the differences in survival experience between the groups. The Cox proportional hazards regression was performed to estimate the hazard ratio to identify the risk factors for breast cancer-specific mortality and all-cause mortality. To further adjust for potential baseline confounders, a propensity score matching was carried out. A receiver operating characteristic (ROC) curve was also calculated to determine the optimal cutoff that maximises sensitivity and specificity in predicting mortality. All tests of hypotheses were twotailed and conducted at a significance level of 0.05. Statistical analyses were conducted using SAS, version 9.4. The authenticity of this article had been validated by uploading the key raw data onto the Research Data Deposit public platform (www.researchdata.org.cn), with the approval RDD number as RDDA2017000137. 3. Results 3.1. Demographic and clinical characteristics of the LCIS patients who underwent bilateral mastectomy and partial mastectomy Of the 5964 cases included in the analysis, 208 cases underwent bilateral mastectomy and 5756 cases underwent partial mastectomy without radiotherapy. The

mean (SD) age of the patients who underwent a bilateral mastectomy was less than that of the patients who underwent a partial mastectomy (48.9  7.9 versus 54.1  10.6, P < 0.0001). ROC curve analysis determined that age 59 was the optimal cutoff value that maximised sensitivity and specificity for predicting both all-cause mortality and breast cancer-specific mortality (Supplementary Figs. 2 and 3, available online). The distribution of age was significantly different between the two treatment groups when stratified based on age <60 (59) years and 60 years (P < 0.0001). In the bilateral mastectomy group, 187 (89.9%) of the patients were <60 years old, while 21 (10.1%) patients were 60 years old. In the partial mastectomy group, 4247 (73.78%) of the patients were <60 years old, while 1509 (26.22%) were 60 years old (Table 1). Racial disparity existed between the two treatment groups (P Z 0.01). Among the bilateral mastectomy patients, 190 (91.35%) patients were of white race, six (2.88%) were of black race and 12 (5.77%) were of other races. However, in the partial mastectomy group, 4869 (84.59%) patients were white, 471 (8.18%) were black and 416 (7.23%) were of other races (Table 1). Factors including tumour size category (<20 mm or >20 mm) and breast cancer laterality (left or right), were similarly distributed between the two treatment groups (Table 1). The records for ER status, PR status and differential grade were missing in 80e90% of the patients. Records of tumour size were missing in more than 50% of the patients. Due to rare breast cancer-specific mortality, propensity score matched data alone were used to compare all-cause mortality between the two treatment groups. In this study, we also performed multivariate analysis to control for tumour differential grade, race and age between the two treatment groups. 3.2. Survival analysis Among the 5964 patients, 321 all-cause deaths were observed. The 1-, 5- and 10-year estimated survival rates were 99.7%, 96.7% and 91.7%, respectively (Fig. 1A, Supplementary Table 1). A total of six deaths occurred in the bilateral mastectomy group (6 of 208, 2.88%), whereas 315 deaths occurred in the partial mastectomy group (315 of 5756, 5.47%). The 1-, 5- and 10-year estimated survival rates were 100%, 99.39% and 96.59%, respectively, in the bilateral mastectomy group, as compared with 99.71%, 96.60% and 91.55%, respectively, in the partial mastectomy group (Fig. 2A, logrank P-value Z 0.1151, Supplementary Table 2). A total of 24 patients died due to breast cancer. The 1-, 5- and 10-year breast cancer-specific estimated survival rates were 100%, 99.85% ad 99.21%, respectively (Fig. 1B, Supplementary Table 4). One death in the bilateral mastectomy group (0.48%) and 23 deaths in the partial mastectomy group (0.40%) were attributable to breast cancer. The 1-, 5- and 10-year breast cancer-

Table 1 Demographics for patients who underwent bilateral mastectomy and partial mastectomy (n Z 5964). Covariate

Level

Tumour size

Race

Grade

Laterality

*P-value Bilateral mastectomy (n Z 208)

No Yes No Yes

48.9  7.9, 47 (44.5, 52.5) 187 (89.90%) 21 (10.10%) 76.6  48.5, 72 (34, 118) 202 (97.12%) 6 (2.88%) 202 (99.51%) 1 (0.49%)

54.1  10.6, 52 (47, 60) <0.0001 4247 (73.78%) <0.0001 1509 (26.22%) 75.3  52.0, 65 (28, 117.5) 0.15 5441 (94.53%) 0.12 315 (5.47%) 5441 (99.58%) 0.58 23 (0.42%)

48.9  7.9, 47 (44.5, 52.5) 187 (89.90%) 21 (10.10%) 75.3  52.0, 65 (28, 117.5) 202 (97.12%) 6 (2.88%) 202 (99.51%) 1 (0.49%)

48.8  8.1, 47 (44, 52) 0.00 187 (89.90%) 21 (10.10%) 70.2  48.0, 66 (28, 106) 206 (99.04%) 2 (0.96%) 206 (100.00%) 0 (0.00%)

Positive Negative Unknown Positive Negative Unknown <20 mm 20 mm Unknown White Black Other Well differentiated Moderate Poor or undifferentiated Unknown Left Right

32 (15.38%) 0 (0%) 176 (84.62%) 24 (11.54%) 7 (3.37%) 177 (85.10) 77 (37.02%) 13 (6.25%) 118 (56.73%) 190 (91.35%) 6 (2.88%) 12 (5.77%) 15 (7.21%) 7 (3.37%) 6 (2.88%) 180 (86.54%) 102 (49.28%) 105 (50.72%)

527 (9.16%) 25 (0.43%) 5204 (90.41%) 447 (7.77%) 73 (1.27%) 5236 (90.97%) 2339 (40.64%) 278 (4.83%) 3139 (54.53%) 4869 (84.59%) 471 (8.18%) 416 (7.23%) 235 (4.08%) 243 (4.22%) 58 (1.01%) 5220 (90.69%) 2983 (51.88%) 2767 (48.12%)

30 (14.42%) 0 (0%) 178 (85.58%) 24 (11.54%) 7 (3.37%) 177 (85.10%) 77 (37.02%) 13 (6.25%) 118 (56.73%) 190 (91.35%) 6 (2.88%) 12 (5.77%) 15 (7.21%) 7 (3.37%) 6 (2.88%) 180 (86.54%) 102 (49.28%) 105 (50.72%)

21 (10.10%) 2 (0.96%) 185 (88.94%) 21 (10.10%) 2 (0.96%) 185 (88.94%) 84 (40.38%) 10 (4.81%) 114 (54.81%) 192 (92.31%) 6 (2.88%) 10 (4.81%) 17 (8.17%) 7 (3.37%) 3 (1.44%) 180 (86.54%) 100 (48.8%) 108 (51.92%)

0.01

0.006

0.40

0.01

0.007

0.46

Partial mastectomy (n Z 208)

Standard P-value difference 0.83 1.00 0.30 0.28 0.50

0.44

0.30

0.39

0.46

0.06

0.68

0.91 0.00 0.04

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

PR

Surgery type Partial mastectomy (n Z 5756)

<60 60

Breast cancer-specific death (excluding other death causes, n Z 5667) ER

Matched

Surgery type Bilateral mastectomy (n Z 208) Age at the time of surgery

Survival months All-cause mortality

Unmatched

0.80 0.10 0.01 0.00 0.00

0.81

9

10

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

Fig. 1. KaplaneMeier curve for all-cause mortality (A) and breast cancer-specific mortality (B).

specific estimated survival rates were all 100% in the bilateral mastectomy group, as compared with 100%, 99.84% and 99.18%, respectively, in the partial mastectomy group (Fig. 3, log-rank P-value Z 0.86, Supplementary Table 5). 3.3. Risk factors for all-cause mortality The univariate and multivariate analyses showed no significant differences between the bilateral mastectomy group and the partial mastectomy group in the risk of all-cause death (HR Z 0.528, 95% confidence interval [CI] 0.235e1.185, P Z 0.08, univariate Cox model; HR Z 1.106, 95% CI 0.350e3.500, P Z 0.86,

multivariate Cox model). However, patients who were 60 years old showed a higher risk of all-cause death (P < 0.0001) in both the univariate and multivariate Cox models. The other parameters investigated, including tumour size, tumour differential grade, ER status, PR status, race and laterality, showed no significant effect on all-cause mortality (Table 2). 3.4. Risk factors for breast cancer-specific mortality Univariate analysis showed that an age 60 years was the only risk factor for breast cancer-specific mortality (HR Z 3.378, 95% CI 1.518e7.521, P Z 0.003). However, multivariate analysis did not show a significant

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

11

Fig. 2. KaplaneMeier curves among patients stratified by surgery type for all-cause mortality (A): log-rank test P Z 0.12, unmatched; (B): log-rank test P Z 0.31, matched.

effect of age. The other factors investigated, including surgery type, hormone status, tumour differential grade, tumour size, race and laterality, had no significant effect on breast cancer-specific mortality (Table 3). 3.5. Adjusting for patient characteristics using propensity score matching To account for potential bias due to an imbalance between the bilateral and partial mastectomy groups

regarding patient age, race and differentiation grade, the propensity score matching was carried out. After matching procedure, the propensity score improved to near equality (Supplementary Fig. 4). In the Cox regression model after propensity score matching, bilateral mastectomy was not a protective factor for overall survival (HR Z 2.284, 95% CI Z 0.451e11.200, P Z 0.32, Table 2). Kaplan-Meier survival analysis also showed no significant difference between bilateral and partial mastectomy groups (log-rank test P Z 0.3101, Fig. 2B).

12

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

Fig. 3. KaplaneMeier curves among patients stratified by surgery type for breast cancer-specific mortality (log-rank test P Z 0.86). Table 2 Risk factors for overall survival post surgery: outcome is all-cause mortality. Risk factors

Level

Univariate Cox regression HR (95% CI)

Surgery type

Bilateral mastectomy Partial mastectomy

Age at surgery

ER PR Tumour size Race

Grade

Laterality

60 <60 Positive Negative Positive Negative <20 mm >20 mm White Black Other Well differentiated Moderate differentiated Poor or undifferentiated (Right versus left)

0.528 Ref 1.101 7.533 Ref 0.904 Ref 0.994 Ref Ref 1.311 Ref 0.867 0.441 Ref 1.455 3.736 0.843

a

Multivariate Cox regression a

b

P-value

HR (95% CI)

(0.235, 1.185)

0.08

(1.091, 1.111) (5.909, 9.604)

<0.0001 <0.0001

(0.119, 6.835)

0.92

1.106 (0.350, 3.500) Ref e 7.593 (5.357, 10.764) Ref e

(0.285, 3.462)

0.99

e

(0.812, 2.114)

0.27

(0.562, 1.338) (0.227, 0.855)

0.52 0.02

(0.542, 3.908) (1.355, 10.30) (0.676, 1.051)

0.46 0.01 0.13

Ref 1.314 Ref 0.626 0.561 Ref 1.851 2.607 0.927

Propensity score Cox regression b

P-value

HR (95% CI)c

P-valuec

0.86

2.248 (0.451, 11.200)

0.32

e <0.0001

1.074 (1.006, 1.147) 7.182 (1.694, 30.447)

0.03 0.008

Infinity (0, Infinity) Ref Infinity (0, Infinity) Ref Ref 15.563 (1.393, 173.856) Ref 4.543 (0.529, 39.016) 10.979 (1.977, 60.969) Ref 0.992 (0, Infinity) 1.026 (0, Infinity) 0.566 (0.135, 2.381)

1.00

(0.810, 2.132)

0.27

(0.307, 1.279) (0.229, 1.373)

0.20 0.21

(0.619, 5.537) (0.823, 8.257) (0.682, 1.260)

0.27 0.10 0.63

1.00

0.03 0.17 0.006 1.00 1.00 0.44

Univariate analysis (n Z 5964). Risk of surgery for overall survival post surgery adjusted by age level, size level, race, grade and laterality: outcome is all-cause mortality (n Z 2706). c Risk factor in propensity score matched data (n Z 416). a

b

4. Discussion The incidence of LCIS has increased over the last two decades, and a four-fold increase was observed from 1978 to 1998, particularly among postmenopausal women [12]. Most studies investigating LCIS prognosis have emphasised the incidence of subsequent invasive and/or in situ carcinoma, whereas survival outcomes have rarely been reported. The present study is the first population-based analysis to address the overall

survival and breast cancer-specific survival in patients with LCIS and to compare the survival efficacy between two major initial local management strategies for LCIS: partial mastectomy without radiotherapy and bilateral mastectomy. The majority of the patients (96.5%) in this study underwent a partial mastectomy without radiotherapy, and the 10-year overall survival probability in this cohort was 91.7%, with a breast cancer-specific mortality less than 0.5%, which reflects the indolent nature of

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

13

Table 3 Risk factors for overall survival post surgery: outcome is breast cancer-specific mortality. Risk factors

Level

HR (95% CI)a

P-valuea

HR (95% CI)b

P-valueb

Surgery type

Bilateral mastectomy Partial mastectomy 60 <60 Positive Negative Positive Negative <20 mm >20 mm White Black Other Well differentiated Moderate differentiated Poor or undifferentiated (Right versus left)

1.195 (0.161, 8.853) Ref 3.378 (1.518, 7.521) Ref Infinity (0, Infinity)

0.86

0.09

1.00

6.652 (0.761, 58.136) Ref 2.577 (0.697, 9.522) Ref e

0.448 (0.041, Ref Ref 1.194 (0.151, Ref 1.057 (0.249, 0 (0,.) Ref 0.589 (0.053, 0 (0,.) 0.449 (0.186,

4.945)

0.51

e

9.429)

0.87

4.497)

0.94 0.99

6.493)

0.67 0.99 0.07

Age ER PR Tumour size Race

Grade

Laterality

1.082)

0.003

Ref 1.259 (0.158, 10.035) Ref 1.091 (0.134, 8.866) 0 (0, Infinity) Ref 0.806 (0.072, 8.970) 0 (0, Infinity) 0.459 (0.118, 1.788)

0.16

0.83 0.93 0.99 0.86 1.000 0.26

Univariate analysis (n Z 5667). Risk of surgery for overall survival post surgery adjusted by age level, size level, race, grade and laterality: outcome is breast cancer-specific mortality (n Z 2551). a

b

these lesions. Although future development of subsequent invasive breast cancer is still a concern in patients with a history of LCIS and some patients and physicians choose to use bilateral mastectomy, our investigation did not reveal a benefit of bilateral mastectomy either for overall survival or breast cancer-specific survival. A higher than average risk of developing invasive cancer in the ipsilateral and contralateral breast can last for decades after a diagnosis of LCIS, with an annual risk of 1e2% and a lifetime risk of 30e40% [6,13e15]. However, the results of the present study and two previous smaller series demonstrate that the breast cancerrelated death rate is extremely low at 10 years [5,10]. A trend towards improved outcomes was also observed in these studies. Breast cancer-specific mortality decreased from 2.8% in the 1990 report [10] to 1.1% in the 2004 report [5] and 0.4% in the present study, which may be due to advances in systematic therapies for invasive breast cancers. Bilateral prophylactic mastectomy may reduce the risk for developing breast carcinoma by approximately 90% in women with a family history of breast cancer or germline BRCA1/2 mutation carriers [16,17], and a potential survival benefit of bilateral mastectomy has been observed in germline BRCA1/2 mutation carriers [17]. However, the same potential survival benefit is unlikely to be revealed in patients with LCIS because the majority of LCIS-associated breast carcinomas are characterised by a favourable prognosis and increased responsiveness to endocrine therapy [18,19], whereas breast carcinomas that develop in BRCA1/2 carriers are more likely to have high-risk features and treatment outcome is poorer, particularly in BRCA1 carriers [20e22]. The incidence of subsequent invasive breast cancer increases with increasing age. Chuba et al. reported that

the 10-year invasive breast cancer incidence increased from 5.6% for patients with LCIS who were below 40 years of age to 10.4% for patients between 60 and 69 years of age and 13.9% for patients older than age 70 years old [6]. In the present study, patients who underwent a bilateral mastectomy were younger than patients who underwent a partial mastectomy alone, and age older than 60 years was an independent risk factor for all-cause mortality. However, the deteriorated survival outcome in older patients was not due to an increased incidence of invasive breast cancer or undertreatment because most of the deaths were due to causes other than breast cancer (Supplementary Tables 6 and 7). Furthermore, neither surgery type nor age at diagnosis was an independent risk factor for breast cancer-specific mortality in the subsequent multivariate analysis, which suggests that the increased incidence of other concurrent diseases may convey a higher risk in older patients than the development of breast cancer. Because family history and genetic predisposition, such as BRCA1/2 mutation, may play an important role in the decision to undergo a bilateral mastectomy and this information is not available in the SEER database, a possible bias due to lack of this information cannot be excluded in the current study. However, family history is not a predictive factor for future development of breast cancer in patients diagnosed with LCIS [23]. Furthermore, in the present study, only one of the six deaths in the bilateral mastectomy group was due to metastatic breast cancer 7 years after the bilateral mastectomy (Supplementary Table 7), and none of the patients died of ovarian cancer, which is another common cancerrelated cause of death among BRCA1/2 mutation carriers [24]. Therefore, if there was a patient selection bias in this study, it is unlikely that it would have a

14

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

significant effect on the survival outcome analysis, particularly for all-cause mortality. Application of chemoprevention may also be a confounding factor in the current study. Chemoprevention with selective ER modulators (SERMs) and aromatase inhibitors has been shown to reduce breast cancer incidence in women who had a moderate increased risk of breast cancer and is actively under investigation [25e31]. The NCCN Breast Cancer Risk Reduction Panel recommends tamoxifen, raloxifene, exemestane and anastrozole as risk-reduction agents for most women desiring non-surgical risk-reduction therapy, and LCIS is listed as one of the primary indications for chemoprevention [9]. These recommendations might improve the adoption rate of chemoprevention in the future. However, exemestane and anastrozole were recommended in very recent years and the adoption rate of tamoxifen and raloxifene for prophylactic clinical use was very low [32,33], considering the large sample size of patients underwent partial mastectomy without radiotherapy in this study, chances were remote for chemoprevention to have a confounding effect on survival outcome analysis in our study. Whether a diagnosis of LCIS should be defined as a precursor or a risk factor remains controversial, laterality and histology investigation of subsequent malignancies may have practical significance when addressing this issue. A recent study of LCIS with long-term followup demonstrated that 63% of subsequent cancer (including invasive cancer or ductal carcinoma in situ) developed in the ipsilateral breast, but the proportion of lobular and ductal histologies were equal in patients with invasive cancer, and there were 35% of ductal carcinoma in situ in the cohort [23]. When considering invasive cancer alone, equal risk predisposition was observed in a study of 4853 patients with primary LCIS [6] and a study derived from the National Adjuvant Breast and Bowel Project B-17 study [5]. Equal risk predisposition of invasive cancer in either breast rationalised prophylactic bilateral mastectomy rather than radiotherapy to ipsilateral breast after partial mastectomy or simple mastectomy of the ipsilateral breast, if further local management was pursued after an excisional biopsy with a diagnosis of LCIS. But the excellent breast cancer-specific survival outcome of LCIS underwent partial mastectomy without radiation as observed in our study, and the development of systematic chemoprevention, questioned the necessity of bilateral prophylactic mastectomy. There were some limitations in this study. First, family histories and uptake of chemoprevention were not available in the SEER database, which made it impossible to evaluate the impact of these factors on survival outcome in the present study. Second, due to lack of central pathology review and high missing rate of clinicopathological factors like ER status, PR status and differential grades, some potentially aggressive forms of

LCIS such as pleomorphic LCIS (PLCIS) could not be identified for analysis. The aggressiveness of PLCIS was characterised by higher frequency of concomitant invasive breast cancer and ductal carcinoma in situ. In a systematic review which collected 121 cases with a core needle biopsy positive for PLCIS from literatures, only 40 patients were determined to not have concurrent disease after final pathological diagnosis of the surgical specimen [34]. Recurrence rate following excision of pure PLCIS ranged from 0% to 12.7%, but no evidence of invasive cancer or DCIS was found on recurrence specimens [34]. Managements of PLCIS are of considerable heterogenicity due to the lack of quality evidence, current guideline recommends surgical excision to investigate the existence of concomitant diseases and may consider excision with negative margins for PLCIS without concomitant invasive cancer or DCIS [8], but pursuit of clear margins may lead to more mastectomies without proven benefit and some authors urged caution in these aggressive treatments for PLCIS [35]. 5. Conclusions Survival outcomes of patients with LCIS who underwent partial mastectomy without radiotherapy were not inferior to patients who underwent bilateral prophylactic mastectomy. Breast cancer-specific mortality in patients with LCIS was extremely low, and aggressive prophylactic surgery like bilateral prophylactic mastectomy should not be advocated for most patients with LCIS. Funding This study was funded by the National Natural Science Foundation of China (81272895, 81630079) and the Medical Scientific Foundation of Guangdong Province, China (Grant Number: 20160302). Conflict of interest statement None declared. Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.ejca.2017.05.030.

References [1] Page DL, Kidd TEJ, Dupont WD, Simpson JF, Rogers LW. Lobular neoplasia of the breast: higher risk for subsequent invasive cancer predicted by more extensive disease. Hum Pathol 1991; 22:1232e9. [2] Rosen PP, Kosloff C, Lieberman PH, Adair F, Braun DWJ. Lobular carcinoma in situ of the breast. Detailed analysis of 99

Z.-M. Xie et al. / European Journal of Cancer 82 (2017) 6e15

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14] [15]

[16]

[17]

[18]

[19]

patients with average follow-up of 24 years. Am J Surg Pathol 1978;2:225e51. Salvadori B, Bartoli C, Zurrida S, et al. Risk of invasive cancer in women with lobular carcinoma in situ of the breast. Eur J Cancer 1991;27:35e7. Bodian CA, Perzin KH, Lattes R. Lobular neoplasia. Long term risk of breast cancer and relation to other factors. Cancer 1996;78: 1024e34. Fisher ER, Land SR, Fisher B, Mamounas E, Gilarski L, Wolmark N. Pathologic findings from the National Surgical Adjuvant Breast and Bowel Project: twelve-year observations concerning lobular carcinoma in situ. Cancer 2004;100:238e44. Chuba PJ, Hamre MR, Yap J, et al. Bilateral risk for subsequent breast cancer after lobular carcinoma-in-situ: analysis of surveillance, epidemiology, and end results data. J Clin Oncol 2005;23: 5534e41. Chester R, Bokinni O, Ahmed I, Kasem A. UK national survey of management of breast lobular carcinoma in situ. Ann R Coll Surg Engl 2015;97:574e7. Gradishar WJ, Anderson BO, Balassanian R, et al. NCCN clinical practice guidelines in oncology: breast cancer. Version 2.2016. Available at: NCCN.org. [Accessed 9 November 2016]. Bevers TB, Ward JH, Aran BK, et al. NCCN Clinical Practice Guidelines in Oncology: Breast Cancer Risk Reduction. Version 1.2016. Available at: NCCN.org. [Accessed 9 November 2016]. Bradley SJ, Weaver DW, Bouwman DL. Alternatives in the surgical management of in situ breast cancer. A meta-analysis of outcome. Am Surg 1990;56:428e32. Software: Surveillance Research Program, National Cancer Institute SEER*Stat software (www.seer.cancer.gov/seerstat) version 8.3.2. Li CI, Anderson BO, Daling JR, Moe RE. Changing incidence of lobular carcinoma in situ of the breast. Breast Cancer Res Treat 2002;75:259e68. Haagensen CD, Lane N, Lattes R, Bodian C. Lobular neoplasia (so-called lobular carcinoma in situ) of the breast. Cancer 1978; 42:737e69. Andersen JA. Lobular carcinoma in situ. A long-term follow-up in 52 cases. Acta Pathol Microbiol Scand A 1974;82:519e33. Ottesen GL, Graversen HP, Blichert-Toft M, Zedeler K, Andersen JA. Lobular carcinoma in situ of the female breast. Short-term results of a prospective nationwide study. The Danish Breast Cancer Cooperative Group. Am J Surg Pathol 1993;17: 14e21. Hartmann LC, Schaid DJ, Woods JE, et al. Efficacy of bilateral prophylactic mastectomy in women with a family history of breast cancer. N Engl J Med 1999;340:77e84. Heemskerk-Gerritsen BAM, Menke-Pluijmers MBE, Jager A, et al. Substantial breast cancer risk reduction and potential survival benefit after bilateral mastectomy when compared with surveillance in healthy BRCA1 and BRCA2 mutation carriers: a prospective analysis. Ann Oncol 2013. http://dx.doi.org/10.1093/annonc/mdt134. Published online April 10. Harbhajanka A, Lamzabi I, Syed S, et al. Prognostic value of coexisting lobular carcinoma in situ with invasive lobular carcinoma. Appl Immunohistochem Mol Morphol 2016;24(10):738e43. Metzger Filho O, Giobbie-Hurder A, Mallon E, et al. Relative effectiveness of letrozole compared with tamoxifen for patients

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33] [34]

[35]

15

with lobular carcinoma in the BIG 1-98 trial. J Clin Oncol 2015; 33:2772e9. Foulkes WD, Chappuis PO, Wong N, et al. Primary node negative breast cancer in BRCA1 mutation carriers has a poor outcome. Ann Oncol 2000;11:307e13. Lee E-H, Park SK, Park B, et al. Effect of BRCA1/2 mutation on short-term and long-term breast cancer survival: a systematic review and meta-analysis. Breast Cancer Res Treat 2010;122: 11e25. Zhong Q, Peng H-L, Zhao X, Zhang L, Hwang W-T. Effects of BRCA1- and BRCA2-related mutations on ovarian and breast cancer survival: a meta-analysis. Clin Cancer Res 2015;21: 211e20. King TA, Pilewskie M, Muhsen S, et al. Lobular carcinoma in situ: a 29-year longitudinal experience evaluating clinicopathologic features and breast cancer risk. J Clin Oncol 2015;33: 3945e52. Evans DGR, Ingham SL. Reduced life expectancy seen in hereditary diseases which predispose to early-onset tumors. Appl Clin Genet 2013;6:53e61. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for the prevention of breast cancer: current status of the national surgical adjuvant breast and Bowel Project P-1 study. J Natl Cancer Inst 2005;97:1652e62. Vogel VG, Costantino JP, Wickerham DL, et al. Effects of tamoxifen vs raloxifene on the risk of developing invasive breast cancer and other disease outcomes: the NSABP Study of Tamoxifen and Raloxifene (STAR) P-2 trial. JAMA 2006;295: 2727e41. Cuzick J, Sestak I, Cawthorn S, et al. Tamoxifen for prevention of breast cancer: extended long-term follow-up of the IBIS-I breast cancer prevention trial. Lancet Oncol 2015;16:67e75. Cuzick J, Sestak I, Bonanni B, et al. Selective oestrogen receptor modulators in prevention of breast cancer: an updated metaanalysis of individual participant data. Lancet (London, England) 2013;381:1827e34. Cuzick J, Sestak I, Forbes JF, et al. Anastrozole for prevention of breast cancer in high-risk postmenopausal women (IBIS-II): an international, double-blind, randomised placebo-controlled trial. Lancet (London, England) 2014;383:1041e8. Goss PE, Ingle JN, Ales-Martinez JE, et al. Exemestane for breast-cancer prevention in postmenopausal women. N Engl J Med 2011;364:2381e91. Mocellin S, Pilati P, Briarava M, Nitti D. Breast cancer chemoprevention: a network meta-analysis of randomized controlled trials. J Natl Cancer Inst 2016:108. http://dx.doi.org/10.1093/jnci/djv318. Crew KD. Addressing barriers to uptake of breast cancer chemoprevention for patients and providers. Am Soc Clin Oncol Educ Book 2015:e50e8. Bevers TB. Breast cancer risk reduction therapy: the low-hanging fruit. J Natl Compr Canc Netw 2015;13:376e8. Wazir U, Wazir A, Wells C, Mokbel K. Pleomorphic lobular carcinoma in situ: current evidence and a systemic review. Oncol Lett 2016;12(6):4863e8. Flanagan MR, Rendi MH, Calhoun KE, Anderson BO, Javid SH. Pleomorphic lobular carcinoma in situ: radiologicpathologic features and clinical management. Ann Surg Oncol 2015;22:4263e9.