Breast Cancer Research and Treatment

, Volume 111, Issue 3, pp 541–547

A retrospective review with long term follow up of 11,400 cases of pure mucinous breast carcinoma

Authors

  • Salomone Di Saverio
    • Department of Surgery, Sylvester Comprehensive Cancer Center, Miller School of MedicineUniversity of Miami
  • Juan Gutierrez
    • Department of Surgery, Sylvester Comprehensive Cancer Center, Miller School of MedicineUniversity of Miami
    • Department of Surgery, Sylvester Comprehensive Cancer Center, Miller School of MedicineUniversity of Miami
Epidemiology

DOI: 10.1007/s10549-007-9809-z

Cite this article as:
Di Saverio, S., Gutierrez, J. & Avisar, E. Breast Cancer Res Treat (2008) 111: 541. doi:10.1007/s10549-007-9809-z
  • 639 Views

Abstract

Background Pure mucinous breast carcinoma (PMBC) is a rare histologic type of mammary neoplasm. It has been associated with a better short-term prognosis than infiltrating ductal carcinoma (IDC) but identical long-term survival curves have been reported. The value of tumor size for TNM staging has been challenged because of the mucin content of the lesions. This study presents a large PMBC series with 20 years follow up as compared to IDC. The relative significance of a variety of common prognostic factors is calculated for this uncommon histology. Materials and methods A retrospective analysis of all PMBC cases reported in the SEER database between 1973 and 2002 was conducted. Overall survival (OS) and disease specific survival (DSS) were calculated at 5, 10, 15 and 20 years of follow up. Those curves were compared with all the IDC cases reported into the database during the same period. The prognostic significance of gender, race, laterality, age at diagnosis, T and N status, estrogen and progesterone receptors and administration of radiation therapy was calculated by univariate and multivariate analysis. Results There were 11,422 PMBC patients reported. The median age at diagnosis was 71 years (Range 25–85). Fifty three percent of the tumors were well differentiated, 38% were moderately differentiated and the remaining 9% were poorly differentiated or anaplastic. The majority of the tumors were located in the upper outer quadrant (44%) the other 56% were roughly evenly divided between the upper inner, lower inner, lower outer and central quadrants. Eighty six percent of the patients had only localized disease at the time of surgery without nodal or distant disease while 12% had regional nodal involvement and 2% had distant metastases. The PMBC cases showed a better differentiation with lesions of lesser grade and more frequent ER/PR expression, smaller size and lesser nodal involvement when compared to the IDC cases of the same period. Kaplan Meier survival curves revealed a 5 years. breast cancer specific survival rate of 94%. Although slowly decreasing with time, 10, 15 and 20 years survival were 89%, 85% and 81% respectively compared to 82% (5 year), 72% (10 year), 66% (15 year) and 62% (20 year) for IDC. There were no significant differences in overall survival. Multivariate analysis by Cox regression revealed the nodal status (N) to be the most significant prognostic factor followed by age, tumor size (T), progesterone receptors and nuclear grade. Disease specific survival curves stratified for nodal status revealed a highly significant difference between node negative and node positive patients. The addition of radiation therapy after surgery did not significantly improve overall survival. Conclusions This large retrospective comparative analysis confirms the less aggressive behavior of PMBC compared to IDC. This favorable outcome is maintained after 20 years. This tumor presents typically in older patients and is rarely associated with nodal disease. Positive Nodal status appears to be the most significant predictor of worse prognosis.

Keywords

Breast neoplasmsPure mucinous breast carcinomaInfiltrating ductal carcinomaLymph node statusRisk factorsPrognosis

Background and aims

Pure mucinous breast carcinoma (PMBC) is a rare histologic type of mammary neoplasm, representing 1–4% of all breast cancers and associated with a better prognosis than infiltrative ductal carcinoma (IDC) [13]. In the elderly, a slightly higher incidence rate of 6–7% was reported [4]. PMBC has to be distinct from the histologically mixed forms of ductal carcinoma with a mucinous component (occurring in 2% of cases) because those tumors carry a prognosis identical to the non-mucinous component [3, 5]. The classification of the mucinous histology and the distinction between the two subgroups of pure and mixed mucinous carcinomas is based upon the quantification of cellularity [6], mucinous neoplasms being defined as having a mucinous component of more than 50% of the lesion. The “pure” type consists exclusively of tumor tissue with extracellular mucin production, while the “mixed” form of mucinous carcinoma also contains infiltrating ductal epithelial component without mucin [7]. A precisely defined percentage of mucinous component for the classification and distinction of pure and mixed lesions is however not yet clearly established. [810]. Although, the pure mucinous subtype has been associated with a better short term prognosis than IDC [11], long term survival curves have been reported as identical, noting a worsening of prognosis after longer follow up [12] [13]. PMBC tends to grow very slowly but can reach large diameters at diagnosis [6] [14]. Those lesions are frequently associated with estrogen receptors positivity [15] and despite large tumor sizes, the axillary lymph nodes are rarely involved [3, 10, 16] [17, 18] accounting for the better biological behavior. Axillary nodal involvement, although rare, seems to affect and worsen the prognosis [19, 20].

Because of its relative rarity, most reported PMBC series have a low number of patients and a limited long term follow up. The significance and prognostic value of the clinico-pathological factors commonly used for IDC is also not well established and therefore the best treatment guidelines for optimal local regional and systemic control of this neoplasm are mostly extrapolated from the treatment of IDC without clear validation in a PMBC series of patients.

The objective of this study was to evaluate the long term disease specific and overall survival as well as the significance of the available clinical and pathological prognostic factors for PMBC in a large and robust database. The retrospective analysis also included a comparison between PMBC and IDC recorded in the SEER database in the same period as well as a comparative evaluation of their demographic, clinical and pathological features and the outcomes in terms of Disease Specific and Overall survival.

Patients and methods

A retrospective comparative analysis of all PMBC and IDC cases reported in the SEER database between 1973 and 2002 was conducted. The SEER (Surveillance Epidemiology and End Results) is a cancer registry held by the cancer statistics branch of the surveillance program promoted by the National Cancer Institute. It is an authoritative source of information on cancer incidence and survival in the Unites States. SEER collects and publishes cancer incidence and survival data from population-based cancer registries covering approximately 26% of the US population. The population covered by SEER is epidemiologically comparable and representative of the US population. The SEER database includes and records all the tumor cases from 18 cancer registries from several states and is considered the standard for quality among cancer registries around the world. It began collecting data on cancer cases in 1973 and recorded several demographics, clinical and pathological variables as well as follow up and survival data [21]. Overall survival (OS) and disease specific survival (DSS) were calculated at 5, 10, 15 and 20 years of follow up. Those curves were compared with the 338,479 IDC cases reported into the database during the same period. The prognostic significance of gender, race, laterality, age at diagnosis, T and N status, nuclear grade, estrogen and progesteron receptors and administration of radiation therapy was calculated by univariate and multivariate analysis.

The statistical analysis has been carried out using a statistical software package (SPSS 13.0 SPSS Inc. Chicago. IL). The association between the discrete variables were evaluated by χ2 or Fisher exact test as appropriate. For means in case of continuous numerical data, we used the independent samples t-test and the Mann–Whitney test, respectively for data normally and non-normally distributed. The data was previously tested for normality by the Kolmogorov–Smirnov test. A statistical analysis of the breast cancer-specific survival and overall survival for each group (obtained by Kaplan–Meier curves for univariate analysis and Cox logistic regression for multivariate analysis) was conducted. The comparison for statistically significant differences between survival curves was performed by the log rank test (Mantel Cox). Statistical significance was considered to be reached at a level of < 0.05.

Results

There were 11,422 PMBC patients reported; 56 of them were males. The median age at diagnosis was 71 years (Mean 68 years Range 25–85). The distribution of population in different groups of age revealed 66.3% of patients to be 65 years old or above, 15.5% between 55 and 65 years and 18.2% under 55 (Fig. 1). The demographics were 85.2% Caucasian and about 7% each Asian and African–American (Table 1). Fifty three percent of the tumors were well differentiated, 38% were moderately differentiated and the remaining 9% were poorly differentiated or anaplastic (Table 1). The majority of the tumors were located in the upper outer quadrant (44%) while the other 56% were roughly evenly divided between the upper inner, lower inner, lower outer and central quadrants. The tumor markers, when available, confirmed the PMBC to be a well-differentiated lesion, associated with estrogen receptors positivity in 94% of tumors and progesterone receptors positivity in 81.5% (Table 1). The mean tumor size was almost 22 mm (median 16 mm) (Fig. 2) and the size distribution showed 83.2% of tumors to be 30 mm or less. Eighty-six percent of the patients had only localized disease at the time of surgery without nodal or distant metastases while 12% had regional nodal involvement and 2% had distant metastases (Table 1). Radiation therapy was administered to 31.5% of the patients. Comparison of those clinico-pathological features between the PMBC cases recorded in the SEER database in the period 1973–2002 and the Infiltrating Ductal Carcinoma cases of the same period yielded interesting results (Table 1). While no differences were seen in distributions of sex and race, the age of PMBC patients was significantly older (median 71 years vs. 61 years). The PMBC lesions were of more differentiated nuclear grade and showed wider expression of estrogen and progesteron receptors. The tumor stage was lower in the PMBC patients, with smaller tumor size (median 15.7 mm vs. 18 mm) and less frequent nodal involvement. No significant differences were seen between the two groups in the administration of adjuvant radiotherapy. After a mean follow up period of 84 months 908 breast cancer related deaths were observed out of 11,422 cases and 3539 died of other causes. The Kaplan Meier breast cancer specific survival curves revealed a 5 year survival rate of 94%. Although slowly decreasing with time, 10, 15 and 20 year survival rates were 89%, 85% and 81% respectively compared to 82% (5 year), 72% (10 year), 66% (15 year) and 62% (20 year) for IDC; the difference between those curves was highly statistically significant (Fig. 3). The stratification in groups by stage of disease at diagnosis showed the same highly significant differences for each stage: the mean cause specific survival period was of 327 months for PMBC cases and 298 months for IDC lesions in the localized stage, 241 versus 208 months in the regional stage and 78 versus 58 months in the distant stage groups (Fig. 4). The same comparison for overall survival curves did not reach statistical significance (Fig. 3). Multivariate analysis by Cox regression revealed the nodal status (N) to be the most significant prognostic factor followed by age, tumor size (T), progesteron receptors and nuclear grade (Table 2). Furthermore we conducted a comparison of disease specific survival curves (Kaplan Meier) stratifying the population by pathological lymph node status. The analysis revealed a worse survival rate with a statistically highly significant difference between the node negative and node positive cases; the difference was shown to be higher with longer follow up (Fig. 5). The disease specific survival for node negative patients was 97% at 5 years, slowly decreasing after longer follow up to 93% (10 year), 90% (15 year) and 86% (20 year) while rapidly worsening survival percentages, 85% (5 year), 73.4% (10 year), 64% (15 year) and 60% (20 year) were observed for node positive cases. Adjuvant radiotherapy was associated with a small disease specific survival advantage, significant at P > 0.05 degree (Fig. 6) but on multivariate analysis adjuvant radiotherapy was not associated with a better outcome. Specifically it did not significantly improve the DSS for patients younger than 50 years, with lesions greater than 50 mm, node positive, poorly differentiated histology and ER negative.
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9809-z/MediaObjects/10549_2007_9809_Fig1_HTML.gif
Fig. 1

Age distribution curve of the population of 11,422 PMBC cases (source SEER database 1973–2002)

Table 1

Clinico-pathologic features of 11,422 PMBC patients (SEER database 1973–2002) and comparison with 338,479 IDC patients (SEER database 1973–2002)

 

PMBC (%)

IDC (%)

P

Age

   

    Mean

68.3

61

<0.01

    Median

71

61

<0.01

Race

  

ns

    White

85.2

85.5

 

    Black

7.0

7.7

 

    Other

7.4

6.4

 

    Unknown

.4

.4

 

Sex

  

ns

    Females

99.5

99.3

 

    Males

.5

.7

 

Grade

  

<0.01

    Well differentiated

53.4

14.7

 

    Moderately differentiated

37.6

41.8

 

    Poorly differentiated––anaplastic

9.0

43.5

 

ER

  

<0.01

    Positive

94.1

74.9

 

    Negative

5.3

24.2

 

    Borderline

.6

.9

 

PR

  

<0.01

    Positive

81.5

65.2

 

    Negative

17.6

33.7

 

    Borderline

.9

1.1

 

Stage

  

<0.01

    Localized

85.8

64.8

 

    Regional

12.1

31.7

 

    Distant

2.1

3.5

 

T size (mm)

  

<0.01

    Mean

21.9

23.5

 

    Median

15.7

18

 

Nodal Status

  

<0.01

    Negative

87.7

64

 

    Positive

12.3

36

 

RT Adjuv

   

    Yes

31.5

34

ns

    No

68.5

66

 
https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9809-z/MediaObjects/10549_2007_9809_Fig2_HTML.gif
Fig. 2

Size distribution curve of the PMBC cases

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9809-z/MediaObjects/10549_2007_9809_Fig3_HTML.gif
Fig. 3

Comparative Kaplan Meier cause specific and overall survival curves between PMBC (11,422 cases) and IDC (338.479 cases)

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9809-z/MediaObjects/10549_2007_9809_Fig4_HTML.gif
Fig. 4

Comparison of the Kaplan Meier cause specific survival curves and mean cause specific survival period of PMBC and IDC, stratified by stage at diagnosis

Table 2

Prognostic significance of 10 clinico-pathological factors and influence on survival

Covariates (Ca = Categorical; Co = Continous; D = Dichotomous)

Univariate analysis

Multivariate analysis Cox regression of survival

χ2 score

P

B Coeff

Wald

Exp (B)

P

Gender (D)

1.942

.163

    

Race (Ca)

3.802

.284

    

Laterality (Ca)

1.017

.313

    

Age at diagnosis (Co)

10.063

.002

.036

15.712

1,037

.000

T size (Co)

54.418

.000

.013

11.436

1,013

.001

N pathological status (D)

66.286

.000

−1.253

20.293

.286

.000

Nuclear grade (Ca)

27.717

.000

−.938

7.445

.392

.024

ER status (Ca)

4.530

.104

    

PR status (Ca)

15.446

.000

.733

9.497

2,082

.009

Adjuvant radiotherapy (D)

6.433

.012

.592

.028

1.807

.119

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9809-z/MediaObjects/10549_2007_9809_Fig5_HTML.gif
Fig. 5

Comparison of disease-specific survival curves of node negative (0) and node positive (1) PMBC

https://static-content.springer.com/image/art%3A10.1007%2Fs10549-007-9809-z/MediaObjects/10549_2007_9809_Fig6_HTML.gif
Fig. 6

Adjuvant radiotherapy and Disease specific survival in the PMBC patients (0 = no RT; 1 = RT)

Discussion

The pure histotype of mucinous breast carcinoma is rare. The few and small series described in the literature indicate an indolent biological behavior and a better long-term prognosis than Invasive Ductal Carcinoma [3, 6, 10, 12, 13, 18, 19, 22]. Recently Fujii et al. investigated the genomic features of PMBC and demonstrated that it does not have the extensive genomic alterations typically found in the more common variants of breast cancer. Moreover mucinous carcinoma has less genetic instability than the other forms of breast cancer and its molecular pathogenesis seems substantially different compared to the usual breast carcinoma [23]. Other studies carried out histochemical and immunohistochemical examinations of mucin revealing that the mucinous growth and the pure type of mucinous carcinoma may originate from intraductal carcinoma [24]. Mammographically PMBC tends to present as a well circumscribed lesion [2527], isoechogenic to the breast fat on ultrasonography [28]. For that reason a significant number of lesions could be misinterpreted as benign on screening mammograms. Interestingly, a delay in diagnosis however may not cause a significant adverse outcome for most women [29]. A careful diagnostic investigation of the radiological characteristics may be helpful in the identification of PMBC. On Magnetic Resonance Imaging, PMBC is associated with a very specific appearance showing a gradually enhancing contrast pattern and a very high signal intensity on T2-weighted images [30, 31]. Although fine needle aspiration cytology (FNA) can diagnose mucinous carcinoma and can demonstrate nicely the mucin component, the sensitivity of FNA is 56% [32] while the histology from core biopsy achieved 100% sensitivity and accuracy [33, 34].

Pure Mucinous Breast Carcinoma is a disease of older women and only 1% of PMBC patients are younger than 35 years [4, 12]. Our series confirms this finding demonstrating significantly older age at diagnosis for PMBC compared to IDC.

Our large series of PMBC from SEER demonstrates a high percentage of hormone receptor expression, indicating a better differentiation than the ductal lesions. Those findings are in agreement with Komenaka's findings reporting a rate of positivity of 91% and 79% for estrogen and progesteron receptors respectively [35].

Although the size of PMBC lesions at diagnosis varies widely and has been described from non-palpable lesions to as large as 20 cm [18, 3540], the median size of PMBC lesions in our study was significantly smaller than IDC lesions. A potential explanation could be the relatively slow growth of PMBC usually allowing diagnosis at a smaller size.

The prognostic significance of tumor size is an especially interesting point with PMBC. Komenaka et al. did not find size to be a significant prognostic factor in their series suggesting that because the majority of the tumor volume consists of mucin it may not impact survival [35]. In our multivariate analysis of a large number of PMBC, tumor size was an independent prognostic indicator but it was less significant than nodal status and age. Although some authors found that the incidence of node positivity was directly related to tumor size [41], other studies did not find any correlation between tumor size and the incidence of axillary nodal metastases [17, 42]. In this study we did find a significant correlation between tumor size and incidence of nodal involvement in PMBC.

Nodal positivity is rare and was detected in 12% of the patients in our review. The same percentage was noticed by Avisar et al. [17], with the rate being 14% in the series from Komenaka et al., and ranging from 2% to 14% in other series [5, 9, 10, 28, 43] compared to 46% to 64% for mixed mucinous lesions [3, 5, 43].

Previous studies have shown the nodal status to be the strongest predictor of disease-specific survival [1, 18, 35] and the majority of patients presenting with axillary metastases proceeded to develop distant metastases [9]. Our multivariate analysis confirms those findings and supports the important prognostic value of sentinel node biopsy in those patients despite the relatively low rate of nodal involvement.

Finally the prognosis of PMBC and the overall and disease specific survival are very favorable. Previous series have suggested that the observed favorable five years survival is temporary and that PMBC tends to recur after 10 years similarly to IDC [9]. Komenaka et al. reported disease-specific survival rates of 95% and 79% respectively at 5 and 10 years [35]. Our study clearly shows that the DSS is significantly higher for PMBC compared to IDC and that survival advantage is maintained after 20 years.

Diab et al. demonstrated a significantly better disease-free survival for the mucinous lesions versus the NOS carcinomas even in the subset of node negativity [41]. Also Thurman et al. [44] compared ductal NOS carcinomas, tubular and mucinous carcinomas and confirmed a better prognosis, lesser nodal involvement and wider Estrogen and Progesteron receptors expression associated with the mucinous lesions. This series confirms a highly statistically significant DSS advantage for PMBC compared to IDC for localized, regional and even for metastatic disease claiming for a truly better tumor biology.

Conclusions

This large retrospective comparative analysis confirms the less aggressive behavior of PMBC compared to IDC. This favorable outcome is maintained after 20 years. This tumor typically presents in older patients and is rarely associated with nodal disease. Positive Nodal status appears to be the most significant predictor of worse prognosis. An axillary staging by sentinel lymph node biopsy could guarantee a reliable staging of disease, identification of the more aggressive lesions and planning of further therapies.

Copyright information

© Springer Science+Business Media, LLC. 2007