Important increase of invasive lobular breast cancer incidence in Geneva, Switzerland

Authors


Abstract

A recent paper from the United States reported a sharp and unexplained increase in invasive lobular breast cancer incidence since 1977 (Li et al., Cancer 2000;88:2561–9). We investigated if this trend was also present in Geneva, Switzerland, where breast cancer incidence is one of the highest in Europe. We analyzed trends in breast cancer incidence according to histologic subtype, age and stage, to clarify the pattern. Our population-based study includes all histologically confirmed invasive breast carcinomas (n = 6,247) recorded between 1976 and 1999 at the Geneva Cancer Registry. Breast histology was classified as ductal carcinoma, lobular carcinoma and other. Incidence trends were studied by log-linear regression analyses. Models including effects of age, period and birth cohorts were used to describe rising incidence trends. The incidence of ductal carcinoma increased 1.2% per year (ptrend < 0.001) from 85.2 to 110.1/100,000. This increase concerned women aged 50–69 years and early-stage tumors. Lobular cancer incidence increased disproportionately (14.4% per year, ptrend < 0.01) and rose from 2.9 to 20.5/100,000. This increase affected all age categories and both localized and advanced stages. In addition, a strong age-cohort effect was present (p < 0.05), and women aged 50–59 years born after 1944 experienced the most marked increase. Our study shows a disproportionate increase of lobular breast cancer incidence compared to ductal cancer incidence. Contrary to ductal cancer, trends for lobular cancer are unlikely to be explained by increased use of screening mammography. Other explanations must be researched, in particular the role played by hormone replacement therapy. © 2003 Wiley-Liss, Inc.

In Switzerland, breast cancer is a major public health concern as it is the most frequent cancer among women and the leading cause of death among middle-aged women. It is estimated that 4,000 women develop breast cancer each year and 1,600 die of the disease. The canton of Geneva is particularly affected, presenting the highest European incidence rates.1

The most common histologic subtype of breast cancer is invasive ductal cancer, which accounts for 70–80% of cases.2 Invasive lobular carcinoma of the breast is the second most frequent histologic subtype and comprises generally 5–10% of all invasive breast carcinomas.2, 3

In invasive lobular carcinoma, the tumor is accompanied by a desmoplastic stromal reaction and a linear arrangement of tumor cells (Indian file pattern).4 Invasive lobular carcinoma generally presents as an ill-defined thickening in the breast rather than a clearly distinct dominant mass, which makes it more difficult to diagnose mammographically and on physical examination than invasive ductal cancer.5, 6

Li et al.7 reported that the incidence of invasive lobular breast cancer in the United States has been increasing disproportionately compared to the incidence of other invasive breast carcinomas. They suggested that hormone replacement therapy may have played a role in this increasing incidence.

Our aim was to describe the trends in breast cancer incidence in Geneva by histologic subtype and to discuss the possible factors contributing to the pattern.

MATERIAL AND METHODS

The data considered for the current analysis were derived from the Geneva Cancer Registry, which includes information regarding all incident cases of malignant neoplasms occurring in the canton (approx. 400,000 inhabitants). Registration is based on several sources of information and is very accurate, as attested by its low percentage (<2%) of cases recorded from death certificate only.8 Individual clinical files from all university public hospitals are systematically consulted, and inquiry forms are addressed regularly to physicians for patients treated in the private sector. Recorded data include sociodemographic characteristics, tumor characteristics coded according to the International Classification of Diseases for Oncology (ICD-O)9 and stage of disease at diagnosis.

We included all incident cases of invasive breast carcinoma (ICD code 174) diagnosed between 1976 and 1999 among the resident population. Women with pure in situ lesions (ductal carcinoma in situ or lobular carcinoma in situ) were excluded. Histologic review of microscopic slides was not performed.

Breast cancers were classified into the following histologic subcategories: ductal carcinoma (ICD-O code 8500), lobular carcinoma (lobular carcinoma not otherwise specified, IDC-O code 8520, and invasive duct and lobular carcinoma, ICD-O code 8522) and other (all other cancers including those with no microscopic confirmation). Estrogen and progesterone receptors were classified as positive (≥10%) or negative (<10%). Information on hormone receptor status was routinely collected only after 1995.

Incidence rates (European age-standardized) from 1976 to 1999 for 5 different time periods (1976–1979, 1980–1984, 1985–1990, 1991–1994 and 1995–1999) were calculated. Five age groups were analyzed: <50, 50–59, 60–69, 70–79 and 80+ years. For stage analysis, we used the 3 Surveillance, Epidemiology and End Results (SEER) staging categories for invasive carcinoma: localized (disease confined to the breast, including nipple and areola), regional (disease invading surrounding tissues or involving axillary or internal mammary lymph nodes) and distant (disease involving infraclavicular, supraclavicular, cervical or other distant lymph nodes or other systemic metastases).10

Variations in incidence rates were studied by log linear regression analyses based on the maximum likelihood method.11 Models including the effects of age, period and birth cohorts were used to better describe incidence trends.12, 13

RESULTS

Between 1976 and 1999, 6,247 cases of invasive breast cancer were diagnosed in Geneva. Of those, 4,465 (71.5%) were classified as invasive ductal cancer, 427 (6.8%) as invasive lobular breast cancer and 1,355 (21.7%) as other. Most prevalent histologic subtypes in the “other” category were (adeno) carcinoma not otherwise specified (n = 461), tubular carcinoma (n = 50), mucinous carcinoma (n = 112), medullary carcinoma (n = 92), inflammatory carcinoma (n = 71) and breast cancer without microscopic confirmation (n = 208).

The characteristics of the 3 histologic categories of breast cancer are presented in Table I. Most of the characteristics of invasive lobular carcinoma and invasive ductal carcinoma were quite similar, but tumors in the “other” category occurred more often among older, widowed women and presented more often with distant metastases. For the period 1995–1999, invasive lobular carcinomas were more often estrogen receptor-positive (i.e., 93% compared to 82% of ductal cancers and 72% of other cancers) and progesterone receptor-positive (i.e., 76% compared to 64% of ductal cancers and 51% of other cancers).

Table I. Characteristics of Invasive Ductal Carcinoma (IDC), Invasive Lobular Carcinoma (ILC) and Other Carcinoma of the Breast Diagnosed Between 1976 and 1999 in Geneva, Switzerland
 IDC (n = 4,465)ILC (n = 427)Other (n = 1,355)
Age (years)   
 <501,088 (24%)98 (23%)230 (17%)
 50–591,129 (25%)142 (33%)200 (15%)
 60–69998 (22%)90 (21%)223 (16%)
 70–79845 (19%)70 (16%)278 (21%)
 80+415 (9%)27 (6%)424 (31%)
 Mean (range)60 (22–98)60 (34–95)68 (23–101)
Place of birth   
 Switzerland2,667 (60%)247 (59%)902 (68%)
 Latin Europe1,063 (24%)97 (23%)241 (18%)
 Other684 (15%)77 (18%)191 (14%)
 Unknown51621
Civil status   
 Single612 (14%)45 (11%)203 (15%)
 Married2,442 (54%)242 (57%)492 (36%)
 Widowed843 (19%)86 (20%)468 (35%)
 Separated/divorced587 (13%)54 (13%)192 (14%)
 Unknown1
Sector of care   
 Private2,100 (47%)196 (46%)467 (35%)
 Public2,358 (53%)229 (54%)863 (65%)
 Unknown725
Stage at diagnosis   
 Localized2,426 (55%)239 (57%)634 (52%)
 Regional1,764 (40%)160 (38%)372 (31%)
 Distant214 (5%)22 (5%)212 (17%)
 Unknown616137

Overall breast cancer incidence increased from 116.3 to 156.4/100,000 between 1976–1979 and 1995–1999 (ptrend < 0.001). Figure 1 presents the trends of invasive breast cancer by histologic subtype. The incidence of invasive ductal breast cancer increased from 85.2 to 110.1/100,000. The mean annual increase was 1.2% (ptrend < 0.001), and the mean annual number of cases rose from 155 to 232 between 1976–1979 and 1995–1999. Figure 2a shows that the trend was different among the 5 age categories. Invasive ductal breast cancer incidence increased significantly among women aged 50–59 and 60–69 but decreased significantly among women older than 80 years. The increase of ductal cancer concerned only localized cases (Fig. 3). The age-adjusted incidence of localized cases increased 2.2% per year (ptrend < 0.001), while the age-adjusted incidence of ductal cancer with regional or distant spread remained stable (ptrend not significant).

Figure 1.

Trends in incidence rates of breast cancer according to histologic subtype, Geneva Cancer Registry, 1976–1999. Incidence rates are adjusted for age (European standard). Above the curve, the mean annual increase or decrease is presented.

Figure 2.

Trends in incidence of invasive ductal (a), lobular (b) and other types of (c) cancer according to age, Geneva Cancer Registry, 1976–1999.

Figure 3.

Trends in incidence of invasive ductal cancer (IDC) and invasive lobular cancer (ILC) by stage, Geneva Cancer Registry, 1976–1999. Nonlocalized refers to regional or distant spread of disease. Incidence rates are adjusted for age (European standard). With the curves, the mean annual increase or decrease is presented.

The incidence of invasive lobular carcinoma increased from 2.9 to 20.5/100,000 between 1976–1979 and 1995–1999. The mean annual increase was 14.4% (ptrend < 0.01), and the mean annual number of cases rose from 5 to 43 between 1976–1979 and 1995–1999 (Fig. 1). The increasing incidence was present among all age categories (Fig. 2b) but more prominent after the age of 50 years. The incidence increase affected both localized stages (ptrend < 0.001) as well as cases with regional or distant spread (ptrend < 0.001) (Fig. 3). Age-period-cohort analysis showed strong differences in invasive lobular cancer incidence according to birth cohort and age (p < 0.05) (Fig. 4). For women under the age of 50 or over the age of 70 years, the risk of invasive lobular cancer according to birth cohort remained stable. For the age categories 50–59 and 60–69, the pattern was completely different. For women born before 1925, the risk of invasive lobular cancer was low. For the generation born between 1925 and 1944, the risk of invasive lobular carcinoma increased markedly, especially among women aged 60–69 years. The highest risk was observed for women aged 50–59 years who were born after 1944.

Figure 4.

Incidence rates of invasive lobular cancer for different birth cohorts according to age.

The incidence of other breast cancers decreased from 28.2 to 25.5/100,000 between 1976–1979 and 1995–1999 (mean annual decrease 1.3%, ptrend < 0.05, mean annual number of cases 52 and 54 in 1976–1979 and 1995–1999, respectively) (Fig. 1). Figure 2c shows that this decrease was almost exclusively present among women 80 years and older. This pattern was due to a sharp decrease of breast cancer not otherwise specified (ICD-O code 8010 or 8140) and breast cancer without microscopic confirmation (ICD-O code 9990) among the elderly (data not shown).

DISCUSSION

Breast cancer incidence in Geneva is among the highest in the world, and it is still increasing.1 Our observations show that the increased risk concerns both invasive ductal and lobular breast cancers. The sharp increase in invasive lobular carcinoma incidence, especially after 1995, was particularly remarkable and demands to be further addressed.

A large proportion of the increased ductal breast cancer incidence is likely due to the widespread use of screening mammography. Since 1985, regular screening mammography, either in the context of the organized screening program or by opportunistic screening, is increasingly being offered to the female Geneva population between 50 and 69 years of age. That the incidence of invasive ductal cancer concerned exclusively localized cases, was present only in the ‘screening’ age categories (i.e., 50–59 and 60–69 years) and decreased in women over 80 years indicates that screening plays a major role in the incidence pattern for invasive ductal cancer.

Increased mammographic screening is unlikely to explain the observed rise of invasive lobular carcinoma incidence because this rise concerned both localized cases and cases with regional and distant spread and because the rise was present in all age categories and not only in screening age groups. Moreover, invasive lobular carcinoma is difficult to detect on mammograms. Lobular carcinoma cells classically infiltrate into the surrounding breast tissue as single cells and often elicit a minimal stromal response, thereby resulting in tumors that are generally less dense than their ductal counterparts.2, 7

One can hypothesize that (part of) the invasive lobular carcinoma incidence increase was caused by a change in histologic diagnostic criteria as in the late 1970s the definition of invasive lobular carcinoma was broadened after new histologic variants were defined.14, 15 We did not review the microscopic slides and cannot rule out that such a change in diagnostic criteria influenced the invasive lobular carcinoma incidence. As this expansion of criteria took place in the late 1970s, one can expect the consequential increase in invasive lobular carcinoma incidence to manifest in the 1980s. Although our series shows an incidence increase between 1980 and 1989, the most important incidence increase was observed a long time after the criteria expansion, in particular after 1995. In addition, if changed diagnostic criteria were the cause of the lobular cancer incidence, one would expect a visible decrease in another histologic category. Although there was a slight decrease in the “other” histology category, closer analysis showed that this decrease was due to a decrease of carcinoma not otherwise specified and carcinoma without microscopic confirmation among the elderly. As lobular cancer incidence affects all age categories, especially those between 50–69, it is unlikely that this phenomenon explains the increasing lobular cancer incidence.

Hormone replacement therapy should be considered as a plausible contributor to the increase of invasive lobular carcinoma. Hormone replacement therapy helps to alleviate menopausal symptoms and has a favorable influence on osteoporosis.16, 17 In Switzerland, as in the rest of Europe and the United States, hormone replacement therapy is increasingly being prescribed.7, 18, 19, 20, 21 In the Swiss canton of Geneva, the proportion of women using hormone replacement therapy is very high: in 1996, over 50% of women aged 45–59 were reported to use hormone replacement therapy.18

In 1997, the Collaborative Group on Hormonal Factors in Breast Cancer combined the results of 51 studies on hormone replacement therapy and found that the risk of having breast cancer diagnosed is increased among women using hormone replacement therapy and increases with longer duration of use.22 More recent case-control studies indicate that hormone replacement therapy use is particularly related to lobular breast cancer.23, 24, 25 One study showed that women who used or had been using hormone replacement therapy for more than 57 months in the past 5 years had a 3-fold increased risk of developing invasive lobular carcinoma and a 1.5-fold increased risk of developing invasive ductal carcinoma compared to women who never used hormone replacement therapy.23 The risk of developing invasive lobular carcinoma was particularly high among women using combined hormone replacement therapy (containing both estrogen and progesterone) for more than 11 months.25 Their risk of developing invasive lobular carcinoma was increased 6-fold compared to women not using hormone replacement therapy, while their risk of developing invasive ductal cancer was not significantly increased. Likewise, Li et al.24 reported that users of combined hormone replacement therapy were at a 2.6-fold increased risk of developing invasive lobular carcinoma but that their risk of developing invasive ductal cancer remained unchanged.

Our study provides descriptive data that strongly suggest a role of hormone replacement therapy in the lobular cancer increase. The risk of invasive lobular carcinoma remained stable among women who went through menopause before or around 1975, a period in which the use of hormone replacement therapy was quite uncommon. The risk of lobular carcinoma increased markedly for the generation who went through menopause between 1975 and 1995, a period in which hormone replacement therapy was introduced and increasingly prescribed in Geneva. The incidence of invasive lobular carcinoma was extremely high for women born after 1944 who went through menopause after 1995 and are very likely to have received hormone replacement therapy.

An increase in lobular breast cancer incidence could have several clinical implications. Firstly, lobular carcinomas are relatively more difficult to diagnose mammographically and therefore more likely to be escape mammographic detection.3 In addition, many of the prognostic factors (e.g., nuclear grade, lymphatic vessel invasion, hormone receptor status) used in ductal carcinoma may not be applicable to invasive lobular carcinoma.26 Invasive lobular carcinoma also tends to have different metastatic patterns compared to other breast tumors. Invasive lobular carcinoma has a higher probability of metastasizing to bone, gastrointestinal, genitourinary and endocrine tissues, whereas the incidence of lung and pleural metastases is lower.3

Our study demonstrates that the risk of invasive lobular breast cancer increased disproportionately among the female Geneva population. Additional studies are needed to study the cause of this increase, in particular the role played by hormone replacement therapy.

Acknowledgements

We thank Mr. L. Raymond and Mr. P. Pury for their expertise in age-period-cohort analyses and Mrs. S. Blagojevic for editorial assistance.

Ancillary