Trends in Incidence Rates of Invasive Lobular and Ductal Breast Carcinoma

Context  Research has suggested that use of combined estrogen and progestin hormone replacement therapy (CHRT) increases breast cancer risk and that CHRT use is more strongly associated with the risk of invasive lobular breast carcinoma than that of invasive ductal carcinoma. Lobular carcinoma is less common than ductal carcinoma but can be more difficult to diagnose because of its subtle elusive infiltrative pattern.

Objective  To evaluate trends in invasive lobular and ductal carcinoma incidence rates from 1987 through 1999, during which time use of CHRT increased in the United States.

Design  Descriptive epidemiologic study.

Setting  Nine cancer registries that participate in the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute and that cover Atlanta, Ga; Detroit, Mich; San Francisco-Oakland, Calif; Seattle, Wash; and Connecticut, Hawaii, Iowa, New Mexico, and Utah.

Population  Women 30 years of age and older residing in the areas covered by the 9 SEER registries.

Main Outcome Measures  Proportional changes in incidence rates of invasive lobular and ductal carcinoma among women with no prior history of breast cancer.

Results  A total of 190 458 women were included in this analysis who were identified through the registries as having invasive breast cancer; 7682 of the 198 140 potentially eligible women (ie, those identified as not having in situ breast cancer) were excluded from this analysis because stage of cancer was unknown. Invasive breast cancer incidence rates adjusted for age and for SEER historic stage increased 1.04-fold (95% confidence interval [CI], 1.004-1.07) from 1987-1999 (206.7/100 000 to 214.1/100 000, age-adjusted). However, incidence rates of tumors classified as lobular increased 1.52-fold (95% CI, 1.42-1.63), and those classified as mixed ductal-lobular increased 1.96-fold (95% CI, 1.80-2.14); rates of these types combined increased 1.65-fold (95% CI, 1.55-1.78) (19.8/100 000 to 33.4/100 000, age-adjusted). In contrast, ductal carcinoma rates remained largely constant (153.8/100 000 to 155.3/100 000, age-adjusted; proportional change, 1.03 [95% CI, 0.99-1.06]). The proportion of breast cancers with a lobular component increased from 9.5% in 1987 to 15.6% in 1999.

Conclusions  Ductal carcinoma incidence rates remained essentially constant from 1987-1999 while lobular carcinoma rates increased steadily. This increase presents a clinical challenge given that lobular carcinoma is more difficult to detect than ductal carcinoma by both physical examination and mammography.

The Collaborative Group on Hormonal Factors in Breast Cancer has reported that current users of combined estrogen and progestin hormone replacement therapy (CHRT) or progestin alone for 5 years or longer have a 53% increase in risk of breast cancer.¹ Similarly, the Women's Health Initiative (WHI), a randomized controlled trial, found that CHRT use is associated with a statistically nonsignificant 26% increase in risk of breast cancer after 5.2 years of follow-up.² However, a growing number of studies report that the risk associated with use of CHRT differs by histological type. Specifically, 5 separate studies have shown that ever use and current use of CHRT are associated with 2.0-fold to 3.9-fold increased risks of invasive lobular carcinoma (ILC), the second most common histological type of breast cancer, but have little impact on risk of the most common histological type, invasive ductal carcinoma (IDC).^3-7 These same 5 studies also found that use of unopposed estrogen hormone replacement therapy was not strongly associated with risk of either ILC or IDC. Two of these studies had the power to assess duration of use, and both found that risk of ILC increased as duration of CHRT use increased.^5,7 Taken as a whole, these data suggest that ILC is more hormonally responsive than IDC.

Beyond their etiologic importance these results also have clinical significance because CHRT use increased steadily in the United States from the 1970s to the 1990s^8-12 and because ILC and IDC have different clinical features. For example, ILC is more likely to be hormone receptor–positive¹³ and to have a better prognosis than IDC,¹⁴ though it is more difficult to detect by mammography and clinical breast examination.^15-17 This is because ILC cells tend to grow in sheets rather than solid masses and to present as subtle thickening of the breast.¹⁸ Trends in incidence rates differ by histological type; we previously reported that incidence rates of ILC increased steadily from 1987 through 1995, while those of IDC remained relatively constant.¹⁹

The Surveillance, Epidemiology, and End Results (SEER) data from the National Cancer Institute are now available through 1999, allowing us to update our original analysis. This update is warranted because CHRT use continues to be widely prevalent among women who have not had a hysterectomy. For example, among postmenopausal women younger than 65 years who served as controls in a large recent US multicenter population-based case-control study, 45.5% had ever used CHRT for 6 months or longer.⁷ Additionally, our original study did not report on invasive mixed ductal-lobular carcinoma (IDLC), which is another histology with lobular features that has since been linked to CHRT use.^5,7

There were 190 458 women aged 30 years and older with no prior history of breast cancer included in the analysis herein having diagnoses of invasive breast cancer reported from 1987 through 1999 to 9 population-based cancer registries in the United States that participate in the SEER program. There were 7682 cases (3.9% of the 198 140 total potentially eligible women [ie, those identified as not having in situ breast cancer]) categorized as having an unknown stage that were excluded from this analysis. We included the registries for Atlanta, Ga; Detroit, Mich; San Francisco-Oakland, Calif; Seattle, Wash; and for Connecticut, Hawaii, Iowa, New Mexico, and Utah.²⁰ The standard for ascertainment of cases of cancer in the SEER registries is 98%.²¹ Individual patient medical records are the source of SEER data on patient and tumor characteristics. Informed consent was not obtained from women included in this study because by law cancer is a reportable disease, and because release of information regarding all diagnoses made, including all of the data collected by SEER, is mandatory in the areas covered by each of the SEER registries. Personal identifying information, such as names, addresses, and Social Security numbers, are not available from the SEER public use database and were not used or accessed in the conduct of this study. In general, the populations covered by SEER are representative of the whole United States with regard to socioeconomic status and education level, though they do include higher proportions of people who live in urban areas and who are foreign born.²²

Numbers and age distribution estimates of registry populations involved data from the US Bureau of the Census. For these estimates, SEER*Stat 4.2 statistical software (National Cancer Institute, Bethesda, Md) was used to calculate incidence rates of breast cancer among women with no prior history of the disease. The incidence rates were age-adjusted to the 2000 US population, and annual rates from 1987 through 1999 were compiled. The year 1987 was chosen as the starting point for this analysis because we reported previously that, while IDC incidence rates increased steadily from 1977 through 1987, these rates have plateaued since 1987.¹⁹ In addition to considering women of all ages together, we analyzed 6 age groups separately: 30-39, 40-49, 50-59, 60-69, 70-79, and 80 years or older. Women younger than 30 years were excluded from this study because breast cancer is rare among these women. Rates among 4 different histological groups of breast cancer cases were assessed, including IDC, ILC, IDLC, and all histologies combined. Histological classifications were based on International Classification of Disease for Oncology codes (ILC, 8520; IDC, 8500; IDLC, 8522). We estimated linear trends in incidence rates by histological category from 1987 through 1999, including proportional changes (ie, the proportions that rates changed [fold change] over time) and associated 95% confidence intervals (CIs), using negative binomial regression.²³ All proportional change analyses were adjusted for age and for SEER historic stage (localized, regional, or distant disease) because of variations in detection of stage of disease over time.

From 1987 through 1999, 190 458 invasive breast cancer cases with a known stage were reported to the 9 SEER registries. Of these cases, 138 625 were IDC (72.8%), 14 486 were ILC (7.6%), and 8860 were IDLC (4.7%). Overall, invasive breast cancer incidence rates increased 1.04-fold (95% CI, 1.004-1.07) from 1987 through 1999 (from 206.7/100 000 to 214.1/100 000) (Table 1). The incidence rates of ILC and IDLC increased 1.52-fold (95% CI, 1.42-1.63) and 1.96-fold (95% CI, 1.80-2.14), respectively, from 1987 through 1999. When these 2 histological groups were combined, the observed increase was 1.65-fold (95% CI, 1.55-1.78) (from 19.8/100 000 to 33.4/100 000). The proportion of invasive breast cancers with a lobular component (either ILC or IDLC) increased from 9.5% in 1987 to 15.6% in 1999. Alternatively, rates of IDC increased only slightly during this same time period (from 153.8/100 000 to 155.3/100 000; proportional change, 1.03 [95% CI, 0.99-1.06]). These trends were also observed within each of the individual SEER registries except for Hawaii, where rates of ILC remained essentially constant (C.I.L., unpublished data, January 2003).

Increases in incidence rates of ILC and IDLC were observed throughout the age range, but were most pronounced for women 50 years of age and older, with increases ranging from 1.45-fold to 1.76-fold and from 1.96-fold to 2.75-fold among these older women, respectively (Table 2). Incidence rates of IDC did increase 1.26-fold among 50 through 59-year-olds, but no increases in rates of IDC were observed among women 60 years of age and older.

Our results demonstrate that, in a sample of the US population drawn from 9 cancer registries, incidence rates of ILC continue to rise while rates of IDC continue to remain largely constant through 1999. Furthermore, we document that incidence rates of IDLC, which represents a heterogeneous group of tumors that consist of varying proportions of lobular and ductal cancer cells, are also increasing. As a result, the proportion of all breast cancers with a lobular component rose steadily from 9.5% in 1987 to 15.6% in 1999. Though rates of ILC and IDLC increased throughout the age range, these increases were most pronounced among women 50 years of age and older. While we could not directly address CHRT use in this national registry database, CHRT use has risen over the past several decades,^8-12 and our data are consistent with the hypothesis that CHRT use is associated with an increased risk of ILC.

A potential limitation of this study was that the histological categories used were based on diagnoses made by multiple pathologists in multiple institutions, and diagnostic criteria may vary somewhat by both individual pathologists and institutions resulting in a certain degree of misclassification error. We are unaware of any data demonstrating changes in clinical and biological features of ILC over time; however, studies addressing this issue are needed. Studies evaluating the concordance between classification of tumors as ILC or IDC ascertained by SEER registries and those made through a centralized pathology review are needed to quantify the magnitude of this misclassification, as none have been reported in the literature. However, our study was limited to a recent finite period when major changes in the diagnostic criteria for lobular carcinomas did not occur, and pathological criteria defining ILC were reasonably well established and well defined.

Although ILC is more difficult to detect by mammography than is IDC,^15-17 we cannot rule out a systematic surveillance bias that may have occurred over time as a result of improvements in the skills of radiologists to identify ILC tumors on mammograms and in the accuracy and quality of mammographic technology. However, studies evaluating this possibility are lacking. It has been recently shown, though, that additional radiographic imaging tools, including ultrasound^24,25 and magnetic resonance imaging,^26,27 are more effective than mammography in detecting ILC, but these tools are not widely used and thus are unlikely to have had an impact on our observations. Surveillance bias also could result from increased scrutiny of mammograms performed on women receiving HRT. However, this is unlikely to be an important contributor to the trends that we observed. One of the first studies identifying HRT as a potential risk factor for breast cancer among US women was published in 1980²⁸ (shortly followed by numerous subsequent studies), and because this was well before 1987, the earliest year included in our study, radiologists would have been cognizant of this association throughout our study's duration.

Results from the WHI have demonstrated a clear association between use of CHRT and an increased risk of breast cancer (though this increase was statistically nonsignificant).² In addition, data from several recent case-control studies suggest that CHRT is more strongly associated with an increased risk of tumors with lobular histologies (both ILC and IDLC) than it is with an increased risk of ductal tumors.^3-7 Our findings are consistent with these observations because the use of CHRT has increased in the United States at the same time that the incidence rates of ILC and IDLC have increased.^8-12 We encourage other investigators assessing the relationship between CHRT use and breast cancer to stratify their analyses by histological type, as a greater understanding of the etiology and epidemiology of lobular breast cancer is of growing importance. The rising incidence rates of lobular breast tumors also present a clinical challenge because these tumors are more difficult to detect both by clinical examination and by mammography.^15-17

Corresponding Author: Christopher Li, MD, PhD, Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, MP 381, PO Box 19024, Seattle, WA 98109-1024 (e-mail: cili@fhcrc.org).

Author Contributions:Study concept and design: Li, Anderson.

Acquisition of data: Li.

Analysis and interpretation of data: Li, Anderson, Daling, Moe.

Drafting of the manuscript: Li, Anderson.

Critical revision of the manuscript for important intellectual content: Li, Anderson, Daling, Moe.

Statistical expertise: Li.

Obtained funding: Daling.

Administrative, technical, or material support: Daling.

Study supervision: Li, Anderson, Daling.

Funding/Support: This study was supported in part by National Cancer Institute (NCI) grants T32 CA 09168 and R01 CA 85913. The data set used for this analysis was obtained from the Surveillance, Epidemiology, and End Results (SEER) Program Public-Use Data (1973-1999), NCI, Department of Cancer Control and Population Sciences, Surveillance Research Program, Cancer Statistics Branch, released April 2002, based on the November 2001 submission of data by each of the registries. The NCI and the SEER program were not involved in the design and conduct of this study, in the analysis or interpretation of the data, or in the preparation and review of this article.

Financial Disclosures: Dr Li has received funding from the National Cancer Institute and from a private foundation (Life Possibilities Fund) for cancer research studies. Dr Anderson has provided expert legal testimony involving breast cancer and is a consultant for Calypso, a company developing cancer diagnosis technology. Dr Daling has received an honorarium for attending a scientific symposium on hormones and health sponsored by Wyeth in 2001.