Background:

In absence of definitive molecular risk markers, clinical management of patients diagnosed with ductal carcinoma in situ (DCIS) remains largely guided by patient and tumor characteristics. In this study, we analyzed recent trends in DCIS incidence and compared them against trends in mammography use.

Methods:

The Surveillance, Epidemiology, and End Results registry was queried for patients diagnosed with DCIS from 2000 to 2014 (18 registries). Joinpoint regression analyses were used to compute age- and race-stratified trends in age-adjusted incidence of DCIS. The patterns of DCIS incidence were compared against mammography utilization data from the National Health Interview Survey.

Results:

Between 2000 and 2014, overall DCIS incidence in the U.S. population was stable (P = 0.24). Among age groups 20 to 44 years and 45 to 55 years, DCIS incidence increased by 1.3% (P = 0.001) and 0.6% (P = 0.02) per year, respectively. Although stable among white women, DCIS incidence increased among black women and women of other races by 1.6% (P < 0.001) and 1.0% (P = 0.002) per year, respectively. Mammography uptake correlated well with DCIS incidence, with the exception of women ages 40 to 49 years and black women who experienced an increase in DCIS incidence despite stagnating and decreasing mammography uptake, respectively.

Conclusions:

Overall DCIS incidence rates have remained stable between 2000 and 2014. However, subgroup analyses revealed an increase in incidence among both younger women and black women.

Impact:

DCIS incidence trends did not correlate with the mammography uptake patterns, suggesting that etiologic factors other than screening may be leading to an increased DCIS incidence in these groups.

This article is featured in Highlights of This Issue, p. 1273

The introduction of mammographic screening in the early 1980s has led to a dramatic increase in the detection of ductal carcinoma in situ (DCIS) over the past three decades (1, 2). In the United States, more than 50,000 women are diagnosed with DCIS each year and DCIS accounts for an estimated 18%–25% of the total number of newly diagnosed breast tumors (3).

Consisting of a heterogeneous group of preinvasive neoplastic lesions, DCIS is primarily characterized by clonal proliferation of malignant cells that are confined to the lumen of breast ducts and lobules (4). The phenotypic spectrum of DCIS ranges from indolent and slowly growing disease to fast growing, aggressive subtypes that can quickly invade the surrounding stroma and metastasize to distant organs (5, 6). However, due to a lack of reliable prognostic markers that delineate the different risk groups, the vast majority of patients undergo invasive treatment in the form of surgery, either alone or in combination with radiation and hormone therapy (7). To enable effective and personalized management of patients with DCIS, there is an ongoing effort to identify clinical and biologic markers that predict the propensity to progress to invasive cancer (8). As a more concise picture of patient-specific risk profiles is bound to emerge, it becomes necessary to characterize the potential impact of risk-stratified management strategies at the population level. In particular, a better understanding of subtype-specific incidence patterns by age and race will inform cost-effectiveness analyses and the estimation of overdiagnosis rates in the population.

In this study, we analyzed data from the U.S. population–based cancer registry Surveillance, Epidemiology and End Results (SEER) to characterize recent trends in (i) DCIS incidence by age and race, and (ii) the distribution of disease subtypes among patients diagnosed with DCIS. Because DCIS is a primarily screen-detected disease, we further compared the incidence trends against uptake of mammography screening.

SEER data

Women diagnosed with DCIS in the United States between 2000 and 2014 were identified through the population-based cancer registries that participate in the NCI's SEER Program, which represents approximately 28% of the U.S. population (9). Women ages 20 years or older were included if diagnosed with behavior code ICD-O-3 “in situ” and one of following histology codes for DCIS: 8050, 8140, 8200, 8201, 8211, 8230, 8246, 8260, 8343, 8401, 8480, 8481, 8490, 8500, 8501, 8503, 8504, 8507, 8508, 8521, 8522, 8523, 8543, or 8550. Patients without microscopic confirmation of the diagnosis, those identified at autopsy or on death certificate only, and patients for whom DCIS was not the first cancer diagnosis were excluded from analyses. Ethics approval for this study was obtained from Duke University (Durham, NC).

Patient-specific variables included age at diagnosis (20–44 years, 45–54 years, 55–69 years, and 70 years or older), race (white, black, other, unknown), and year of diagnosis (2000–03, 2004–07, 2009–11, and 2012–14). Tumor-specific variables included nuclear grade (low, intermediate, high, unknown), tumor size (≤10 mm, 11–30 mm, >30 mm, unknown), estrogen receptor (ER), and progesterone receptor (PR) status (positive, negative, unknown). In addition to race- and age-specific subgroups, we further defined a subgroup of “low-risk” patients with DCIS based on the eligibility criteria of the COMET active surveillance trial (10). Patients were included in this subgroup if they were 40 years or older at diagnosis and had an ER- or PR-positive lesion of low or intermediate nuclear grade.

To evaluate the long-term relationship between mammography screening and DCIS incidence, we compared crude mammogram rates (see below) to the corresponding crude DCIS incidence rates over the extended interval from 1992 to 2014 (SEER, 13 registries).

Mammography data

Mammography utilization data for women ages 40 years or older was obtained from the National Health Interview Survey (NHIS; ref. 11). More precisely, we used the fraction of women who self-reported having had a mammogram in the two years prior to the survey round as a proxy for mammogram uptake. To include a historical perspective on mammography uptake data we included survey rounds from 1987, 1993, 1994, 2000, 2005, 2008, 2010, 2013, and 2015. The age-unadjusted (crude) rates were further stratified by age category (40–49, 50–64, and ≥65 years) and race (black and white).

Statistical analyses

SEER*Stat software version 8.34 (NCI, Bethesda, MD, released March 22, 2017) was used to calculate age-adjusted DCIS incidence rates from 2000 to 2014 using the SEER 18 database. Rates were age-adjusted to the 2000 U.S. standard population. The magnitude and direction of DCIS incidence trends by grade, ER status, PR status and tumor size were evaluated using the Joinpoint Regression Program version 4.5.0.1 (NCI, Bethesda, MD, released June 2017). This program uses permutation analysis to fit a series of joined straight lines on a logarithmic scale to observe rates and estimate the annual percent change (APC). The study period was divided into intervals 2000–2003, 2004–2007, 2008–2011, and 2012–2014, and associations between time intervals and patient and tumor characteristics were evaluated using χ2 tests. Trends in the proportion of specific tumor subtypes between 2000 and 2014 were evaluated using Cochrane–Armitage tests.

Crude screening rates from the NHIS survey (1987–2015) were plotted alongside crude DCIS incidence rates (SEER 13, 1992–2014), stratified by race and age groups, and trends in crude DCIS incidence and mammography uptake rates were evaluated using Kendall tau correlation test, separately for the time periods 1992–1999 and 2000–2014.

All P values were calculated as two-sided, with statistical significance declared for P values below 0.05. Unless otherwise stated, statistical analyses were performed in statistical software SAS version 9.4 (SAS Institute Inc).

Interactive visualizations

Trends in age-adjusted DCIS incidence for different combinations of age, race, and tumor features (nuclear grade, tumor size, ER, and PR status) are presented in an online supplementary in the form of an interactive web-page. The visualizations were created using Tableau software (version 9.1, Tableau Software Inc): https://public.tableau.com/profile/yiling.liu#!/vizhome/tableau_new_0/Story1?publish=yes.

Incidence and mortality of colorectal, cervical, and female breast cancer

To provide a broader context for our findings, we compared the US incidence and mortality rates of female breast, colorectal, and cervical cancers between 1999 and 2015. Annual incidence and mortality rates were obtained from the U.S. Cancer Statistics Data Visualizations Tool (12). Relative rates were obtained by rescaling the absolute rates with respect to the corresponding baseline rates from 1999.

Between 2000 and 2014, 145,670 women with DCIS met the inclusion criteria. Demographic and clinicopathologic characteristics of the study population are summarized in Table 1. The majority of diagnoses were recorded in white women (n = 113,396, 77.8%), followed by black women (15,712, 10.8%), and women of other races (15,112, 10.4%).

Table 1.

Trends in patient and tumor features

Total (N = 145,670)2000–03 (N = 34,554)2004–07 (N = 37,188)2008–11 (N = 41,783)2012–2014 (N = 32,145)
n (%)n (%)n (%)n (%)n (%)Pa
Age      <0.001 
 20–44 years 17,860 (12.3) 4,503 (13) 4,725 (12.7) 4,999 (12) 3,633 (11.3)  
 45–54 years 40,512 (27.8) 9,534 (27.6) 10,710 (28.8) 11,767 (28.2) 8,501 (26.4)  
 55–69 years 56,512 (38.8) 12,510 (36.2) 13,782 (37.1) 16,554 (39.6) 13,666 (42.5)  
 ≥70 years 30,786 (21.1) 8,007 (23.2) 7,971 (21.4) 8,463 (20.3) 6,345 (19.7)  
Race      <0.001 
 White 113,396 (77.8) 28,125 (81.4) 29,440 (79.2) 32,084 (76.8) 23,747 (73.9)  
 Black 15,712 (10.8) 3,222 (9.3) 3,897 (10.5) 4,647 (11.1) 3,946 (12.3)  
 Other 15,112 (10.4) 2,978 (8.6) 3,561 (9.5) 4,609 (11.0) 3,964 (12.0)  
 Unknown 1,450 (1.0) 229 (0.7) 290 (0.8) 443 (1.1) 488 (1.5)  
ER status      <0.001 
 Positive 83,168 (57.1) 4,857 (14.1) 21,276 (57.2) 30,865 (73.9) 26,170 (81.4)  
 Negative 14,592 (10.0) 1,198 (3.5) 4,521 (12.2) 5,125 (12.3) 3,748 (11.7)  
 Unknown 47,910 (32.9) 28,499 (82.5) 11,391 (30.6) 5,793 (13.9) 2,227 (6.9)  
PR status      <0.001 
 Positive 68,961 (47.3) 3,909 (11.3) 17,314 (46.6) 25,676 (61.5) 22,062 (68.6)  
 Negative 22,783 (15.6) 1,731 (5) 6,732 (18.1) 8,212 (19.7) 6,108 (19)  
 Unknown 53,926 (37.0) 28,914 (83.7) 13,142 (35.3) 7,895 (18.9) 3,975 (12.4)  
Tumor grade      <0.001 
 Grade I 17,359 (11.9) 3,854 (11.2) 4,282 (11.5) 5,081 (12.2) 4,142 (12.9)  
 Grade II 50,862 (34.9) 10,301 (29.8) 12,553 (33.8) 15,450 (37) 12,558 (39.1)  
 Grade III/IV 55,060 (37.8) 11,848 (34.3) 14,803 (39.8) 16,177 (38.7) 12,232 (38.1)  
 Unknown 22,389 (15.4) 8,551 (24.7) 5,550 (14.9) 5,075 (12.1) 3,213 (10)  
Tumor size      <0.001 
 1–10 mm 49,926 (34.3) 9,865 (28.6) 12,612 (33.9) 15,305 (36.6) 12,144 (37.8)  
 11–30 mm 37,351 (25.6) 7,510 (21.8) 9,027 (24.3) 11,038 (26.4) 9,776 (30.4)  
 >30 mm 13,162 (9.0) 2,204 (6.4) 3,055 (8.2) 4,156 (9.9) 3,747 (11.7)  
 Unknown 45,231 (31.1) 14,891 (43.2) 12,494 (33.6) 11,284 (27) 6,478 (20.2)  
Total (N = 145,670)2000–03 (N = 34,554)2004–07 (N = 37,188)2008–11 (N = 41,783)2012–2014 (N = 32,145)
n (%)n (%)n (%)n (%)n (%)Pa
Age      <0.001 
 20–44 years 17,860 (12.3) 4,503 (13) 4,725 (12.7) 4,999 (12) 3,633 (11.3)  
 45–54 years 40,512 (27.8) 9,534 (27.6) 10,710 (28.8) 11,767 (28.2) 8,501 (26.4)  
 55–69 years 56,512 (38.8) 12,510 (36.2) 13,782 (37.1) 16,554 (39.6) 13,666 (42.5)  
 ≥70 years 30,786 (21.1) 8,007 (23.2) 7,971 (21.4) 8,463 (20.3) 6,345 (19.7)  
Race      <0.001 
 White 113,396 (77.8) 28,125 (81.4) 29,440 (79.2) 32,084 (76.8) 23,747 (73.9)  
 Black 15,712 (10.8) 3,222 (9.3) 3,897 (10.5) 4,647 (11.1) 3,946 (12.3)  
 Other 15,112 (10.4) 2,978 (8.6) 3,561 (9.5) 4,609 (11.0) 3,964 (12.0)  
 Unknown 1,450 (1.0) 229 (0.7) 290 (0.8) 443 (1.1) 488 (1.5)  
ER status      <0.001 
 Positive 83,168 (57.1) 4,857 (14.1) 21,276 (57.2) 30,865 (73.9) 26,170 (81.4)  
 Negative 14,592 (10.0) 1,198 (3.5) 4,521 (12.2) 5,125 (12.3) 3,748 (11.7)  
 Unknown 47,910 (32.9) 28,499 (82.5) 11,391 (30.6) 5,793 (13.9) 2,227 (6.9)  
PR status      <0.001 
 Positive 68,961 (47.3) 3,909 (11.3) 17,314 (46.6) 25,676 (61.5) 22,062 (68.6)  
 Negative 22,783 (15.6) 1,731 (5) 6,732 (18.1) 8,212 (19.7) 6,108 (19)  
 Unknown 53,926 (37.0) 28,914 (83.7) 13,142 (35.3) 7,895 (18.9) 3,975 (12.4)  
Tumor grade      <0.001 
 Grade I 17,359 (11.9) 3,854 (11.2) 4,282 (11.5) 5,081 (12.2) 4,142 (12.9)  
 Grade II 50,862 (34.9) 10,301 (29.8) 12,553 (33.8) 15,450 (37) 12,558 (39.1)  
 Grade III/IV 55,060 (37.8) 11,848 (34.3) 14,803 (39.8) 16,177 (38.7) 12,232 (38.1)  
 Unknown 22,389 (15.4) 8,551 (24.7) 5,550 (14.9) 5,075 (12.1) 3,213 (10)  
Tumor size      <0.001 
 1–10 mm 49,926 (34.3) 9,865 (28.6) 12,612 (33.9) 15,305 (36.6) 12,144 (37.8)  
 11–30 mm 37,351 (25.6) 7,510 (21.8) 9,027 (24.3) 11,038 (26.4) 9,776 (30.4)  
 >30 mm 13,162 (9.0) 2,204 (6.4) 3,055 (8.2) 4,156 (9.9) 3,747 (11.7)  
 Unknown 45,231 (31.1) 14,891 (43.2) 12,494 (33.6) 11,284 (27) 6,478 (20.2)  

aχ2 test.

DCIS incidence trends

While the overall DCIS incidence rate remained stable between 2000 and 2014 (P = 0.24; Fig. 1A), significant changes were observed for specific subgroups. When stratified by race (Fig. 1C), DCIS incidence rates increased among black women and women of other races by 1.6% (P < 0.001) and 1.0% (P = 0.002) per year, respectively. In an age-stratified analysis (Fig. 1B), incidence rates increased among women of ages 20–44 years and 45–54 years by 1.3% (P = 0.001) and 0.6% (P = 0.02) per year, respectively. Among women of ages 55–69 years, the incidence rates trended downward by 0.3% per year (P = 0.08).

Figure 1.

Age-adjusted DCIS incidence trends, overall and by age and race. APC in the DCIS incidence rate in the United States, 2000–2014, overall (A), age group (B), and by race (C). Asterisk (*) indicates that the APC is statistically significantly different from zero (P < 0.05). To highlight trends, the scales of the y-axes vary.

Figure 1.

Age-adjusted DCIS incidence trends, overall and by age and race. APC in the DCIS incidence rate in the United States, 2000–2014, overall (A), age group (B), and by race (C). Asterisk (*) indicates that the APC is statistically significantly different from zero (P < 0.05). To highlight trends, the scales of the y-axes vary.

Close modal

DCIS tumor characteristics at diagnosis

The proportions of incident DCIS cases with unknown grade, unknown ER and PR status, and unknown size decreased drastically between 2000 and 2005, and then continued to decrease at a slower rate until 2014 (Supplementary Fig. S1). During the period 2000–2003, 82.5% of women had unknown ER status, 83.7% had unknown PR status, 24.7% had unknown grade, and 38% had unknown tumor size (Table 1). By 2012–2014, these proportions had decreased to 6.9%, 12.4%, 10.0%, and 19.0%, respectively.

Among newly diagnosed cases with known tumor characteristics, the distributions of nuclear grade, tumor size, and hormone receptor status underwent significant changes (Fig. 2 and online materials). With respect to hormone receptor status, the proportions of both ER- and PR-positive tumors increased significantly (P < 0.001 for both; Fig. 2A). While there was no measurable trend among grade I tumors, the proportions of grade II and grade III tumors increased and decreased, respectively (P < 0.001; Fig. 2B). Finally, the proportion of small tumors (1–10 mm) decreased (P < 0.001), while the proportion of tumors larger than 30 mm increased (P < 0.001); there was no measurable trend among tumors of size 11–20 mm (Fig. 2C). Changes in tumor characteristics for different age and race groups are found in Fig. 3 and online materials. While the trends among white women (Fig. 3A) mirrored those of the overall population (Fig. 2), differential trends were found in black women (Fig. 3B). First, the proportion of grade I tumors decreased among black women (P < 0.05) while being stable among white women. Second, there were no noticeable trends in the tumor size distributions among black women, in contrast to the observed shift from smaller (≤1 cm) to larger (>3 cm) tumors in white women.

Figure 2.

Trends in tumor characteristics among women diagnosed with DCIS. Proportions of women diagnosed with ER-positive and PR-positive DCIS (A), grade I, II, and III DCIS (B), and small (1–10 mm), intermediate (11–30 mm), and large (>30 mm) DCIS (C) in the United States, 2000–2014. Only cases with known tumor features are included in the analyses.

Figure 2.

Trends in tumor characteristics among women diagnosed with DCIS. Proportions of women diagnosed with ER-positive and PR-positive DCIS (A), grade I, II, and III DCIS (B), and small (1–10 mm), intermediate (11–30 mm), and large (>30 mm) DCIS (C) in the United States, 2000–2014. Only cases with known tumor features are included in the analyses.

Close modal
Figure 3.

Trends in tumor characteristics among women diagnosed with DCIS; subgroup analyses. Proportions of women diagnosed with ER-positive and progesterone (PR)-positive DCIS, grade I, II, and III DCIS, and small (1–10 mm), intermediate (11–30 mm), large (>30 mm) DCIS among white women (A) and black women (B), in the United States, 2000–2014. Only cases with known tumor features are included in the analyses.

Figure 3.

Trends in tumor characteristics among women diagnosed with DCIS; subgroup analyses. Proportions of women diagnosed with ER-positive and progesterone (PR)-positive DCIS, grade I, II, and III DCIS, and small (1–10 mm), intermediate (11–30 mm), large (>30 mm) DCIS among white women (A) and black women (B), in the United States, 2000–2014. Only cases with known tumor features are included in the analyses.

Close modal

During the most recent time period from 2012 to 2014 (Table 2; Supplementary Table S1), 14.3% of patients with DCIS with known tumor grade had a grade I lesion, 43.4% a grade II lesion, and 42.3% a grade III lesion. Among women with known ER status, 87.5% had an ER-positive lesion and 12.5% an ER-negative lesion. Among lesions of known size, 46.6% were 1–10 mm, 37.5% were 11–30 mm, 14.4% were larger than 30 mm, and 1.5% were microinvasive. Among patients with known tumor grade and ER or PR status, the proportion of women who satisfied the criteria of the low-risk subgroup (age ≥ 40 years with grade I/II and ER- or PR-positive tumor) increased from 47.2% (2000–2004) to 52.5% (2012–2014; P < 0.001).

Table 2.

Tumor features in patients diagnosed between 2012–2014

n (%)
Known tumor grade (n = 28,932) 
 Grade I 4,142 (14.3) 
 Grade II 12,558 (43.4) 
 Grade III/IV 12,232 (42.3) 
Known ER Status (n = 29,918) 
 ER-positive 26,170 (87.5) 
 ER-negative 3,748 (12.5) 
Known PR Status (n = 28,170) 
 PR-positive 22,062 (78.3) 
 PR-negative 6,108 (21.7) 
Known tumor size (n = 25,667) 
 1–10 mm 12,144 (47.3) 
 11–30 mm 9,776 (38.1) 
 >30 mm 3,747 (14.6) 
Known tumor grade, ER and/or PR Statusa (n = 26,552) 
 Low riskb 14,762 (55.6) 
 Non-low risk 11,790 (44.4) 
n (%)
Known tumor grade (n = 28,932) 
 Grade I 4,142 (14.3) 
 Grade II 12,558 (43.4) 
 Grade III/IV 12,232 (42.3) 
Known ER Status (n = 29,918) 
 ER-positive 26,170 (87.5) 
 ER-negative 3,748 (12.5) 
Known PR Status (n = 28,170) 
 PR-positive 22,062 (78.3) 
 PR-negative 6,108 (21.7) 
Known tumor size (n = 25,667) 
 1–10 mm 12,144 (47.3) 
 11–30 mm 9,776 (38.1) 
 >30 mm 3,747 (14.6) 
Known tumor grade, ER and/or PR Statusa (n = 26,552) 
 Low riskb 14,762 (55.6) 
 Non-low risk 11,790 (44.4) 

aRestricted to age ≥ 40 years.

bGrade I/II, ER-positive, and/or PR-positive.

Mammography utilization and DCIS incidence

To provide historical context, we performed a comparison of mammography utilization and DCIS incidence over an extended time period. Prior to 2000, both DCIS incidence and self-reported mammography uptake increased among all age groups and races (Fig. 4; P < 0.01 for all subgroups). After 2000, more complex patterns emerged. Among women ages 40–49 years (Fig. 4A), DCIS incidence continued to increase (P = 0.015), while mammography uptake decreased (P = 0.01). Among women ages 50–64 (Fig. 4B), both mammography uptake and DCIS incidence decreased (P = 0.01 and P = 0.004). Screening uptake and DCIS incidence were stable among women of age 65 years and older (Fig. 4C). Among black women of all ages (Fig. 4D), we found that the DCIS incidence continued to increase during the period 2000–2014 (P < 0.0001) despite unchanged mammogram use (P = 0.9). Among white women, mammogram use decreased (P = 0.006) yet there was no accompanying downward trend in DCIS incidence (P = 0.2).

Figure 4.

Mammogram utilization (1987–2015) and DCIS incidence (1992–2014), by age groups and race. Crude mammogram utilization (NIHS data) and DCIS incidence (SEER 13) rates among women of all races ages 40–49 years (A), 50–64 years (B), and greater than 65 years (C), and for all ages by race (D). Scales for DCIS incidence and mammogram utilization on the left and right sides of the graphs, respectively.

Figure 4.

Mammogram utilization (1987–2015) and DCIS incidence (1992–2014), by age groups and race. Crude mammogram utilization (NIHS data) and DCIS incidence (SEER 13) rates among women of all races ages 40–49 years (A), 50–64 years (B), and greater than 65 years (C), and for all ages by race (D). Scales for DCIS incidence and mammogram utilization on the left and right sides of the graphs, respectively.

Close modal

Incidence and mortality of colorectal, cervical, and female breast cancers

The relative U.S. incidence and mortality rates of colorectal, cervical, and female breast cancers between 1999 and 2015 are shown in Fig. 5. While mortality rates consistently decreased in all three cancer sites, a concurrent decrease in incidence was only observed in colorectal and cervical cancers; the incidence rate of female breast cancers remained unchanged after 2005.

Figure 5.

Incidence and mortality rates of female breast, cervical, and colorectal (male and female) cancers in the United States, 1999–2015. Relative rates are obtained by rescaling the actual rates with respect to the baseline rates in 1999. Data extracted from the U.S. Cancer Statistics Data Visualizations Tool (12).

Figure 5.

Incidence and mortality rates of female breast, cervical, and colorectal (male and female) cancers in the United States, 1999–2015. Relative rates are obtained by rescaling the actual rates with respect to the baseline rates in 1999. Data extracted from the U.S. Cancer Statistics Data Visualizations Tool (12).

Close modal

It is well established that widespread mammographic screening has resulted in a marked increase in DCIS detection since the mid-1980s (13–15). In this study, we found that the overall incidence rate of DCIS has now stabilized between 2000 and 2014. However, we noted an unanticipated discrepancy between breast cancer screening and incidence rates of DCIS in some groups. NHIS survey data have shown that between 2000 and 2008, the age-adjusted rate of breast cancer screening among women ages 40–49 and 50–64 years fell by 2.7% and 4.4%, respectively (16). Paradoxically, our analysis shows that the incidence rates of DCIS have in fact increased among women of ages 20–44 years and 45–54 years (1.3% and 0.6% per year, respectively; P < 0.05 for both). The increase in DCIS incidence among women ages 20–44 is especially notable, as most of these women are younger than the age at which screening is recommended. Moreover, although screening rates have not markedly changed among blacks, the DCIS incidence rate is disproportionately increasing in this group as well (1.6% annually; P < 0.05). This divergence between DCIS incidence rate and breast cancer screening rate has not been previously observed and suggests an effect in some groups of either endogenous or environmental risk factors for DCIS that have not yet been recognized or identified (17, 18).

We found that the biologic characteristics of DCIS and relationship to screening prevalence could not be easily determined, as many DCIS features were not reported prior to 2000. Between 2000 and 2014, reporting of histopathologic tumor features in the cancer registry SEER increased substantially, at least partly due to evolving coding guidelines and practices of the cancer registry. In particular, the fraction of cases with unknown ER status decreased from 84% (2000–2003) to 7% (2012–2014). This change has been attributed to increased ER testing after FDA approval of tamoxifen for DCIS in 2000 (19).

However, among cases with known tumor features, we noted changes in the distribution of tumor size, grade, and hormone receptor status. As there is little evidence to suggest a reporting bias, these observations likely represent real underlying trends. The most marked change was noted in hormone receptor status, with an increase in the fraction of ER-positive DCIS from 80% in 2000 to 87% in 2014. In contrast, it has been reported that the incidence rate of ER-positive invasive cancers fell during this period, while the incidence of ER-negative invasive cancers did not change significantly (20). This finding has been widely attributed to a lower rate of exogenous hormone use resulting in fewer women being exposed to the proliferative effects of combined estrogen plus progestin on the breast epithelium (21). The disproportionate rise in ER-positive, and not ER-negative DCIS during a time interval when use of exogenous hormones dropped, again suggests that etiologic risk factors for DCIS and invasive cancer may differ.

Comparing screening prevalence and DCIS incidence against invasive breast cancer incidence and mortality provides insights into the respective roles of screening and treatment in prevention. Over the past decade, breast cancer mortality has dropped substantially, while invasive cancer incidence remained stable after 2003 (Fig. 5A). These trends, in conjunction with relatively stable screening uptake and DCIS incidence patterns (Figs. 1A and 4), make it difficult to attribute more than a negligible impact of DCIS detection and treatment on recent breast cancer mortality reductions and suggest a larger effect of treatment improvements on reducing breast cancer mortality (22). This observation stands in contrast to other cancers, such as cervical (Fig. 5B) and colorectal cancer (Fig. 5C), for which effective screening programs have successfully reduced both incidence and mortality.

A limitation of this study is the use of self-reported mammography utilization data. Indeed, prior studies have suggested that self-reported data overestimate the true utilization rate, and the degree of overestimation varies by age and race (23). However, because we used these data to estimate relative trends only, the impact of potential systematic reporting biases is unlikely to affect our conclusions. Another limitation is that the study populations for cancer incidence (SEER) and mammography utilization (NHIS) are not identical. This limitation is partially addressed by the fact that both the NHIS survey and SEER registry draw from samples that are representative of the U.S. population.

Finally, we note that during the most recent study period (2012–2014), the majority of DCIS lesions diagnosed were of low to intermediate nuclear grade (58%) and ER positive (87%). In particular, 55% of newly diagnosed patients with DCIS satisfied the three main criteria for low-risk DCIS as used by the ongoing COMET trial, which randomizes patients with DCIS to either usual care or active surveillance (10): age ≥ 40 years, nuclear grade I/II, and hormone receptor–positive. Therefore, should active surveillance become part of usual care, a sizable fraction of patients with DCIS may have the option to deescalate treatment and delay surgery unless there is progression to invasive cancer.

In an era of increasingly tailored treatment options for patients with cancer, it is essential to gain a clear understanding of the different disease subtypes and their incidence patterns by age and race groups. For this reason, we developed a web-based visualization tool that enables the user to explore the complex incidence patterns for DCIS during the period from 2000 to 2014 (see Methods section for details). Altogether, the results of this study data provide the basis for future cost-effectiveness studies and other model-based research activities for DCIS.

In conclusion, we observed an emerging divergence between DCIS incidence and mammographic screening patterns in some subgroups. Specifically, we found an increase in ER-positive DCIS and DCIS diagnosed in young patients and among African American individuals. Importantly, these trends did not correlate with the mammography uptake patterns, suggesting that etiologic factors other than screening may be leading to an increased DCIS incidence in these groups.

No potential conflicts of interest were disclosed.

Conception and design: M.D. Ryser, L.H. Hendrix, M. Worni, E.S. Hwang

Development of methodology: M.D. Ryser, L.H. Hendrix, M. Worni, E.S. Hwang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.H. Hendrix, E.S. Hwang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.D. Ryser, L.H. Hendrix, M. Worni, Y. Liu, T. Hyslop, E.S. Hwang

Writing, review, and/or revision of the manuscript: M.D. Ryser, L.H. Hendrix, M. Worni, T. Hyslop, E.S. Hwang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.H. Hendrix

Study supervision: M.D. Ryser, L.H. Hendrix, E.S. Hwang

This work was supported by the NIH (K99 CA207872 to M.D. Ryser; R01 CA185138-01 to E.S. Hwang; 5P30CA014236 to T. Hyslop), the National Science Foundation (DMS-1614838 to M.D. Ryser), the Patient-Centered Outcomes Research Institute (1505-30497 to E.S. Hwang and T. Hyslop), the Department of Defense (BC132057 to E.S. Hwang) and the Breast Cancer Research Fund (to E.S. Hwang). This work was also supported in part by NIH grants U01 CA152958 (to J. Mandelblatt) and U01 CA157224 (to R. Etzioni) to the Cancer Intervention and Surveillance Modeling Network Breast and Prostate Cancer Working Groups, respectively, for important background discussions relevant to the scientific areas covered in this article.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Bleyer
A
,
Welch
HG
. 
Effect of three decades of screening mammography on breast-cancer incidence
.
N Engl J Med
2012
;
367
:
1998
2005
.
2.
Ozanne
EM
,
Shieh
Y
,
Barnes
J
,
Bouzan
C
,
Hwang
ES
,
Esserman
LJ
. 
Characterizing the impact of 25 years of DCIS treatment
.
Breast Cancer Res Treat
2011
;
129
:
165
73
.
3.
Ward
EM
,
DeSantis
CE
,
Lin
CC
,
Kramer
JL
,
Jemal
A
,
Kohler
B
, et al
Cancer statistics: breast cancer in situ
.
CA Cancer J Clin
2015
;
65
:
481
95
.
4.
Allred
DC
. 
Ductal carcinoma in situ: terminology, classification, and natural history
.
J Natl Cancer Inst Monogr
2010
;
2010
:
134
8
.
5.
Roses
RE
,
Arun
BK
,
Lari
SA
,
Mittendorf
EA
,
Lucci
A
,
Hunt
KK
, et al
Ductal carcinoma-in-situ of the breast with subsequent distant metastasis and death
.
Ann Surg Oncol
2011
;
18
:
2873
8
.
6.
Ryser
MD
,
Weaver
DL
,
Zhao
F
,
Worni
M
,
Grimm
LJ
,
Gulati
R
, et al
Cancer outcomes in DCIS patients without locoregional treatment
.
J Nat Cancer Inst
2019
Feb 13 [Epub ahead of print]
.
7.
Worni
M
,
Akushevich
I
,
Greenup
R
,
Sarma
D
,
Ryser
MD
,
Myers
ER
, et al
Trends in treatment patterns and outcomes for ductal carcinoma in situ
.
J Natl Cancer Inst
2015
;
107
:
djv263
.
8.
Groen
EJ
,
Elshof
LE
,
Visser
LL
,
Rutgers
EJT
,
Winter-Warnars
HAO
,
Lips
EH
, et al
Finding the balance between over- and under-treatment of ductal carcinoma in situ (DCIS)
.
Breast
2017
;
31
:
274
83
.
9.
Surveillance Epidemiology and End Results (SEER) Program
.
SEER*Stat Database: SEER 13 Regs Research Data, Nov 2016 Sub (1992-2014) <Katrina/Rita Population Adjustment>- Linked To County Attributes - Total U.S., 1969-2015 Counties. National Cancer Institute, DCCPS, Surveillance Research Program, released April 2017, based on the November 2016 submission
.
Available from:
https://seer.cancer.gov.
10.
Youngwirth
LM
,
Boughey
JC
,
Hwang
ES
. 
Surgery versus monitoring and endocrine therapy for low-risk DCIS: the COMET trial
.
Bull Am Coll Surg
2017
;
102
:
62
3
.
11.
National Center for Health Statistics
. 
Health, United States, 2016: with chartbook on long-term trends in health. Hyattsville, MD: National Center for Health Statistics; 2017.
12.
U.S. Department of Health and Human Services; Centers for Disease Control and Prevention; and National Cancer Institute
.
U.S. Cancer Statistics Data Visualizations Tool, based on November 2017 submission data (1999-2015)
.
Available from:
https://gis.cdc.gov/Cancer/USCS/DataViz.html.
13.
Feuer
EJ
,
Wun
LM
. 
How much of the recent rise in breast cancer incidence can be explained by increases in mammography utilization? A dynamic population model approach.
Am J Epidemiol
1992
;
136
:
1423
36
.
14.
Wun
LM
,
Feuer
EJ
,
Miller
BA
. 
Are increases in mammographic screening still a valid explanation for trends in breast cancer incidence in the United States?
Cancer Causes Control
1995
;
6
:
135
44
.
15.
Ernster
VL
,
Barclay
J
. 
Increases in ductal carcinoma in situ (DCIS) of the breast in relation to mammography: a dilemma
.
J Natl Cancer Inst Monogr
1997
:
151
6
.
16.
Breen
N
,
Gentleman
JF
,
Schiller
JS
. 
Update on mammography trends: comparisons of rates in 2000, 2005, and 2008.
Cancer
2011
;
117
:
2209
18
.
17.
Engmann
NJ
,
Golmakani
MK
,
Miglioretti
DL
,
Sprague
BL
,
Kerlikowske
K
, 
Breast Cancer Surveillance Consortium. Population-attributable risk proportion of clinical risk factors for breast cancer
.
JAMA Oncol
2017
;
3
:
1228
36
.
18.
Kerlikowske
K
. 
Epidemiology of ductal carcinoma in situ
.
J Natl Cancer Inst Monogr
2010
;
2010
:
139
41
.
19.
Nichols
HB
,
Bowles
EJ
,
Islam
J
,
Madziwa
L
,
Sturmer
T
,
Tran
DT
, et al
Tamoxifen initiation after ductal carcinoma in situ
.
Oncologist
2016
;
21
:
134
40
.
20.
Cardoso
F
,
van't Veer
LJ
,
Bogaerts
J
,
Slaets
L
,
Viale
G
,
Delaloge
S
, et al
70-gene signature as an aid to treatment decisions in early-stage breast cancer
.
N Engl J Med
2016
;
375
:
717
29
.
21.
Chlebowski
RT
,
Hendrix
SL
,
Langer
RD
,
Stefanick
ML
,
Gass
M
,
Lane
D
, et al
Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women's Health Initiative Randomized Trial
.
JAMA
2003
;
289
:
3243
53
.
22.
Plevritis
SK
,
Munoz
D
,
Kurian
AW
,
Stout
NK
,
Alagoz
O
,
Near
AM
, et al
Association of screening and treatment with breast cancer mortality by molecular subtype in U.S. women, 2000-2012
.
JAMA
2018
;
319
:
154
64
.
23.
Cronin
KA
,
Miglioretti
DL
,
Krapcho
M
,
Yu
B
,
Geller
BM
,
Carney
PA
, et al
Bias associated with self-report of prior screening mammography
.
Cancer Epidemiol Biomarkers Prev
2009
;
18
:
1699
705
.