Participants included 292 black (10%), 263 Hispanic (9%), 370 Asian or Pacific Islander (12%), and 2075 (69%) non-Hispanic white women. We identified only 1 woman as being non-white Hispanic (this patient was included in the black category). Asian or Pacific Islanders tended to be younger at the time of DCIS diagnosis, and 38% of those women were diagnosed before age 50 years (). Within each racial/ethnic group, the percentage of women diagnosed as having DCIS increased over time, with the sharpest rise occurring after 1997 among Asian or Pacific Islanders. Within this group, the rate of diagnosis increased from 14% in 1996–1997 to 29% in 1998–1999. The percentage of women with a known family history (first-degree blood relative) of breast cancer was similar in black, Hispanic, and white women (19% in each group) and was lower in Asian or Pacific Islanders (11%).
Characteristics of Women at Initial Ductal Carcinoma In Situ Diagnosis Between 1990 and 2001a
For participants with known addresses, 77% lived in census tracts where the geocoded educational attainment included some college. Geocoded educational attainment varied by race/ethnicity, with Asian or Pacific Islander women and white women representing the highest proportions of women who lived in census tracts with higher educational attainment (P <.001). Similarly, geocoded median family income varied by race/ethnicity, with higher percentages of black and Hispanic women living in census tracts with lower median family income (P <.001). Among participants with known weight, 58% of Asian or Pacific Islanders were of normal weight. A considerable percentage of white women were of normal weight (41%), although 57% were in the overweight and obese categories.
Detection method of DCIS was associated with race/ethnicity (P = .02) and varied by age, with more younger women having palpable lesions (P <.001). Although most lesions (83%) were detected by screening mammography, 17% were detected by palpation. Among Asian or Pacific Islanders, almost one-fourth of the lesions were detected by palpation, whereas among white women 16% were detected by palpation.
gives adjusted ORs from the multinomial logistic regression models for the association of patient and tumor factors with DCIS treatment regimens. The reference group for each treatment regimen (the dependent variable) comprised women who underwent BCS alone. The reference group for race/ethnicity (the main independent variable) was white women. Of 3000 women, 43% (n = 1284) underwent BCS alone, whereas 57% (n = 1716) underwent adjuvant radiation therapy or tamoxifen treatment. As expected, none of the women in our cohort were treated with chemotherapy. Radiation therapy alone was the most common adjuvant regimen. About 4% (n = 129) of women underwent adjuvant tamoxifen treatment only, while 11% (n = 337) underwent a combination of adjuvant radiation therapy and tamoxifen treatment.
Ductal Carcinoma In Situ Adjuvant Treatment Modalities by Characteristics of Women
Race/ethnicity was not strongly associated with use of adjuvant treatments; variation differed little by these subgroups of women, as the confidence intervals (CIs) were wide and included the null. Treatment regimens also did not vary by family history of breast cancer, geocoded educational attainment, BMI, or history of diabetes. Year of diagnosis was the strongest correlate of adjuvant radiation therapy use. For example, compared with women who underwent BCS alone, the odds of adjuvant radiation therapy receipt increased over the years; adjusted ORs varied from 2.55 (95% CI, 1.75–3.71) in 1992–1993 to 4.08 (95% CI, 2.86–5.83) in 2000–2001. Women who lived in census tracts with a low geocoded median family income were significantly less likely to receive adjuvant radiation therapy (OR, 0.65; 95% CI, 0.48–0.89). Older women (>70 years) were 2.5 times more likely to receive adjuvant tamoxifen therapy (without radiation therapy) than younger women (<50 years) (OR, 2.52; 95% CI, 1.29–4.90). However, older women were dramatically less likely to receive adjuvant radiation therapy (OR, 0.40; 95% CI, 0.31–0.51). The ORs for all treatment regimens increased over the years, with the sharpest rise occurring for combined radiation therapy and tamoxifen treatment.
Of the subset of women with pathologic data (n = 986), patients with large tumors (OR, 2.45; 95% CI, 1.23–4.88) and comedo histologic growth pattern (OR, 1.89; 95% CI, 1.35–2.64) were more likely to receive adjuvant radiation therapy than BCS alone. We did not have a large enough sample size to examine the association of pathologic factors and other treatment combinations (ie, BCS with tamoxifen treatment or BCS with radiation therapy and tamoxifen treatment). During the study period (1990–2001), estrogen receptor testing was not routinely performed. Therefore, this information was not captured in relation to tamoxifen use.