Frequency, determinants, and temporal trends in lymph node assessment in women with ductal carcinoma in situ

Abstract

11050 Background: This study examined factors associated with lymph node (LN) assessment in women with ductal carcinoma in situ (DCIS) at two sites in the U.S. Methods: Detailed clinical, pathologic, and radiographic information was collected through medical record review for 1090 patients with pure DCIS treated between 1985 and 2000 in Monroe County, New York and the Henry Ford Health System (HFHS), Detroit, Michigan. Sociodemographic, clinical, pathologic, and radiologic factors associated with LN assessment were examined using logistic regression models. Covariates were interacted with a site indicator; models were tested down to include only interactions with treatment and year. Results: 395 (36.2%) women had LN assessment; 43 of these were sentinel node dissections. Of the 438 women who had a mastectomy, 341 (86.3%) had LN assessment compared with 54 (13.7%) of the 598 women who had breast conserving surgery (p = 0.0001). In multivariate logistic regression, factors positively associated with LN assessment were mastectomy (OR, 45.5 95% CI 23.9, 86.5), private insurance (OR 2.3, 95% CI 1.2, 4.4), and high nuclear grade (OR 3.7, 95% CI 1.5, 9.2). Smoking was negatively associated with LN assessment (OR 0.5, 95% CI .3, .9). Age, race, method of detection (palpable mass vs. mammography), histologic subtype, and presence of multifocal disease were not associated with LN assessment. Statistically significant downward trends over time existed and did not differ between the two sites. LN assessment rates differed by treatment by site. Women who had mastectomy at HFHS were more likely to have LN assessment than women who had mastectomy in Monroe County (OR 5.0, 95% CI 1.5, 16.7), but women at HFHS who had autogenous reconstruction following mastectomy were substantially less likely to have LN assessment than their counterparts in Monroe County (OR 0.2, 95% CI .04, .65). Conclusions: The finding that LN assessment rates in women with DCIS vary according to non-clinical factors, such as treatment site and insurance type, suggests uncertainty about the role of LN assessment exists. Local treatment culture and financial incentives may play a role in the variations in LN assessment patterns. No significant financial relationships to disclose.