Factors predicting tumor size at presentation in breast cancer patients

摘要: 10609 Background: Early diagnosis of breast cancer improves prognosis. A high percentage of women at our institution present with large breast cancers. Using tumor board records, we investigated the potential covariates associated with tumor size at presentation, including age, race, marital status, and insurance status in order to identify populations at risk for presenting with late stage disease. Methods: Data were obtained from tumor board records of University Health Systems, Greenville, NC, a regional referral center in rural Eastern North Carolina. We analyzed data from patients presenting between 1999 and 2005. Relationships between tumor size at diagnosis and explanatory variables (marital status, age, ethnicity, and insurance status) were explored. Results: 1568 patients with breast cancer were identified of which 32% were African American (AA), 67% were White (W), and 0.6% other. 8.3% of patients had Medicaid; 27%, Medicare; 53.5%, private insurance; 3.4%, HMO; and 6.1% were uninsured. Tumor size was 0.01 to 25 cm and highly skewed so that statistical inferences were conducted on the log of tumor size. Tumors of >5 cm were found in 13.8% of AA women as compared to 6.4% of W women. Overall tumor size in AA women was larger than in W women (p-value = .0000). Marital status was not statistically significant (p-value = .5148). A linear regression model for log tumor size at time of diagnosis using explanatory variables marital status, age, ethnicity, and insurance status showed that only race (p-value = .0000) and age (p-value = .0004) were statistically significant. Older patients presented with somewhat smaller tumors. Most of the variability in log tumor size remained unexplained (r-squared = 4%). Conclusions: AA women were more likely to present with large tumors. While Medicaid patients had larger tumors, insurance status was not statistically significant in the linear regression model due to partial confounding of insurance and race: 20.8% of AA women were on Medicaid compared to 2.6% of W women. No significant financial relationships to disclose.