Abstract #4134

Background
 Medial (inner quadrants) tumor location has been shown to adversely influence survival in breast cancer. We investigated whether tumor location should alter the choice of local-regional treatments (surgery, with or without radiation therapy - RT), in women with node-negative breast cancer.
 Materials and methods
 Data were abstracted from the Surveillance, Epidemiology, and End Results (SEER) 2006's database for 58,709 women aged 25-95 years, diagnosed between 1988-1997 with non-metastatic T1-2 node-negative breast cancer, who underwent breast conserving surgery (BCS) and axillary dissection. We used Cox proportional hazards to examine the effect of tumor location (medial versus other) on overall mortality after accounting for other prognostic factors (covariates listed in footnote table 1) and interactions identified by the Akaike Information Criteria.
 Results
 There were no notable differences in patients' characteristics according to tumor location. Among them, 25,232 had BCS, 33,477 had mastectomy. However the distribution according to the use of adjuvant RT showed marked heterogeneity. RT was used in 88% of BCS and in 3% of mastectomy patients. In the multivariate analysis that adjusted against the imbalances, medial location was associated with an increased mortality, hazard ratio (HR) 1.08 (95% confidence interval 1.04-1.13). Mastectomy had no significant impact on survival, HR=0.99 (0.92-1.07). RT was associated with a significant mortality reduction, HR=0.81 (0.75-0.88). There was no interaction between tumor location and surgery, or between tumor location and RT, while the interaction between surgery and RT was significant, HR=1.31 (1.14-1.51). The corresponding factorial summary showed no subgroup effect of tumor location: RT after BCS was associated with the same mortality reduction of 19% regardless of tumor location, and RT after mastectomy was associated with the same relative increase of 7% regardless of tumor location (Table 1).
 Conclusion
 We confirm the poorer prognosis associated with tumor medial location. However there was no subgroup effect. The survival outcomes of local-regional treatments were not affected by tumor location, arguing that tumor location is not a sufficient indication to modify local-regional treatments in node-negative patients. Local-regional treatment should be based on tumor characteristics and not tumor location.


graphic

Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4134.

Thirty-first San Antonio Breast Cancer Symposium Dec 10-14, 2008; San Antonio, TX