Impact of postmastectomy radiotherapy on the outcomes of breast cancer patients with T1-2 N1 disease : An individual patient data analysis of three clinical trials.

PURPOSE

To assess the impact of postmastectomy radiotherapy (PMRT) on overall survival and relapse-free survival among breast cancer patients with T1-T2 N1 disease who received standard adjuvant systemic therapy.

METHODS

This is an individual patient data pooled analysis of 1053 breast cancer patients referred for adjuvant therapy in three clinical trials (BIG 02/98, BCIRG001, and BCIRG005). Overall survival was assessed according to whether or not patients received adjuvant radiotherapy through Kaplan-Meier analysis. Univariate and multivariate analyses of predictors of overall and relapse-free survival were conducted through Cox regression analysis.

RESULTS

Locoregional relapse rates (after a median follow up of 116 months) were 5.6% among patients who received adjuvant radiotherapy vs. 6.6% among patients who did not receive adjuvant radiotherapy. Actuarial 5‑ and 10-year locoregional relapse-free survival rates were 94 and 93%, respectively, among patients who did not receive adjuvant radiotherapy versus 95 and 92% among patients who received adjuvant radiotherapy. The following factors were associated with worse overall survival in multivariate Cox regression analysis: age < 40 years (P < 0.0001), T2 stage (P = 0.004), higher lymph node ratio (P < 0.0001), and negative hormone receptor status (P < 0.0001). Likewise, the following factors were predictive of shorter locoregional relapse-free survival: age ≤ 40 (P < 0.0001), no PMRT (P = 0.034), fluorouracil/adriamycin/cyclophosphamide (FAC) chemotherapy (P = 0.001), and higher T stage (P = 0.002).

CONCLUSION

The current analysis does not show a beneficial impact of PMRT on overall or relapse-free survival among patients with T1-T2 N1 disease who received standard adjuvant systemic therapy. There is, however, evidence of improvement in locoregional relapse-free survival with PMRT. These findings need to be prospectively validated.