A Nomogram to Estimate Locoregional Recurrence-Free Survival for Women With T1-2 N0 Breast Cancer Managed With or...

Postmastectomy radiotherapy (PMRT) in women with T1-2 N0 breast cancer is not routinely recommended by guidelines but is sometimes considered when patients have multiple predictors of locoregional relapse (LRR). This study re-evaluates the role of PMRT in the era of modern systemic therapy.

Patients with pT1-2 pN0 M0 breast cancer treated from 2005 to 2014 with total mastectomy were identified. Patients who had prior or synchronous breast cancer, neoadjuvant chemotherapy, or incomplete radiotherapy courses were excluded. LRR was analysed with a Fine-Gray subdistribution hazard model, and overall survival (OS) was analysed with Cox regression. A nomogram for estimating LRR was devised and validated with propensity score matching.

The study cohort included 3752 women with negative margins and a median follow-up of 11.4 years. As systemic therapy, 32.4% had chemotherapy, 70.2% had hormone therapy, and 10.7% had targeted therapy. The 10-year LRR for the study cohort was 3.7%. The 358 PMRT patients had more adverse features than the 3394 no-PMRT patients. LRR was increased by seven predictors: younger age, larger primaries, luminal B or triple-negative subtypes, lymphovascular invasion (LVI), close margins, omission of chemotherapy, and omission of hormone therapy. PMRT receipt was associated with decreased LRR (HR=0.53, P = 0.04) but did not affect OS. Our nomogram only predicted absolute LRR risks >10% without PMRT when combinations of four or more predictors were present. Only 5% of patients had four or more predictors. The nomogram's predictions for PMRT benefit were within 1% of the predictions based on the higher-risk, propensity-score-matched cohort.

For T1-2N0 breast cancers treated in the era of modern systemic therapy, PMRT did not influence OS. However, there is a small subgroup of high-risk patients for whom PMRT offers a meaningful LRR reduction. Our nomogram can help individualise PMRT decision-making for patients with multiple LRR predictors.