Risk factors for early locoregional recurrence among young-onset breast cancer patients: Findings from a single institutional prospective dataset

Introduction: While women diagnosed with breast cancer at young ages remain at risk of locoregional recurrence in long-term survivorship, predictors of risk remain unclear. Methods: Women ≤40 years of age diagnosed with breast cancer for which they underwent surgery between January 2016 and April 2023 (N=1,088) were identified from a prospectively maintained database with detailed clinicopathologic and treatment information collected. The primary outcome was locoregional recurrence (≥6 months after primary BC diagnosis) without concurrent distant metastasis. Univariable associations between patient characteristics, primary tumor and treatment variables, and outcome were assessed via Fine and Gray subdistribution models, considering distant metastasis and death as competing risks. Backward selection of variables by AIC was performed using crr step in R to build a multivariable model predicting risk of locoregional recurrence. A sensitivity analysis was completed among individuals who did not have a mastectomy. Results: A total of 1,053 women ≤40 years of age who did not experience a competing risk within the first 6 months were included. Mean age at first BC diagnosis was 35.3 (SD=4.2) years. Most patients were diagnosed with Stage 1 or 2 disease (77%) and had (HR)+/HER2- tumors (51%) [HR+/HER2+ (18%), HR-/HER2- (16%), HR-/HER2+ (10%), unknown HER2 (5%)]. Over a median of 3.6 (IQR=3.3-3.8) years of follow up, 34 women had a local or regional recurrence without concurrent metastasis. In univariable models, higher hazard of locoregional recurrence was associated with in situ disease (vs. invasive, p=0.05), smaller tumor size (p=0.005), lumpectomy (vs. mastectomy, p<0.001), non-receipt of endocrine therapy (ET) if HR+ (vs. HR+ w/ ET, p=0.009), and non-receipt of adjuvant chemotherapy (p=0.004). The multivariable model was tested among individuals with non-missing grade and stage (N=1037, events=34). Variables selected included stage (stage 3 vs. stage 1: aSHR=3.41, p=0.04), HR and ET combined (HR+ w/o ET vs. HR+ w/ET: aSHR=3.60, p=0.007), and surgery and RT combined (mastectomy only vs. lumpectomy+RT: aSHR=0.22, p=0.003; mastectomy+RT vs. lumpectomy+RT: aSHR=0.08, p<0.001). Other variables (age, race, BMI, known germline mutations, tumor grade, HER2 status, and adjuvant chemotherapy) were not selected. When restricted to individuals without mastectomy (N=405, events=26), receipt of ET emerged as the key predictor of locoregional recurrence (HR+ w/o ET vs. HR+ w/ET: aSHR=3.09, p=0.02). Conclusion: In a modern cohort of young BC patients, receipt of ET for HR+ disease emerged as the most important factor in predicting hazard of early locoregional recurrences. This highlights the importance of recommending ET for young BC patients with HR+ disease, while finding ways to increase tolerability of ET for young BC patients to encourage adherence.