Prediction of RecurIndex locoregional recurrence score using clinicopathological variables in pN1 breast cancer: an analysis from a prospective cohort

RecurIndex is a multigene profiling assay developed for the Chinese population that predicts the risk of local recurrence and distant metastasis in early-stage breast cancer. While clinicopathological features have traditionally guided prognosis and adjuvant therapy, their relationship with RecurIndex local-regional score remains to be fully elucidated. This prospective cohort study aimed to investigate the association between histopathological parameters and RecurIndex recurrence score (RS), and to develop a clinicopathological model to facilitate clinical decision-making in pN1 breast cancer patients. A total of 335 patients with pT1-2N1M0 breast cancer and available RecurIndex test results were initially included. Based on preliminary analysis showing that patients with triple-negative and HER2-enriched subtypes exhibited near-universal high genomic risk (> 98%), subsequent analyses were restricted to hormone receptor-positive (HR+) patients ( n = 239). Genomic risk was defined using a prespecified RecurIndex score threshold (low-risk: RS < 44; high-risk: RS ≥ 44). Univariate and multivariate logistic regression analyses were performed to identify clinicopathological variables associated with high-risk RS. A parsimonious clinical model was constructed based on regression coefficients from variables selected via stepwise regression and clinical rationale. Model performance was evaluated using area under the receiver operating characteristic curve (AUC), sensitivity, specificity, and predictive values. A nomogram and Excel-based calculator were developed to facilitate clinical application. Based on the LRR risk score (≥ 44 defined as high-risk), 163 patients (48.7%) were classified as low-risk and 172 (51.3%) as high-risk. High-risk patients more frequently had tumors > 2 cm (72% vs. 62%, p = 0.048), grade III histology (80% vs. 54%, p < 0.001), HER2 positivity (49% vs. 17%, p < 0.001), elevated Ki67 (45.5% vs. 27.6%, p < 0.001), and a higher number of resected lymph nodes (16.3 vs. 14.2, p = 0.011). Conversely, low-risk patients showed higher rates of BMI ≥ 24 (43% vs. 31%, p = 0.021) and higher ER/PR expression (both p < 0.001). HER2-enriched and triple-negative subtypes exhibited the highest proportions of high-risk patients (98.1% and 97.7%, respectively). Among HR+ patients, multivariate analysis identified HER2 positivity (OR = 2.895), PR expression (OR = 0.980 per 1% increase), and Ki67 index (OR = 1.027 per 1% increase) as independent predictors of high-risk RS (all p < 0.01). A clinical model incorporating age, HER2 status, PR, and Ki67 achieved an AUC of 0.819 (95% CI: 0.766–0.872) in the test set, with sensitivity 86.4%, specificity 67.3%, and negative predictive value 91.7%. Our clinical model, based on age, HER2 status, PR expression, and Ki67 index, effectively predicts high RecurIndex risk in pN1 HR+ breast cancer patients. The high negative predictive value supports its potential utility in identifying patients who may safely forgo RecurIndex testing.