Candidates of Genomic Tests in HR+/HER2- Breast Cancer Patients With 1-2 Positive Sentinel Lymph Node Without Axillary Lymph Node Dissection: Analysis From Multicentric Cohorts

Abstract

BackgroundThe genomic tests such as the MammaPrint and Oncotype DX test are being gradually applied for hormone receptor positive/HER-2 negative (HR+/HER2-) breast cancer patients with up to three positive axillary lymph nodes (ALNs). The first results from RxPONDER trial suggested that Oncotype DX could be applied to patients with 1-2 positive sentinel lymph nodes (SLNs) without axillary lymph node dissection (ALND), which constituted 37.4% of the intent-to-treat population. However, there was no distinctive research on how to apply genomic tests precisely to HR+/HER2- patients with 1-2 positive SLNs without ALND. The purpose was to construct a nomogram using the multi-center retrospective data to predict precisely which HR+/HER2- candidates with 1-2 positive SLNs could be subjected to genomic tests (≤ 3 positive lymph nodes).MethodsWe conducted a retrospective analysis of 18,600 patients with stage I-III breast cancer patients treated with sentinel lymph node biopsy (SLNB) in Shandong Cancer Hospital, Fudan University Shanghai Cancer Center, and West China Hospital. The univariate and multivariate logistic regression analysis was conducted to identify the independent predictive factors of having ≤ 3 positive nodes among patients with 1-2 positive SLNs. A nomogram was developed based on variables in the final model with p<0.05. Calibration of the nomogram was carried out by internal validation using the bootstrap resampling approach and was displayed using a calibration curve. The discrimination of the model was evaluated using the ROC curve.ResultsBased on the database of the three institutions, a total of 18,600 breast cancer patients were identified undergoing SLNB between May 2010 and 2020. Among the 1817 HR+/HER2- patients with 1-2 positive SLNs undergoing ALND, 84.2% harbored ≤ 3 totals metastatic ALNs. The multivariate logistic regression analysis identified imaging abnormal nodes (OR=0.197, 95%CI: 0.082-0.472), the number of positive SLNs (OR=0.351, 95%CI: 0.266-0.464), the number of negative SLNs (OR=1.639, 95%CI: 1.465-1.833), pathological tumor stage (OR=0.730, 95%CI: 0.552-0.964), and lympho-vascular invasion (OR=0.287, 95%CI: 0.222-0.398) as independent predictors for the proportion of patients with ≤ 3 total metastatic ALNs (all p<0.05). These five predictors were used to create a predictive nomogram. The AUC value was 0.804 (95%CI: 0.681-0.812, p<0.001). The calibration curve showed a satisfactory fit between the predictive and actual observation based on internal validation with a bootstrap resampling frequency of 1000.ConclusionThe nomogram based on the multi-centric database showed a good accuracy and could assist the oncologist in determining precisely which HR+/HER2- candidates with 1-2 positive SLNs without ALND could perform genomic tests. In the era of SLNB and precision medicine, the combined application of genomic tests and SLNB could provide patients with a better strategy of dual de-escalation management, including the de-escalation of both surgery and systemic treatment.