Validación de la disección axilar dirigida por clip y radiotrazador como técnica de estadificación axilar del cáncer de mama tras quimioterapia neoadyuvante

Abstract

Introduction and aims: The standard treatment of the axilla in breast cancer (BC) with affected lymph nodes at diagnosis (cN1-3) and neoadjuvant chemotherapy (NACT) is an axillary lymph node dissection (ALND). In order to avoid it, new (post-NACT) axillary staging techniques were developed, such as the localization of the affected lymph node prior to NACT with a marker, subsequently performing a sentinel lymph node biopsy (SLNB) and the marked node biopsy (MNB), named as "Targeted Axillary Dissection" (TAD). The aim of this research is to study the feasibility and value of this technique in patients with BC cN1 undergoing NACT, and to determine predictive factors of the result of the ALND, in order to identify patients in which it could be omitted. Material and method: Prospective observational study between January 2016 and August 2019. Patients with BC, and histologically confirmed cN1 staging, treated with NACT, and marked with a metallic clip prior to neoadjuvant treatment were included. We performed: (1) A feasibility analysis of clinical, radiological, pathological variables, difficulties and complications of the TAD. (2) A diagnostic test study of the SLNB, MNB and the combination of their results (TAD), using ALND as Gold Standard. (3) A case-control study between patients with and without metastatic involvement in ALND, and between those with a "false" and "true" positive result in TAD, performing univariate and multivariate analysis (logistic regression) with the mentioned variables, in order to know predictive factors of axillary state after NACT. Results: 60 patients were included. 22 patients (36.7%) had a complete clinical response to NACT, and 43 patients (71.7%) were ycN0 (in the post-neoadjuvant clinical and ultrasound evaluation). Neither limitations nor complications in clip placement were found. 3 patients (5%) presented problems in wire placement due to difficulties in post-neoadjuvant ultrasound imaging. Intraoperative location of the clipped node was problematic in 7 cases (11.7%). The pathological complete response rate (pCR) was 30.5% (18 patients) and the pathological axillary complete response rate (ypN0) was 38.3% (23 patients). Sensitivity (SLNB: 80.9% (95% CI 61.8-100); MNB: 80.8% (95% CI 63.7-97.8); TAD: 92.6% (95% CI 80.9- 100)) and negative predictive value (SLNB: 84.6% (95% CI 68.8-100); MNB: 81.0% (95% CI 63.7-97.8); TAD: 91.3% ( 95% CI 77.6-100), figure 4.9) were higher than specificity (SLNB: 68.7% (95% CI 51.1-86.4); MNB: 63.6% (95% CI 45.7-81 , 6); TAD: 63.6% (95% CI 45.7-81.6)) and positive predictive value (SLNB: 62.9% (95% CI 42.9-83.0); MNB: 63.6 % (95% CI 45.7-81.6); TAD: 67.6% (95% CI 51.1-84.0)). The tumor size at diagnosis (Odds Ratio (OR) = 1.67; 95% CI 1.02-2.74), the number of suspected nodes in ultrasound at diagnosis (OR = 2.20; 95% CI 1.01-4 , 77), HER2 positive (OR 0.04; 95% CI 0.003-0.54), a complete clinical-radiological response to NACT (OR = 0.07; 95% CI 0.01-0.75), and a positive TAD (OR 15.48; 95% CI 1.68-142.78) were independent predictors of a positive result in LA. Based on the logistic regression model, we developed a "positive ALND predictive score", with a good calibration (Hosmer-Lemeshow test: p = 0.65), and discrimination (area under the curve = 0.93; 95% CI 0, 87-0.99), and with the highest Youden index (0.7) at cut-point of 17% risk of positive ALND (sensitivity = 100%; specificity = 70%). The only independent predictor of a false positive result in the TAD was the presence of HER2 (OR = 22.63; 95% CI 1.14-450.98). Conclusion: TAD is a feasible and valid technique to rule out axillary metastatic involvement in patients with BC cN1 with response to NACT. In patients with a small tumor and low number of suspected lymph nodes in ultrasound at diagnosis, HER2 positive, complete clinical-radiological response to NACT, a negative result in TAD, or more than one of these factors, omission of ALND could be considered.