Preoperative axillary nodal staging of invasive lobular breast cancer with ultrasound guided fine needle aspiration in patients with suspicious ultrasound findings versus aspiration in all patients - A retrospective single institutional analysis.

INTRODUCTION: - At present, surgical strategies for breast cancer patients with >2 lymph nodes (LN) involved differ from those with no or lower degree of nodal involvement. Preoperative assessment of the axilla is less sensitive in patients with lobular carcinoma (ILC) than patients with other histological tumour types. MATERIALS AND METHODS: - A retrospective analysis of axillary staging by palpation, axillary ultrasound (AXUS) and AXUS-guided fine-needle aspiration cytology (FNAC) of 153 patients with ILC diagnosed and operated on between January 2013 and December 2020 was performed. Patients had either sentinel node biopsy or axillary lymph node dissection according to current practice. In period 1, patients had FNAC only when AXUS suggested nodal involvement (n = 106), and in period 2, all ILC patients had axillary FNAC (n = 47). RESULTS: - Of the factors associated with >2LNs involvement, logistic regression suggested only AXUS/FNAC based staging as independent variable for all patients. Patients with AXUS-guided FNAC had a significantly higher proportion of true negative and lower proportion of true positive cases in the P2 period (0 vs 55% and 72% vs 11% for >2 LNs involvement, respectively; both p < 0.0001). CONCLUSIONS: - AXUS-guided FNAC of all ILC patients did not result in improved preoperative identification of patients with >2 metastatic LNs but increased the false-negative rate of the assessment by producing false-negative results in patients who would not have undergone a biopsy due to negative AXUS findings.

International multicenter tool to predict the risk of four or more tumor-positive axillary lymph nodes in breast cancer patients with sentinel node macrometastases.

Recently, many centers have omitted routine axillary lymph node dissection (ALND) after metastatic sentinel node biopsy in breast cancer due to a growing body of literature. However, existing guidelines of adjuvant treatment planning are strongly based on axillary nodal stage. In this study, we aim to develop a novel international multicenter predictive tool to estimate a patient-specific risk of having four or more tumor-positive axillary lymph nodes (ALN) in patients with macrometastatic sentinel node(s) (SN). A series of 675 patients with macrometastatic SN and completion ALND from five European centers were analyzed by logistic regression analysis. A multivariate predictive model was created and validated internally by 367 additional patients and then externally by 760 additional patients from eight different centers. All statistical tests were two-sided. Prevalence of four or more tumor-positive ALN in each center's series (P = 0.010), number of metastatic SNs (P < 0.0001), number of negative SNs (P = 0.003), histological size of the primary tumor (P = 0.020), and extra-capsular extension of SN metastasis (P < 0.0001) were included in the predictive model. The model's area under the receiver operating characteristics curve was 0.766 in the internal validation and 0.774 in external validation. Our novel international multicenter-based predictive tool reliably estimates the risk of four or more axillary metastases after identifying macrometastatic SN(s) in breast cancer. Our tool performs well in internal and external validation, but needs to be further validated in each center before application to clinical use.

Unifocal, multifocal and diffuse carcinomas: a reproducibility study of breast cancer distribution.

Multifocality of invasive breast carcinoma has been associated with prognostic disadvantage. Unifocal, multifocal and diffuse distributions have been recently defined for both inasive carcinomas and in situ components, and these have been combined into categories of prognostic relevance. Eight observers analyzed the same series of 30 megaslides from 29 carcinomas, and had to classify the lesions into the three distribution patterns of unifocal, multifocal or diffuse (or not present/non influential). The reproducibility of the distribution patterns of invasive carcinomas was better than that of the in situ carcinoma components, but was still only fair to moderate on the basis of kappa values. The reproducibility of DCIS was poor to slight with some kappa values reflecting agreement by chance only. The results suggest the definitions of these distribution patterns require refinements for a more reliable and reproducible diagnosis if one wants to associate prognostic information with this variable.