Histolopathologic approach to direct the management of the axilla in early breast cancer patients.

Screening mammography improved early detection of breast cancer and since then, the percentage of patients with node involvement is declining. Sentinel lymph node biopsy (SLNB), although helpful in the diagnostic armamentarium of these patients, has consequential limitations. In these patients, moreover, lymphatic invasion (LI) is of utmost importance to determine the risk of local recurrence. To find an adjuvant histologic approach to assist in pre-operative analysis of the patient's risk for having LI and thus sentinel ± non-sentinel axillary lymph node metastasis, one hundred and twenty patients with early invasive duct carcinoma without axillary lymph node metastasis were evaluated. Logistic regression predictive models were built from 80 patients and validated in the remaining 40 patients. The final stepwise multi-regression analysis identified four sensitive models. In the validation group, model 1 [applicable to tumors grade 3 having micropapillary differentiation (MPD)] correctly identified 92.31% of patients with LI and confirmed LI [positive predictive value (PPV) = 83%], but with moderate sensitivity. Model 2 [applicable to tumors without MPD], model 3 [applicable to tumors grade 1/2], and model 4 [applicable to tumors grade 1/2 having no MPD] all had moderate PPV and a high negative predictive value (NPV) rendering these models reliable negative tests. However, as they had high sensitivity, positive results confirm the presence of LI. Patient with tumors grade 3 and MPD might need SLNB and/or axillary lymph node dissection (ALND). Patients having tumors grade 1/2, size <2 cm, and no MPD nor extensive retraction artifact, SLNB and/or ALND could be omitted. In tumors grade 1/2 (model 2) and those without MPD (model 3), the proposed models are reliable negative tests rather than a definitive positive one.