engleski

Predictors of response to neoadjuvant treatment in estrogen receptor-positive breast cancer at the University Hospital for Tumors, Sestre milosrdnice University Hospital Center ; a retrospective cohort study

Breast cancer is not a single disease, but it is made up of diff erent subtypes of tumors that diff er in their biological behavior, aggression, tendency of metastasis and response to therapy. Among the patients with early breast cancer there is a subset of a patients who, with regard to disease biology or tumor size, is at increased risk for recurrence of the disease, ie subgroup with a worse prognosis. With the aim of reducing the risk of disease recurrence, systemic therapy has been traditionally used in the adjuvant setting, ie performed following surgical treatment. For the last 20 years, neo-adjuvant or inductive systemic therapy has been increasingly used as a fundamental strategy for treating patients with early breast cancer. Studies conducted so far indicate that the clinical and pathological response is largely dependent on the subtype of breast cancer. The studies consistently showed that the response to neoadjuvant treatment (achieving complete pathological response / pCR) was signifi cantly lower in patients with ER-positive versus ER-negative breast cancer patients. Aim of our study was to try to create a model for early prediction of pathological full response (pCR) to neoadjuvant treatment of ER- positive breast cancer. This retrospective cohort study was performed at the University Hospital for Tumors in a consecutive sample of 48 women treated with neo-adjuvant systemic antineoplastic therapy during 2015 and 2016. The test was performed on 68 patients. Finally, the study included 48 patients, out of which 13 (27%) received complete response to neo-adjuvant systemic antineoplastic therapy. After adjusting for all involved variables with multivariate binary logistic regression, the full response was independent, statistically signifi cant tumor size and lymph node involvement. Patients with a tumor ≥2.0 cm had 99% less chance of PCR than women with tumors <2 cm (OR = 0.01, 95% CI (0.00-0.28, p = 0.009). lymph nodes had a 20.7-fold higher chance for pCR than women without aff ected lymph nodes (OR = 20.7 ; 95% CI 1.19- 360.67 ; p = 0.038). Predictive model with only two variables: tumor size and involvement lymph nodes were signifi cant (Omnibus test, p <0.001), well- aligned with empirical data (Hosmer’s and Lemeshow’s test, p = 0.334), relatively high predictive values (Nagelkerke R2 = 0.42, 81% of the exact classifi cations). Classifi cation and regression strain revealed two segments of patients with the best prospects for pCR: 1) Patients with a tumor size <2.0 cm including 83% of pCR ; 2) Women with tumor size ≥2.0 cm with aff ected lymph nodes and HER2 negative tumor, including pCR 44%. In conclusion, tumor size <2 cm and lymph node involvement are promising predictors of response to neo-adjuvant systemic antineoplastic therapy of ER-positive breast cancer. The model is necessarily validated on an independent sample. As a limitation of the study, it is to be stressed out hat the size of the available population did not allow us to validate the model on an independent sample. At the University Hospital for Tumors, a validation prospective cohort study is currently underway, which we expect to complete in April 2018. The retrospective setup of our study has prevented us from analyzing more potentially important predictors such as lymphovascular invasion, gradus, cellularity, type or tumor immunophenotype.

hrčak, hrcak, znanstveni časopisi, hrvatski, scientific journals, croatian

nije evidentirano