Nuevos factores pronósticos y predictores de respuesta al tratamiento neoadyuvante en cáncer de mamaestudio del papel del sistema inmune con biomarcadores locales y sistémicos

Zusammenfassung

NTRODUCTION AND RATIONALE. Pathologic complete response (pCR) is the main prognostic factor for breast cancer (BC) patients treated with neoadjuvant chemotherapy (NCT). Uniquely, it is insufficient for the prognostic estimation of these patients. The tumor microenvironment (TME) plays a key role in the treatment response and prognosis of BC patients, but the integrity of the peripheral immune system is also relevant, so both compartments should be considered to define prognostic and predictive models that integrate immune markers. OBJECTIVE. The main objective of this doctoral thesis was to determine whether the combined assessment of tumor and peripheral immune response by tumor lymphocyte infiltration (sTIL) and neutrophil-lymphocyte ratio (NLR) at diagnosis is able to improve, compared to classical clinicopathological predictive and prognostic factors, response prediction and prognostic stratification in a consecutive series of early BC patients treated with NCT. MATERIAL AND METHODS. Observational, single-center, retrospective study of a cohort of patients diagnosed with BC and treated with NCT between 2008 and 2019. Quantification of sTIL was performed using a standardized method. A predictive and prognostic model was developed, including sTIL and NLR biomarkers. Internal validation of the models was performed using random sampling techniques (cross-validation and Bootstrap resampling) and subgroup analysis. RESULTS. After analyzing a total of 466 patients with BC treated with NCT, we found differences in infiltration according to immunohistochemical subtype (higher in HER2+ and triple-negative BC). The sTIL, as a continuous variable, was significantly associated with pCR, it was not a significant prognostic factor for disease-free survival (DFS), overall survival (OS) or BC-specific overall survival (OS) in the overall series, but it was for the triple-negative subtype. NLR values did not differ by subtype. NLR, as a continuous variable, was not associated with pCR and was not a significant prognostic factor for DFS, OS and BC-OS. No correlation was found between sTIL and NLR levels at diagnosis. In the pooled analysis of both biomarkers, only sTIL remained as an independent predictor of response to NCT. No significant differences were found in the analysis of the prognostic capacity for DFS, OS and BC-OS of the sTIL and NLR biomarkers together. The generation of a predictive model that included clinical (tumor size and lymph node involvement at diagnosis) and pathological (molecular subtype and grade 3) variables, together with the quantification of sTIL, was shown to be the optimal model for the prediction of pCR after NCT. Internal validation of the model by random sampling and bootstrapping techniques confirmed the robustness of the selected model. In the subgroup analysis, for the non-luminal BC, the optimal predictive model also included NLR, although the difference with the reference model did not reach statistical significance. For the three survival events, the optimal model for prognostic prediction was the one that included exclusively clinical variables (tumor size and lymph node involvement at diagnosis) and pathological variables (molecular subtype, achievement of pCR and post-treatment extracapsular nodal invasion). The inclusion of immune biomarkers (sTIL and NLR) did not prove to improve the discriminatory capacity of the clinicopathological model. CONCLUSIONS. The most relevant finding of this doctoral thesis has been the generation of a predictive model with clinical and pathological variables, together with the quantification of sTIL, which was shown to be the optimal model for the prediction of pCR, and this prediction was maintained in the exploratory analysis according to molecular subtype. Although this finding has not been confirmed in the survival analysis, the limited number of events and the predictive capacity of pCR do not allow us to completely rule out its impact in the longer term.