This multicenter retrospective observational study aims to clinically validate the PORTENT prognostic algorithm for predicting the risk of distant metastases in women with early-stage breast cancer. Female patients with Stage I-III disease from three independent cohorts with available residual tumor tissue and follow-up data will be included.
The primary endpoint of the study is the discriminatory performance of the PORTENT algorithm, assessed by the area under the receiver operating characteristic curve (AUC) for the prediction of distant metastases at 5 and 10 years from diagnosis.
The algorithm integrates established clinicopathological prognostic factors (tumor stage, histological grade, and Ki67-MIB1) with the expression levels of miR-3916 and miR-3613-5p. MicroRNA expression and target protein expression will be evaluated using RT-qPCR and immunohistochemistry.
Secondary and exploratory analyses will include model calibration assessed using calibration curves and the Integrated Calibration Index (ICI), as well as survival analyses (overall survival, progression-free survival, and metastasis-free survival) performed using Cox proportional hazards models.
