Abstract Background: HER2+ breast cancer (BC) is a heterogeneous disease. Despite the novel targeted therapies against HER2, 20% of patients relapse. In this context, the identification of prognostic and predictor biomarkers is needed.MicroRNAs are involved in BC development and prognosis. We aim at assessing the prognostic significance of two micro-RNAs (named as A and B) in HER2+ BC tissue, to propose signature to predict relapse in this setting of patients.Methods: This study was conducted in a cohort of 46 non-consecutive HER2+ BC patients diagnosed at Hospital Clínico València-INCLIVA. All BC patients received standard treatment for localized disease. RNA was isolated from formalin-fixed paraffin-embedded primary tumor and retro-transcribed to cDNA. MicroRNA expression levels were measured by real-time qPCR and results were normalized according to the expression of housekeeping mir-16 miRNA.Non-parametric Mann-Whitney U test was used to ascertain the statistical significance of differences in miRNA expression levels between disease-free and relapse. Receiver operator characteristics (ROC) curves were constructed and AUC, accuracy, specificity and sensitivity, were calculated as a biomarker performance parameter. The best cutoff value was established based on the highest value obtained in ROC curve analysis according to the maximization of the sum of sensitivity and specificity. Kaplan-Meier curves were plotted for disease-free survival (DFS) based on the best cutoff value. As both microRNA showed high correlation (r=0.85), a genetic signature was developed using Principal Component Analysis (PCA) and extracting the first component. MicroRNA targeted genes were downloaded from miRTarBase (release 7.0). Cytoscape (v3.8.0) and ClueGo (v2.5.6) and were used to perform functional enrichment analysis with the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway database. To select significant terms/pathways, p-values were adjusted by the Benjamni-Hochberg method (p<0.05). Results: The expression of both microRNA-A and B was higher in the primary tumor of those patients who relapsed after treatment compared to those free of disease (p= 0.003 and 0.0002, respectively). ROC curves demonstrated that microRNA-A and B might discriminate those HER2+ BC patients who relapse. MicroRNA-A predicts relapse with an AUC 0.740, sensitivity of 0.810, specificity of 0.583 and accuracy of 0.689, and microRNA-B presented an AUC 0.821, sensitivity of 0.810, specificity of 0.750, and accuracy of 0.778. The survival analysis showed that microRNA-A and B strongly predicted shorter DFS (p= 0.014 and 0.0012, respectively).Two-microRNA signature combining (PCA1= 99.4% absorbed variance) microRNA-A and microRNA-B expression levels had high performance in discriminating relapsed from non-relapsed HER2+ BC patients (AUC 0.748, accuracy 0.689, sensitivity 1, specificity 0.417). Signature predicted DFS (p= 0.009). Functional enrichment analysis returned 55 significant pathways. Of these, important pathways related to tumorigenesis and response to treatment were enriched. Interestingly, P53 pathway, apoptosis, and cell cycle were in the top 5 enriched pathways. Conclusions: We demonstrated the value of microRNA A and B as potential prognostic biomarkers. High expression was related to poor prognosis in over cohort of HER2+ BC patients. Both microRNAs were able to strongly discriminate patients who will relapse and predict shorter DFS. Moreover, our signature was able to predict the outcome, suggesting a potential prognostic role at diagnosis. Both microRNAs are related to important biological pathways associated to BC progression. Further investigations to optimize the micro-RNA signature are on-going. Citation Format: Anna Adam-Artigues, Juan Antonio Carbonell-Asíns, Sheila Zúñiga-Trejos, Santiago Moragón, Maria Teresa Martínez, Cristina Hernando, Eduardo Tormo, Octavio Burgués, Ana Lluch, Begoña Bermejo, Pilar Eroles, Juan Miguel Cejalvo. Identification of a novel two-microRNA signature for recurrence prediction in HER2 positive breast cancer [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS19-20.