Ricerc@Sapienza

Background
Breast magnetic resonance imaging (MRI) is of fundamental importance in breast cancer care nowadays, having widely recognized indications, from locoregional staging to the evaluation of the response to NAC, and more recently has been employed to investigate biomarker information. Radiogenomics refers to the integration of image-extracted phenotypes and genomic data of the same tumor, which has the potential to improve both diagnosis and therapeutic strategies for evaluating the individualized disease signatures with higher accuracy. MicroRNAs (miRNAs), which are short (18-22 nucleotides) non-coding RNA sequences, are novel biomarkers of breast tumorigenesis with the ability to regulate hundreds of genes and biological pathways. They are appealing targets for screening, diagnosis, prognosis and for the choice of correct treatment.
Hypothesis
Our purpose is a computational approach of breast cancer (BC), that correlates phenotypes from magnetic resonance imaging (MRI) with genotypes and miRNAs expression, developing a radiomiRNomic map. In order to discover the relationships between MRI texture analysis and miRNA expression data derived, the radiogenomic strategy could open new frontiers in Breast Cancer Precision Diagnosis and Prognosis.
Aims
The project has two aims:
1.precision lesion assessment and diagnosis. To noninvasively predict the molecular subtype of BCs, attributing luminal A, luminal B HER+, luminal B HER-, non-luminal HER2+, or triple-negative/basal-like.
2.precision diagnosis and prognosis. To noninvasively correlate radiomics information with molecular subtype of BCs and miRNAs profiling.
Expected Results
1.To noninvasive predict molecular BC subtypes by an artificial intelligence-based algorithm of radiomics features.
2.To identify new radiomiRNomic profiles of multi-omics biomarkers, which have the potential to improve both diagnosis and prognosis with the ultimate aim of personalizing therapeutic strategies.