Speakers
Description
Medical image analysis provides a non-invasive framework to investigate tumor heterogeneity in lung cancer. Our work focuses on radiomics applied to Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) data, extracting quantitative features to study their association with biomarker expression and lung cancer subhistotypes. These approaches aim to complement conventional imaging and pathological assessment by capturing information not readily accessible through visual inspection. More recently, we have initiated radiogenomics analyses to explore links between imaging phenotypes and underlying molecular and genomic characteristics in lung cancer. While these methods show potential for improving biological interpretation and predictive modeling, their clinical translation remains challenging. In particular, issues related to image acquisition variability, feature reproducibility, normalization, and workflow standardization require careful investigation. Ongoing efforts therefore emphasize methodological rigor, harmonization strategies, and validation across datasets, with the goal of developing reliable and interpretable imaging biomarkers rather than immediate clinical deployment.
