Texture Analysis of Non-Small Cell Lung Cancer on Unenhanced CT and Blood Flow Maps: a Potential Prognostic Tool

• Main Authors: Serena Baiocco, Domenico Barone, Giampaolo Gavelli, Alessandro Bevilacqua
• Format: Article
• Language: English
• Published: FRUCT 2019-04-01
• Series: Proceedings of the XXth Conference of Open Innovations Association FRUCT
• Subjects: Texture analysis; Heterogeneity; CT; CT perfusion; Image processing; Biomarker; Tumour; Lung
• Online Access: https://fruct.org/publications/fruct24/files/Bai.pdf
• ISSN: 2305-7254; 2343-0737

The presence of tumour heterogeneity makes the clinical oncological practice very challenging, since introduces a great variability in tumours’ response to available therapies. For this reason, in the last decade, quantifying the salient features of the intra-tumoural heterogeneity has gained a great attention, also leading to a re-emerging of the texture analysis. Tumour heterogeneity represents the complex biology of tumour microenvironment, characterised by both spatial and temporal variability, increased by the presence of chaotic blood vessels within tumour tissue. Computed Tomography (CT) has always been considered one of the reference technologies for morphological analysis of organs and tissues, permitting to capture the “in vivo” spatial tumour heterogeneity. The need to also detect hemodynamic tumour features has stimulated the use of CT perfusion (CTp), a promising functional imaging technique in the oncological field. CTp allows detecting the presence of tumour abnormal hemodynamic patterns, by analysing the tissue temporal variations occurring after an intravenous administration of contrast medium. This work presents the extraction of meaningful statistical and texture features from both baseline CT images and perfusion maps of lung tumours, which could work as prognostic image-based biomarkers.