Keywords:
Thorax, Lung, Computer applications, CT, CT-Quantitative, Image manipulation / Reconstruction, Computer Applications-Detection, diagnosis, Segmentation, Technical aspects, Cancer, Tissue characterisation, Multidisciplinary cancer care
Authors:
E. Barabino1, G. Ficarra1, C. Genova1, M. Verda2, S. Casella3, S. Caprioli1, F. Grossi1, G. Cittadini1; 1Genoa/IT, 2Imperia/IT, 3Savona/IT
DOI:
10.26044/ecr2019/C-3625
Aims and objectives
Non–small cell lung cancer (NSCLC) comprises approximately 85% of all lung cancer cases [1].
Since its introduction in clinical practice,
immunotherapy revolutionized the treatment of NSCLC but nowadays there are no reliable clinical or biological markers that can predict response to immune checkpoint inhibitors.
Expression of Programmed-Death Ligand 1 (PD-L1) is considered unreliable to predict response to therapy due to intratumoral heterogeneous expression and non-standardized techniques of immune assay [2,3]. Texture Analysis (TA) is a technique that can be applied on digital images to capture the framework of pixel usually invisible to the human eye and to translate it in numbers.
Several features,
each describing a definite aspect of a digital image,
are extracted in texture analysis.
These features are often redundant and highly correlated with each other: a statistical model elaborated from these data could be unstable and not reflect a true textural signature [9].
Machine learning (ML) offers the possibility to process and select textural features to create performant predictive models.
The aim of our study is to evaluate the feasibility and added value of Computed Tomography Texture Analysis (CTTA) and ML approaches to predict response to immunotherapy in NSCLC.