Description:

Radiomicsis a rapidly developing and expanding field in oncologic imaging. It aims to maximize the prognostic utility of measurable and quantifiable imaging properties, also known as radiomicfeatures. Radiomicfeatures are designed to capture the unique characteristics of cancerous tissue through the quantification of various image-texture properties like randomness, coarseness, directionality. It is currently assumed that tumor-image heterogeneities are associated with the underlying tumor genotype and phenotype. Therefore, quantification of the tumor texture can provide a more detailed and distinctive description of tumor image properties in comparison to anatomic imaging biomarkers. The main objective of this study is to examine the interaction between CT-derived 3D texture features and anatomic tumor volume as a measure of tumor size.

Dr. Velichkois Research Assistant Professor of Radiology at Northwestern University-Feinberg School of Medicine. His scientific interests overlap in the areas of quantitative imaging assessment of the therapy response, development of imaging biomarker, medical image processing and application of ML and AI techniques in oncologic imaging. With a background in statistical physics and informatics, he applies concepts from one field to questions in another. In particular, he is interested in bridging computer vision and machine learning research with medical diagnosis and classification.

Dr. Velichkois the manager of the Quantitative Imaging Core Laboratory (QICL) in the Department of Radiology. The QICL is affiliated with the Robert H. Lurie Comprehensive Cancer Center (RHLCCC) of Northwestern University and provides standardized and unbiased tumor measurements for patients enrolled in RHLCCC oncologic clinical trials, including patients on NCTN clinical trials.