| Autor: |
Amir L. Rifi, Ines Dufait, Chaimae El Aisati, Mark De Ridder, Kurt Barbe |
| Přispěvatelé: |
Clinical sciences, Faculty of Medicine and Pharmacy, Radiation Therapy, Artificial Intelligence supported Modelling in clinical Sciences, Digital Mathematics, Biostatistics and medical informatics, Public Health Sciences, Translational Radiation Oncology and Physics |
| Jazyk: |
angličtina |
| Rok vydání: |
2022 |
| Popis: |
Radiomics has the potential of characterizing the tumor phenotype hidden in medical images, allowing us to get more out of medical images than the eyes can see and liberating us from only using lesion size as a tumor response criterion. The extracted radiomics features are typically used in machine learning models to predict tumor responses. However, the inherent non-biological-interpretability of the features strongly hinders their clinical application. Therefore, our group aims to discover the biological meaning of radiomics features by performing dedicated in vivo experiments. Here, as a proof of concept, the radiomics features extracted from the CT scans of three widely used and well-characterized murine tumor models (CT26, 4T1 and EMT6) were analyzed and compared using an exploratory factor analysis. The results suggest that the features were able to differentiate between the different tumor models. To the best of our knowledge, this is the first attempt to directly link biological meaning to radiomic features using controlled in vivo experiments. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
