Keywords:
Computer applications, Lung, PET-CT, Image manipulation / Reconstruction, CT-Quantitative, Computer Applications-General, eLearning, Structured reporting, Cancer
Authors:
K. S. Mader, A. W. Sauter, G. Sommer, B. Stieltjes; Basle/CH
DOI:
10.1594/ecr2018/C-2728
Methods and materials
For this study,
we take a group of 179 patients diagnosed with Lung Cancer with available imaging data.
For each of these patients we take the first PET and CT image.
The lesions within the images are then manually marked using a 3D Slicer-based Annotation Tool and include annotations of Tumor,
Lymph Nodes,
and Metastases following the TNM-atlas.
We take the established library PyRadiomics (radiomics.io) and complement it with a number of other common image features to generate 194 features per lesion covering both PET and CT images.
Each of these features is then calculated for each lesion in each patient (over 2000 lesions).
For each metric in each patient we show similar lesions (based on volume and the value of the metric).
The lesions for the maximum and minimum value for that metric are also saved.
Finally the current lesion is shown on the distribution for reference of where it lies on the spectrum.