Program Information

Radiomics and Image Registration Updates for the Computational Environment for Radiotherapy Research (CERR)

A Apte*, Y Wang , J Deasy , Memorial Sloan Kettering Cancer Center, New York, NY

Presentations

SU-E-I-66 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

Purpose:
To present new tools in CERR for Radiomics, image registration and other software updates and additions.

Methods:
Radiomics: CERR supports generating 3-D texture metrics based on gray scale co-occurance. Two new ways to calculate texture features were added: (1) Local Texture Averaging: Local texture is calculated around a voxel within the user-defined bounding box. The final texture metrics are the average of local textures for all the voxels. This is useful to detect any local texture patterns within an image. (2) Image Smoothing: A convolution ball of user-defined radius is rolled over an image to smooth out artifacts. The texture metrics are then computed on the smooth image. Image Registration: (1) Support was added to import deformation vector fields as well as non-deformable transformation matrices generated by vendor software and stored in standard DICOM format. (2) Support was added to use image within masks while computing image deformations. CT to MR registration is supported. This registration uses morphological edge information within the images to guide the deformation process. In addition to these features, other noteworthy additions to CERR include (1) Irregularly shaped ROI: This is done by taking intersection between infinitely extended irregular polygons drawn on any of the two views. Such an ROI is more conformal and useful in avoiding any unwanted parts of images that cannot be avoided with the conventional cubic box. The ROI is useful to generate Radiomics metrics. (2) Ability to insert RTDOSE in DICOM format to existing CERR plans. (3) Ability to import multi-frame PET-CT and SPECT-CT while maintaining spatial registration between the two modalities. (4) Ability to compile CERR on Unix-like systems.

Results:
The new features and updates are available via https://www.github.com/adityaapte/cerr.

Conclusion:
Features added to CERR increase its utility in Radiomics, Image-Registration and Outcomes modeling.

Contact Email: