MENU

Click here to

×

Are you sure ?

Yes, do it No, cancel

Updates to the Computational Environment for Radiological Research (CERR) Software: Inter-Software Radiomics Test Suite, Data Import and Segmentation

R Pandya*, A Iyer , A Apte , J Deasy , Memorial Sloan-Kettering Cancer Center, New York, NY

Presentations

(Sunday, 7/29/2018) 2:05 PM - 3:00 PM

Room: Davidson Ballroom B

Purpose: To present updates to CERR’s capabilities for Radiomics, Data Import and Segmentation.

Methods: Major updates include (1) Inter-software Radiomics Test suite: A test suite was added to CERR to compare feature calculation against the PyRadiomics library. All 7420 radiomics features tested between CERR and PyRadiomics were in close agreement. Tests were evaluated on tumors of 10 H&N cancer patients from the TCIA database. On an average, the largest differences were observed for Shape features on the order of 1%. In order to help understand the effects of pre-processing on feature calculation, side-by-side visualization of the original and processed images was developed. Pre-processing options include Wavelets and Laplacian-of-Gaussian filters. The available set of patch-wise radiomics features, useful for investigating sub-regions within the structures of interest, was extended to include first-order statistical features in addition to Haralick and Law’s texture. (2) Data import: DICOM import of Diffusion Weighted Imaging (DWI) and Dynamic Contrast Enhanced (DCE) MRI sequences was added. The DWI and DCE image volumes are grouped by b-value and acquisition time, respectively. DICOM import of images and RTSTRUCT for obliquely scanned patients was added. (3) Segmentation: Contouring tools in CERR were updated to include Brush, Eraser and active-contour-based refinement tools. To help evaluate the results of auto-segmentation algorithms, a graphical tool was developed to label different segments within structures. These labels, assigned by an expert, can then be used to evaluate algorithm performance, and in turn, facilitate further refinement.

Results: The Radiomics Test suite, new features and updates are available as open source, GPL copyright software at https://www.github.com/cerr/CERR.

Conclusion: The inter-software Test suite, updates to CERR’s import and contouring capabilities, and the inclusion of new pre-processing filters and patch-wise radiomics features make CERR an ideal tool for radiological research.

Funding Support, Disclosures, and Conflict of Interest: This research was partially funded by NIH grant 1R01CA198121 and NIH/NCI Cancer Center Support grant P30 CA008748.

Keywords

Not Applicable / None Entered.

Taxonomy

Not Applicable / None Entered.

Contact Email