Abstract

The brilliant therapeutic effect and longer patient survival time has made stereotactic body radiation therapy (SBRT) the major technology in radiation therapy. In the therapy procedure, cone-beam computed tomography (CBCT) imaging is routinely performed for patient set-up and positioning verification. CBCT images also have valuable information about day-to-day changes of the lesion during the treatment procedure, which may have huge potential influence in the therapy outcome. we perform scatter correction on CBCT and analyze the radiomics features from scatter corrected CBCT to explore its feasibility for radiomics analysis in SBRT. The dataset consisted of 12 patients with two-stage abdomen CBCT and planning CT (pCT) images. Firstly, the scatter artifacts are suppressed using a customized ultrafast Monte Carlo simulation-based scatter correction algorithm. The regions of interest (ROIs) are contoured randomly in CBCT, pCT and uncorrected CBCT images to extract the radiomics features. A total number of 271 radiomics features which include 59 unfiltered and 212 wavelet-filtered features are calculated. The Spearman's correlation coefficients (CCs) are calculated for statistical analysis. The percentages of strongly-correlated features (PSCFs) (CC>0.8) are improved by 11.8% and 9.9% in two stages respectively after the scatter correction. Specifically, for the features after applying low-pass filter in both x and y directions (LL-wavelet-filtered features), the PRFs increase from 17.0%, 18.9% to 37.7% and 34.0%, respectively. As the results shows, effective scatter correction can improve the correlation between CBCT and pCT features. The LL-wavelet-filtered features are improved most significantly. The features extracted from CBCT images have the possibility to be applied in radiomics.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call