Program Information
Dose Contrast-Enhanced CT Provide More Information for Liber in Radiomics Study
P Hu*, J Wang , L Shen , R Luo , J Gan , W Hu , Z Zhang , Fudan University Shanghai Cancer Center, Shanghai, shanghai
Presentations
WE-RAM2-GePD-J(A)-3 (Wednesday, August 2, 2017) 10:00 AM - 10:30 AM Room: Joint Imaging-Therapy ePoster Lounge - A
Purpose: We want to investage the radiomics features in liver with or without contrast CT and to analyze whether contrast-enhanced CT can provide more information for liver in radiomics study.
Methods: 166 patients with rectal cancer were included in our study, all patients received a non-contrast CT scan and a contrast-enhanced CT scan of chest protocol after radiotherapy. Both scans were obtained in the same CT scanner with the same protocol. The liver was distinguished and segmented by an experienced radiation oncologist in non-contrast CT scan, rigid assisted alignment and deformable registration were applied for the contrast-enhanced CTs by using MIM software. Then physician manually modified the contour as necessary. 206 features were extracted in our study by using QIAT ₍Quantitative Image Analysis Toolbox₎, an in-house developed radiomics feature extraction software, to calculate radiomics features for analysis. Intra-class correlation coefficients ₍ICC₎ of two CT scans were calculated to evaluate relationship between two CT scans. The relationship was also examined by using ROC curves testing between these features and liver damage.
Results: Over 90 percent features from non-contrast CT images have a strong correlation ₍ICC≥0.8₎ to the contrast-enhanced CT images; just ten features from non-contrast CT images have low correlation ₍ICC≤0.5₎ to contrast-enhanced CT images. The AUC values of the ROC curves of the liver damage prediction model were 0.72 and 0.60 for non-contrast CT and contrast-enhanced CT.
Conclusion: It observed that non-contrast CT is more suitable for liver in radiomics study.
Contact Email: