Click here to

Session: Multi-Disciplinary: Image Guidance: Cone-Beam CT [Return to Session]

Cone Beam CT (CBCT) Based Synthetic CT Generation Using Deep Learning Methods for Dose Calculation of Nasopharyngeal Carcinoma Radiotherapy

X Xue1*, Y Ding1, J Shi2,X Hao2, X Li1, D Li1, Y Wu1, H An2, W Wei1,M Jiang3, X Wang4, (1) Hubei Cancer Hospital, Wuhan, CN, (2) University of Science and Technology of China, Hefei, CN, (3) Huazhong University of Science and Technology, Wuhan, CN, (4) Rutgers-Cancer Institute of New Jersey, NJ


SU-IePD-TRACK 3-1 (Sunday, 7/25/2021) 3:00 PM - 3:30 PM [Eastern Time (GMT-4)]

Purpose: Cone beam CT (CBCT) for the purpose of setup alignment inhibits its further use in adaptive radiotherapy due to artifacts and inaccurate Hounsfield units (HU). We compared three deep learning methods to generate high quality synthetic CT (sCT) images from CBCT and planning CT (pCT) for dose calculation of nasopharyngeal carcinoma (NPC) radiotherapy.

Methods: 169 NPC patients with a total of 20926 slices of CBCT and pCT images were included. The CycleGAN, Pix2pix and U-Net models were used to generate the sCT images. The mean absolute error (MAE), root mean squared error (RMSE), peak signal to noise ratio (PSNR), and structural similarity index (SSIM) were used to quantify the accuracy of the proposed models in a testing cohort of 34 patients. Radiation dose were calculated on pCT and sCT by three deep learning models following the same protocol. Dose distributions were evaluated for 4 patients by comparing the dose-volume-histogram (DVH) and 2D gamma index analysis.

Results: The average MAE and RMSE values between sCT by three models and pCT reduced by 15.4 HU and 26.8 HU at least, while the mean PSNR and SSIM metrics between sCT by different models and pCT added by 10.6 and 0.05 at most, respectively. There were only slight differences for DVH of selected contours between different plans. The passing rates of gamma index analysis for 4 patients under 3mm/3% 3mm/2%, 2mm/3% and 2mm/2% criteria were all higher than 95%. All the sCT had achieved better evaluation metrics than those of original CBCT, while the performance of CycleGAN model was proved to be best among three methods.

Conclusion: The dosimetric agreement confirmed the HU accuracy and consistent anatomical structures of sCT by deep learning methods. The results demonstrated that the sCT could be readily incorporated for further adaptive radiotherapy dose calculation in clinical practice.

Funding Support, Disclosures, and Conflict of Interest: This study was supported by the National Natural Science Foundation of China (No. 12075095), the Natural Science Foundation of Anhui Province (No. 1808085QH281), the Fundamental Research Funds for the Central Universities (No. WK9110000127).



    Not Applicable / None Entered.


    Not Applicable / None Entered.

    Contact Email