Purpose: Megavoltage computed tomography (MVCT) offers an opportunity for adaptive helical tomotherapy. However, high noise and a reduced contrast in the MVCT images resulted from a decrease in the imaging dose to patients limits the usability of MVCT. Therefore, we propose an algorithm to improve the image quality of MVCT with limited training data set.
Methods: The proposed algorithm generates kilovoltage CT (kVCT)-like images from MVCT images using a deep learning-based image synthesis network. A total of 566 and 514 slices of CT images were prepared from 6 training and 5 test cases, respectively. Data augmentation using an affine transformation was used to produce the training data in order to overcome the lack of data diversity in the network training. The mean absolute error (MAE), root mean square error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) were used to quantify the correction accuracy of the images generated by the proposed algorithm. The proposed method was validated by comparing images generated with non-augmented datasets.
Results: The average MAE, RMSE, PSNR, and SSIM values were 13.12 HU, 39.44 HU, 35.99 dB, and 0.96 in the proposed method, whereas the result with non-augmented data showed inferior outcomes (15.65 HU, 45.82 HU, 34.66 dB, 0.95). The voxel intensity of the image obtained by the proposed method also indicated similar distributions to those of the kVCT image.
Conclusion: The proposed algorithm generates synthetic kVCT images from MVCT images using a deep learning network with small patient datasets. The image quality achieved by the proposed method was correspondingly improved to the level of a kVCT image while maintaining the anatomical structures of an MVCT image. With a further evaluation for clinical implementation, the proposed method can ensure an accurate treatment planning in adaptive radiation therapy.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2019R1F1A1062775,No. 2019M2A2B4096537).