ESTRO 2024 - Abstract Book

S3965

Physics - Image acquisition and processing

ESTRO 2024

Conclusion:

This study demonstrates the efficacy of using Pix2Pix GAN for generating high-quality synthetic CBCT images from Gamma Knife CBCT scans. The high SSIM score of 0.93 highlights the model's capability to produce images that are structurally similar to the original CT scans. These results are promising for enhancing the utility of CBCT images in clinical settings, particularly in radiotherapy treatment planning and monitoring. Future work should focus on validating the synthetic images' clinical utility and exploring other deep-learning architectures for further improvements.

References:

1. Ramachandran P, Perrett B, Dancewicz O, Seshadri V, Jones C, Mehta A, Foote M. Use of GammaPlan convolution algorithm for dose calculation on CT and cone-beam CT images. Radiat Oncol J. 2021 Jun;39(2):129-138.

2. Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, Bengio Y. Generative adversarial nets. Advances in neural information processing systems. 2014;27.

3. Isola P, Zhu JY, Zhou T, Efros AA. Image-to-image translation with conditional adversarial networks. InProceedings of the IEEE conference on computer vision and pattern recognition 2017 (pp. 1125-1134).

4. Ronneberger, O., Fischer, P. and Brox, T., 2015. U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18 (pp. 234-241). Springer International Publishing.

5. Wang Z, Bovik AC, Sheikh HR, Simoncelli EP. Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing. 2004 Apr 13;13(4):600-12.