ESTRO 2024 - Abstract Book

S3062

Physics - Autosegmentation

ESTRO 2024

Fig. 2 (a) Violin plots showing the centroid errors between the cGAN segmentation and ground truth in the AP/LAT and SI directions. (b) Violin plots of the DSC between the cGAN segmentation and ground truth. The bars represent the median (thick) and quartiles (thin).

Conclusion:

The first algorithm for liver contrast agent segmentation during radiation therapy treatment was developed and evaluated on patient kV images. The preliminary results demonstrate that real-time segmentation is achievable to a high degree of accuracy using deep learning. This solution could enable a non-invasive method for real-time tumour tracking without the costs, time and side effects for marker placement. The method will undergo evaluation on additional patients as recruitment progresses.

Keywords: hepatocellular carcinoma, deep learning, cGAN

References:

1. Sung, Hyuna, et al. "Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries." CA: a cancer journal for clinicians 71.3 (2021): 209-249.

2. Poulsen, Per Rugaard, et al. "Respiratory gating based on internal electromagnetic motion monitoring during stereotactic liver radiation therapy: first results." Acta Oncologica 54.9 (2015): 1445-1452.

3. Keall, Paul J., et al. "The first clinical implementation of real-time image-guided adaptive radiotherapy using a standard linear accelerator." Radiotherapy and Oncology 127.1 (2018): 6-11.

4. Yue, Jinbo, et al. "Lipiodol: a potential direct surrogate for cone-beam computed tomography image guidance in radiotherapy of liver tumor." International Journal of Radiation Oncology* Biology* Physics 82.2 (2012): 834-841.

5. Keall, Paul. Radio-opaque Contrast Agents for Liver Cancer Targeting With KIM During Radiation Therapy (ROCK RT). Identifier NCT05169177. (2022): https://www.clinicaltrials.gov/study/NCT05169177

6. Isola, Phillip, et al. "Image-to-image translation with conditional adversarial networks." Proceedings of the IEEE conference on computer vision and pattern recognition. (2017): 1125-1134.