Click here to

Session: AI/ML Autoplanning, Autosegmentation, and Image Processing II [Return to Session]

Deep Learning-Based Auto Segmentation Using Generative Adversarial Network On Magnetic Resonance Images for Head and Neck Cancer

D Kawahara1*, A Saito2, Y Nagata3, (1) ,Department of Radiation Oncology, Hiroshima, JP


SU-F-BRB-5 (Sunday, 7/10/2022) 2:00 PM - 3:00 PM [Eastern Time (GMT-4)]

Ballroom B

Purpose: The current study aims to propose an auto-segmentation model using a generative adversarial network (GAN) on magnetic resonance (MR) images of head and neck cancer for magnetic resonance guided radiotherapy (MRgRT).

Methods: A dataset from the AAPM RT-MAC Grand Challenge 2019 was used. Especiallu, eight structures in the MR images, namely lymph node level II and level III, submandibular glands, and parotid glands, were segmented with the deep learning models using a generative adversarial network (GAN) and a fully convolutional network with a U-net. These images were compared with the commercially segmentation tool of atlas-based segmentation.

Results: The mean Dice similarity coefficient (DSC) and Jaccard similarity coefficient (JSC) of the U-net and GAN models was significantly higher than that of the atlas-based method for all the structures (p<0.05). The maximum Hausdorff distance (HD) was significantly lower than that in the atlas method (p<0.05). The DSC was highest for 0.75–0.85, and the HD was lowest within 5.4 mm of the 2.5D GAN model in all the OARs. Comparing the 2.5D and 3D U-nets, the 3D U-net was superior in segmenting the organs at risk (OAR) for HN patients.

Conclusion: The current study investigated the auto-segmentation of the OAR for HN patients using U-net and GAN models on MR images. Our proposed model is potentially valuable for improving the efficiency of head and neck radiotherapy treatment planning.




Not Applicable / None Entered.

Contact Email