Abdominal multi-organ auto-segmentation using 3D-patch-based deep convolutional neural network

Hojin Kim, Jinhong Jung, Jieun Kim, Byungchul Cho, Jungwon Kwak, Jeong Yun Jang, Sang-wook Lee, June-Goo Lee & Sang Min Yoon
Nature/Scientific Reports, 2020
Related Product


Date Published
2020. 04

The study aimed to evaluate the efficiency and accuracy of an auto-segmentation framework using a 3D-patch-based convolutional neural network, U-Net, enhanced with a graph-cut algorithm, in segmenting abdominal organs in radiotherapy. The framework utilized 3D-patch-based CT images for the liver, stomach, duodenum, and right/left kidneys. The segmentation accuracy was assessed by comparing the predicted structures with those produced from atlas-based methods and inter-observer segmentation, using the Dice similarity coefficient, Hausdorff distance, and mean surface distance. The U-Net-based auto-segmentation outperformed atlas-based methods for all abdominal structures and showed comparable results to inter-observer segmentation, particularly for the liver and kidneys. Moreover, the average segmentation time reduced from 22.6 minutes to 7.1 minutes with automation using U-Net, showcasing potential clinical usefulness in terms of accuracy and time-efficiency. The concept of this study has been incorporated into AVIEW RT-ACS by Coreline Soft.


Please leave your inquiry if you have any questions regarding our products, recruitment, investment, or any other matters.

Contact us