Abstract #2313

Automatic Segmentation of Tumor-related Vessels of Breast Cancer on Ultrafast DCE MRI using U-Net

Masako Kataoka¹, Takuto Fukutome², Tomohiro Takemura², Kango Kawase², Kojiro Yano³, Maya Honda⁴, Mami Iima⁴, Akane Ohashi⁴, Masakazu Toi⁵, and Kaori Togashi⁴

¹Radiology (Diagnostic Imaging and Nuclear Medicine), Kyoto Univ. Hospital, Kyoto, Japan, ²Faculty of Medicine, Kyoto University, Kyoto, Japan, ³Osaka Institute of Technology, Osaka, Japan, ⁴Kyoto University Graduate School of Medicine, Kyoto, Japan, ⁵Kyoto University, Kyoto, Japan

We aimed to develop a system for automatic segmentation of tumor-related vessels on ultrafast dynamic contrast (UF-DCE) MRI using U-net. Training set consisted of image dataset of 20 MIP images obtained from -15 to +60 sec after contrast injection from 59 patients. Exclusion criteria were those with poor image quality. The dice similarity coefficient was above 0.8 for training set, 0.6 for validation set. Careful analysis of failed cases revealed that inaccurate segmentation of the vessel were caused by low-contrast images, noisy-images, artifact, bright skin line due to incomplete fat suppression, and non-mass enhancement that may mimic vasculature.

This abstract and the presentation materials are available to members only; a login is required.

Join Here