ESTRO 2024 - Abstract Book

S486

Clinical - Breast

ESTRO 2024

statistically significant factors found in OS and LRFS regression analysis. The types of follow-up imaging, metastases status, estrogen receptor status, and number of metastases for SBRT were statistically significant factors (p < 0.05) in the multivariate Cox regression analysis on PFS.

Conclusion:

There are limited studies published on the efficacy and post treatment toxicities of metastatic breast cancer OM and OP SBRT with adequate length of follow-up. This study confirmed that SBRT was a safe, non-invasive treatment option to patients with extracranial OM and OP diseases originated from primary breast cancer in terms of the acceptable post treatment toxicities .

Keywords: SBRT, Oligometastatic, Oligoprogressive

References:

Nil

561

Digital Poster

Standardized prognosis prediction study for breast cancer through auto segmentation of PET images

Dong Hyeok Choi 1 , Joonil Hwang 2 , So Hyun Ahn 3 , Jin Sung Kim 1 , Hai Jeon Yoon 4 , Hyungju Kwon 5

1 Yonsei University, Medicine, Seoul, Korea, Republic of. 2 KAIST, Nuclear and Quantum Engineering, Daejeon, Korea, Republic of. 3 Ewha Womans University, Medicine, seoul, Korea, Republic of. 4 Ewha Womans University, Nuclear Medicine, Seoul, Korea, Republic of. 5 Ewha Womans University, Surgery, Seoul, Korea, Republic of

Purpose/Objective:

F-18 FDG positron emission tomography/computed tomography (PET/CT) is widely used in dignosis to evaluate tumor metabolism [1]. Through F-18 FDG PET/CT images, systemic inflammatory response can be detected and evaluated [2]. The size and placement of the volume of interest (VOI) used to obtain the standardized uptake value (SUV) is important to improve the precision of prognosis prediction using SUV [3, 4]. This research introduces an innovative method for analyzing and diagnosing nuclear medicine images through the application of deep learning.

Material/Methods:

The presented technique automates the segmentation and analysis of specific organs, allowing for SUV calculation throughout the entire organ volume. Utilizing the Swin UNETR model, essential prognostic indicators for breast cancer such as the breast, liver, spleen, and bone marrow were automatically segmented. In the training of artificial