ESTRO 2024 - Abstract Book

S487

Clinical - Breast

ESTRO 2024

intelligence, three distinct datasets were employed: a test dataset, a validation dataset, and an independent test dataset, each comprising 40, 10, and 10 patients, respectively. Validation of the effectiveness of this approach was achieved by comparing the SUVs obtained through our methodology with those acquired using commercially available software.

Results:

Application of the proposed technique on a dataset of 10 patients yielded auto-segmentation accuracy of 0.9311 for all target organs. A quantitative assessment of SUVmax between our approach and the MIM software unveiled an average divergence of 0.23 ± 0.12. Notably, the liver exhibited the most pronounced discrepancy compared to other organs, a variance potentially attributable to its substantial volume. Conversely, the spleen manifested the least dissimilarity. Furthermore, the comparison of SUVmean revealed an average disparity of about 0.27 ± 0.03. Notably, the spleen exhibited the highest variance, while the left breast and liver demonstrated the least discrepancy (both at 0.26).

Conclusion:

This study successfully standardized the computation of SUV in nuclear medicine imaging through the introduction of a deep learning-based automated organ segmentation and SUV analysis approach. Potential applications of this method are expected to include not only breast cancer prognosis prediction, but also a variety of other carcinomas. In addition, it is expected that it will be possible to introduce it into clinical practice with rapid metabolic evaluation and guaranteed reproducibility.

Keywords: PET/CT, Swin UNETR, breast cancer

References:

[1] Proctor MJ, Talwar D, Balmar SM, O’Reilly DS, Foulis AK, Horgan PG, et al. The relationship between the presence and site of cancer, an inflammation-based prognostic score and biochemical parameters. Initial results of the Glasgow Inflammation Outcome Study. Br J Cancer. 2010;103:870 – 6. pmid:20717110 [2] Lee JW, Choi JS, Lyu J, Lee SM. Prognostic significance of 18F-fluorodeoxyglucose uptake of bone marrow measured on positron emission tomography in patients with small cell lung cancer. Lung Cancer. 2018;118:41 – 7. pmid:29572001

[3] Vali R, et al. SNMMI Procedure Standard/EANM Practice Guideline on Pediatric (18)F-FDG PET/CT for Oncology 1.0. J Nucl Med. 2021;62(1):99 – 110.

[4] Zwezerijnen, Gerben JC, et al. "Reproducibility of [18F] FDG PET/CT liver SUV as reference or normalisation factor." European Journal of Nuclear Medicine and Molecular Imaging 50.2 (2023): 486-493.