Development of low-dose digital breast tomosynthesis system based on hybrid deep learning processing for image quality improvement

2022 Fiscal Year Final Research Report

• Project/Area Number: 20K08143
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 52040:Radiological sciences-related
• Research Institution: Kitasato University
• Principal Investigator: GOMI TSUTOMU 北里大学, 医療衛生学部, 教授 (10458747)
• Co-Investigator(Kenkyū-buntansha): 鯉淵 幸生 独立行政法人国立病院機構高崎総合医療センター(臨床研究部), 臨床研究部, 副院長 (10323346)
• Project Period (FY): 2020-04-01 – 2023-03-31
• Keywords: 乳がん / デジタルトモシンセシス / 深層学習

Outline of Final Research Achievements

In this study, we evaluated the improvement of image quality in digital breast tomosynthesis under low-radiation dose conditions of pre-reconstruction processing using conditional generative adversarial networks [cGAN (pix2pix)]. pix2pix pre-reconstruction processing with filtered back projection (FBP) was compared with and without multiscale bilateral filtering (MSBF) during pre-reconstruction processing. This phantom study revealed that an approximately 50% reduction in radiation-dose is feasible using our proposed pix2pix pre-reconstruction processing. Thus, pix2pix shows promise for integration into the clinical application workflow to reduce image noise while maintaining image quality in breast tomosynthesis.

Free Research Field

医用画像工学

Academic Significance and Societal Importance of the Research Achievements

新しい乳腺デジタルトモシンセシスシステムは、臨床で使用している通常の照射線量より少ない照射線量での撮像において画質改善を含めた有用性が期待できる。深層学習を応用した高解像度化処理に伴う微細病変の抽出能向上を実現できる新しい乳腺デジタルトモシンセシスシステムは、被ばく線量の低減と微細病変の診断能向上を図るものであり、画像診断の精度向上に寄与できる。本研究の成果は、更なる被ばく線量低減と画質改善につながるものであり、乳がんの早期発見に貢献することが期待できる。