development of bony lesion detection system for CT images by unsupervised deep learning

• Project/Area Number: 18K12095
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 90130:Medical systems-related
• Research Institution: The University of Tokyo
• Principal Investigator: Hanaoka Shouhei 東京大学, 医学部附属病院, 講師 (80631382)
• Project Period (FY): 2018-04-01 – 2021-03-31
• Project Status: Completed (Fiscal Year 2020)
• Budget Amount: ¥4,420,000 (Direct Cost: ¥3,400,000、Indirect Cost: ¥1,020,000); Fiscal Year 2020: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000); Fiscal Year 2019: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000); Fiscal Year 2018: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
• Keywords: 医用画像処理 / 深層学習 / X線CT / 骨転移 / 時間差分 / 異常検知 / 骨疾患 / ディープラーニング

Outline of Final Research Achievements

We developed a computer program which can estimate each voxel value of the latest CT examination from voxel values of the previous CT examination. Furthermore, it can also estimate the estimation error. Using these estimated CT value and the error, z-score of each voxel in the latest CT examination is calculated so that the anomaly-highlighted image is displayed. The proposed system was validated with a real film-reading environment. The experiment was performed with 11 radiologists and 80 datasets, and the improve of the receiver operating characteristic (ROC) curve was confirmed using the proposed temporary subtracted CT.

Academic Significance and Societal Importance of the Research Achievements

CT画像においてしばしば主治医や放射線科読影医によって見逃される早期のがん骨転移について、その検出を助ける骨病変抽出・強調表示手法が開発できた。これにより、がん骨転移をより早期に発見し治療できることが期待され、がん患者の予後、quality of lifeの向上に資すことができることと期待される。

Research Products

• [Journal Article] Unsupervised Deep Anomaly Detection in Chest Radiographs (2021)
  • Author(s): Nakao Takahiro、Hanaoka Shouhei、Nomura Yukihiro、Murata Masaki、Takenaga Tomomi、Miki Soichiro、Watadani Takeyuki、Yoshikawa Takeharu、Hayashi Naoto、Abe Osamu
  • Journal Title: Journal of Digital Imaging Volume: n/a Issue: 2 Pages: 418-427
  • DOI: 10.1007/s10278-020-00413-2
  • Peer Reviewed / Open Access
• [Journal Article] Clinical usefulness of temporal subtraction CT in detecting vertebral bone metastases (2019)
  • Author(s): Hoshiai Sodai、Masumoto Tomohiko、Hanaoka Shouhei、Nomura Yukihiro、Mori Kensaku、Hara Tadashi、Saida Tsukasa、Okamoto Yoshikazu、Minami Manabu
  • Journal Title: European Journal of Radiology Volume: 118 Pages: 175-180
  • DOI: 10.1016/j.ejrad.2019.07.024
  • Peer Reviewed
• [Journal Article] HoTPiG: a novel graph-based 3-D image feature set and its applications to computer-assisted detection of cerebral aneurysms and lung nodules (2019)
  • Author(s): Hanaoka Shouhei、Nomura Yukihiro、Takenaga Tomomi、Murata Masaki、Nakao Takahiro、Miki Soichiro、Yoshikawa Takeharu、Hayashi Naoto、Abe Osamu、Shimizu Akinobu
  • Journal Title: International Journal of Computer Assisted Radiology and Surgery Volume: epub ahead Issue: 12 Pages: 2095-2107
  • DOI: 10.1007/s11548-019-01942-0
  • Peer Reviewed
• [Presentation] Development of temporal subtraction CT images using deep learning to detect vertebral bone metastases (2021)
  • Author(s): 星合壮大、花岡昇平、野村行弘、他
  • Organizer: 第80回日本医学放射線学会総会
• [Presentation] Residual network-based unsupervised temporal image subtraction for highlighting bone metastases (2018)
  • Author(s): Shouhei Hanaoka
  • Organizer: CARS 2018, Berlin, 18th May, 2018
  • Int'l Joint Research