Development of automatic detection application for bone metastases in X-ray CT images

2013 Fiscal Year Final Research Report

• Project/Area Number: 24800017
• Research Category: Grant-in-Aid for Research Activity Start-up
• Allocation Type: Single-year Grants
• Research Field: Medical systems
• Research Institution: The University of Tokyo
• Principal Investigator: HANAOKA Shouhei 東京大学, 医学部附属病院, 助教 (80631382)
• Project Period (FY): 2012-08-31 – 2014-03-31
• Keywords: 画像診断システム / コンピュータ支援画像診断 / 医用画像処理 / 放射線診断学

Research Abstract

1. A segmentation method for vertebral bone regions was developed. In the method, five landmarks are detected on each vertebra and these landmark positions are used to initialize the following multi-atlas-based segmentation method which can segment regions of all vertebrae simultaneously.
2. A rough detection and displaying method for bone metastasis regions is developed. The method is based on the temporal subtraction between the recent CT images and the older CT images.
3. A database which includes 35 subjects with multiple bone metastases was established. In each subject, the metastasis regions were inputted manually in a 3-dimensional manner.

Research Products

• [Presentation] Semiautomatic segmentation of whole-spinal vertebrae in CT volumes by multi-atlas method : accuracy improvement by using anatomical landmark position information (2014)
  • Author(s): S Hanaoka, Y Masutani, M Nemoto, Y Nomura, S Miki, T Yoshikawa, N Hayashi, and K Ohtomo
  • Organizer: Computer assisted radiol- ogy and surgery
  • Place of Presentation: Fukuoka, Japan (accepted)
  • Year and Date: 2014-06-27
• [Presentation] Automated detection of anomalous spinal segmenta- tions : A feasibility study by using 300 CT datasets (2014)
  • Author(s): Y Nakano, S Hanaoka, M Nemoto, Y Masutani, N Hayashi, and K Ohtomo
  • Organizer: The 73rd annual meeting of the Japan Radiological Society
  • Place of Presentation: Yokohama, Japan
  • Year and Date: 2014-04-12
• [Presentation] A multiple anatomical landmark detection system for body CT images (2013)
  • Author(s): S Hanaoka, M Nemoto, Y Nomura, S Miki, Yoshikawa, N Hayashi, K Ohtomo, and Y Masutani
  • Organizer: First International Workshop on Bioimage Recognition (BIR)
  • Place of Presentation: Matsuyama, Japan
  • Year and Date: 2013-12-05
• [Presentation] Sparse Gaussian graphical model of spatial distribution of ana- tomical landmarks–whole torso model building with training datasets of partial imaging ranges (2013)
  • Author(s): S Hanaoka, Y Masutani, M Nemoto, Y Nomura, S Miki, T Yoshikawa, N Hayashi, and K Ohtomo
  • Organizer: Fourth MICAI workshop on Mathematical Founadtions of Computational Anatomy (MFCA)
  • Place of Presentation: Nagoya, Japan
  • Year and Date: 2013-09-22