Related Experiment Videos
[Examination of normal database in three-dimensional SSP]
Yasushi Yamamoto1, Masuo Haramoto, Akio Komatsu
1Department of Radiology, Shimane University Hospital.
Nihon Hoshasen Gijutsu Gakkai Zasshi
|April 28, 2005
Summary
Establishing a multi-institutional normal database (NDB) for statistical image analysis is feasible. A single institution
Area of Science:
- Medical imaging analysis
- Nuclear medicine
- Statistical modeling
Background:
- Statistical image analyses, including three-dimensional stereotactic surface projection (3D-SSP), are sensitive to the normal database (NDB) employed.
- Ideally, each institution should develop its own NDB, but this is often limited by resources and expertise.
Purpose of the Study:
- To assess the validity of creating a general-purpose NDB from standard volunteer data at a single institution.
- To determine if such an NDB is suitable for use by other institutions.
Main Methods:
- Three NDBs were established using standard volunteer data: one from E.CAM and two from IRIX.
- IRIX NDBs were created with identical full width at half maximum (FWHM) and image reconstruction algorithms as E.CAM.
- Comparisons involved NDB mean values, standard deviations, and Z-score map effects.
Main Results:
- The NDBs established were found to be suitable for multi-institutional use.
- Identical FWHM (SPECT imaging resolution) and image reconstruction algorithms were key factors for NDB suitability.
- The Z-score maps showed comparable results across the tested NDBs.
Conclusions:
- A normal database (NDB) suitable for multi-institutional use can be established from single-institution volunteer data.
- Standardizing SPECT imaging resolution (FWHM) and image reconstruction algorithms is crucial for creating a transferable NDB.