Development of a medical diagnosis support system for FDG whole body tumor PET

2006 Fiscal Year Final Research Report Summary

• Project/Area Number: 17591311
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Single-year Grants
• Section: 一般
• Research Field: Radiation science
• Research Institution: Tokyo Metropolitan Foundation for Research on Aging and Promotion of Human Welfare
• Principal Investigator: ODA Keiichi Tokyo Metropolitan Foundation for Research on Aging and Promotion of Human Welfare, Tokyo Metropolitan Institute of Gerontology, Research Scientist, 東京都老人総合研究所, 研究員 (70224235)
• Co-Investigator(Kenkyū-buntansha): ISHII Kenji Tokyo Metropolitan Foundation for Research on Aging and Promotion of Human Welfare, Tokyo Metropolitan Institute of Gerontology, Sub-Team Leader, 東京都老人総合研究所, 研究副部長 (10231135) ; TOYAMA Hinako International University of Health and Welfare, Department of Health Service Management, Professor, 医療福祉学部, 教授 (50180188)
• Project Period (FY): 2005 – 2006
• Keywords: positron emission tomography / tumor / FDG

Research Abstract

Introduction : [Fluorine-18]-2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) has widely been used for tumor detection and staging. Dual phase imaging of FDG-PET has reported to be useful to differentiate the accumulation in tumors from that in inflammation or physiological tissues. For that purpose, a visual inspection of the image and manual placement of regions of interest (ROIs) are necessary in order to correlate the standardized uptake value (SUV) of the target regions in early image with those in delayed one. The aim of this work is to develop a new diagnostic assistant system for oncology FDG-PET, that can automatically analyze and display the regional SUV balance between early and delayed scans by clustering method.
Materials and Methods : In clinical study, whole-body FDG PET examinations were performed in 10 patients with lung cancer with two-dimensional (2D) scan mode using ECAT EXACT 47 (Siemens/CTI, Knoxville, TN, USA). The diagnosis of the patients was co nfirmed by the follow-up clinical information. The early scans (7-8 bed positions) started at 57-71 min after injection of 144-249 MBq (4 MBq/kg of body weight). The delayed scans (3-6 bed positions) started at 88-147 min. Each scan time per one bed position was two minutes. The early and the delayed SUV images were calculated and co-registered individually by maximization of mutual information. A clustered map, which demonstrate the pixel-by-pixel relation between the SUVs in early and delayed scans, were created and displayed with four colors.
Results and Discussions : The S1 segment, indicated SUVs of >2.5 and increase of >10% between early and delayed scans, showed malignant lesions clearly, however, high uptake areas mapped with S1 were also observed in the bladder and heart. The S3 segment, indicated 1.0<SUVs【less than or equal】2.5 and increase of >10% between the both scans, suspected a malignancy, but physiological accumulation was possible. The S4 segment, indicated SUVs of >2.5 and decrease between early and delayed scans, showed benignity or physiological accumulations. The clustered regions, which were given voluntary color indexes, were able to demonstrate adequately pixel-by-pixel information of the amount of SUV and its time course between the early and the delayed scans. The present clustering method is useful in clinical diagnosis with dual phase FDG-PET scans by offering an easy way to explore the longitudinal alteration of SUV.