Acquisition time optimization of positron emission tomography studies by use of a regression function derived from torso cross-sections and noise-equivalent counts

• Published in: Radiological physics and technology 2016-07, Vol.9 (2), p.161-169
• Main Authors: Kangai, Yoshiharu, Onishi, Hideo
• Format: Article
• Language: English
• Subjects: Fluorodeoxyglucose F18; Image Processing, Computer-Assisted - methods; Imaging; Medical and Radiation Physics; Medicine; Medicine & Public Health; Nuclear Medicine; Positron-Emission Tomography; Radiology; Radiotherapy; Regression Analysis; Signal-To-Noise Ratio; Time Factors; Tomography, X-Ray Computed
• ISSN: 1865-0333; 1865-0341
• DOI: 10.1007/s12194-016-0345-6

In this study, we aimed to optimize the positron emission tomography (PET) acquisition time for individual patients by employing a regression function derived from torso cross-sections by using computed tomography (CT) attenuation corrections and the noise-equivalent counts (NECs). We initially determined the standard image quality or the standard NEC at our institution by visually assessing the images acquired from 61 patients. We measured the NECs of the livers and the torso cross-sections of 165 patients who were evaluated with PET/CT with 18 F-2-fluoro-2-deoxy- d -glucose on the basis of our standard protocol of 120 s/bed position. The optimal acquisition time (OPT) was calculated as the product of the ratio of the standard NEC to the estimated NEC multiplied by 120 s. The estimated NEC was derived from the oval cross-section of each patient by use of the regression function. We evaluated the validity of the OPT equation in 59 additional patients. We determined 5.83 Mcounts as the standard NEC at our institution. The mean OPTs in a group of 59 patients of whom 20, 19, and 20 were underweight, normal-weight, and overweight, respectively, were 106.3 ± 18.0, 137.1 ± 4.6, and 172.1 ± 24.3 s, respectively. After optimization, the NECs for normal-weight and overweight patients increased by 14 and 43 %, respectively, compared with the NECs attained with use of the conventional acquisition time (120 s). Using the regression function based on the torso cross-sections and the NECs enabled optimizations of the PET acquisition times for individual patients.