Open AccessThe Impact of Varying Number of OSEM Iterations on Standardized Uptake Value and Image Quality of Discovery STE PET/CT Scanner
Author(s) -
Ali Bagherpour,
Asia Mushtaq,
Muhammad Basim Kakakhel
Publication year - 2018
Publication title -
journal of global oncology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.002
H-Index - 17
ISSN - 2378-9506
DOI - 10.1200/jgo.18.33600
Subject(s) - image quality , standardized uptake value , scanner , iterative reconstruction , imaging phantom , nuclear medicine , medicine , image resolution , contrast (vision) , positron emission tomography , artificial intelligence , medical physics , mathematics , image (mathematics) , computer science , radiology
Background: In PET imaging both quantitative and qualitative interpretations are used. Qualitative and quantitative interpretations depend upon PET/CT image quality that along with many biologic factors strongly depends upon image reconstruction parameters. Aim: The objective of this experimental work was to study the impact of one of the key reconstruction parameter, i.e., number of reconstruction iteration, on standardized uptake value and image quality of PET/CT scan. Methods: Images of NEMA IEC Image Quality Phantom were acquired in list mode for 10 mins on Discovery STE PET/CT scanner, using tumor to background ratio of 4:1 and 18 F-FDG as radiotracer. List mode data were further transformed into data sets of varying acquisition time (0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0 mins) per bed position. Transformed data set of 5.0 mins were used to study the impact of varying number of iterations (2, 3, 4, 5, 7, 10, 15, 20) using OSEM approach of iterative reconstruction. Standardized uptake value (SUV) and underestimation in SUV were calculated as quantitative measures, while hot lesion contrast, cold lesion contrast and background variability were calculated as qualitative measures. Results: Standardized uptake value, hot and cold lesion contrast, image spatial resolution and background variability showed increasing trend with increase in reconstruction iterations. Maximum increase of 20.25%, 16.33%, 9.79% and 6.88% was observed in SUV for 10 mm, 13 mm, 17 mm and 22 mm lesions as number of iteration change from 2 to 3. Smallest and the largest diameter lesions showed maximum underestimations of 54.67% and 8.20% at 2 iterations respectively. Percentage hot lesion contrast showed rapid increase as the number of iteration change from 2 to 7 and increased slowly afterward. Background variability range from 4.4% to 6.4%, 4.1%–5.7%, 3.6%–4.6%, 3%–3.8%, 2.7%–3.2%, 2.4%–2.7% for 10.0 mm, 13.0 mm, 17.0 mm, 22.0 mm, 28.0 mm and 37.0 mm sphere respectively. Conclusion: Optimized reconstruction parameters for routine clinical studies 3 iterations with image matrix size of 128 × 128 with filters FWHM of 6 mm and for high resolution studies 3 iterations image matrix size of 256×256 with filters FWHM of 6 mm.