RT Journal Article
SR Electronic
T1 Harmonization of preclinical and clinical PET imaging to support co-clinical imaging research
JF Journal of Nuclear Medicine
JO J Nucl Med
FD Society of Nuclear Medicine
SP 1741
OP 1741
VO 59
IS supplement 1
A1 Richard Laforest
A1 Madhusudan Savaikar
A1 Kooresh Shoghi
YR 2018
UL http://jnm.snmjournals.org/content/59/supplement_1/1741.abstract
AB 1741Introduction: Numerous recent works highlight the limited utility of established tumor cell lines in recapitulating the heterogeneity of tumors in patients. More realistic preclinical cancer models are thought to be provided by transplantable, patient-derived tumor xenografts (PDX). Importantly, the use of PDX ushers new paradigms involving co-clinical imaging trials where novel quantitative imaging (QI) methodologies developed and validated using PDX mice can be implemented in a clinical setting. To that end, there is a need to harmonize preclinical and clinical imaging parameters. This work aims to harmonize imaging performance of animal PET imaging relative to human PET imaging in terms of count recovery and image noise to define data acquisition and image reconstruction guidelines for similar imaging conditions across subject scales. Methods: The NEMA NU-4 2008 image quality and the Micro Hollow Sphere (Data Spectrum, NC) phantoms were imaged on the Siemens Inveon MM PET/CT. The NEMA NU-4 phantom was filled with 250 uCi (9.25MBq). The hollow sphere phantom was filled with a sphere to background ratio of 6.5:1 and contained a total of 200 uCi (7.4 MBq) (sphere sizes: 5.95, 6.95, 8.23, 9.86 mm). Both phantoms were scanned in listmode for 15 minutes. The NEMA IEC Body Phantom (Data Spectrum, NC, sphere sizes: 10,13,17,22,28,27mm) was filled with a 8.6:1 ratio with a total activity of 1.73 mCi and imaged in the Siemens mMR in listmode for 30 min. On both platforms, attenuation correction was provided by scaling a computed tomography scan of the corresponding phantoms. On both platforms, images were reconstructed with a range of parameters typically used in animal and human imaging applications namely with various number of iterations and post-reconstruction smoothing filter parameters, with and without resolution recovery. Results: Small animal hot rod and hot sphere recovery coefficients (RC) were plotted as a function of size and compared to hot sphere RCs from the human scanner. A range of RCs from 20% to 95% was observed for object size range from 1 mm to 5mm in small animal PET imaging. RCs ranged from 20% to 100% for object size ranging from 10 to 37 mm (as defined by SUV peak). On either platforms, significant dependence of the RC was observed on the reconstruction parameters. Comparing both imaging platforms, a lesion with at 60% RC would have a size between 2-3 mm in small animal PET, this object would be 13 to 16 mm on a human scanner. Image noise was evaluated on an image reconstructed using filtered back projection algorithm (FBP) from data reconstructed using various amount of events on each platform. It was observed that approximately similar noise was achieved on non-filtered images if we take into account the relative volumetric spatial resolution between the two imaging systems. This indicates higher activity needs to be injected per body weight in high resolution animal imaging relative to human imaging. Conclusions: Taken together, these data illustrate that data acquisition and image reconstruction parameter can be determined to harmonize image properties across scales. This work will discuss optimization of small animal PET imaging conditions and present scenarios where image quality and quantitative accuracy can be compared across subject size scales.