TY - JOUR T1 - A flow-limited oxygen-dependent diffusion model using heterogeneous perfusion for quantitative analysis of dynamic [<sup>18</sup>F]FMISO PET JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 420 LP - 420 VL - 52 IS - supplement 1 AU - Kuangyu Shi AU - Christine Bayer AU - Constantin Maftei AU - Florian Gaertner AU - Sabrina Astner AU - Jan Wilkens AU - Fridtjof Nüsslin AU - Peter Vaupel AU - Sibylle Ziegler Y1 - 2011/05/01 UR - http://jnm.snmjournals.org/content/52/supplement_1/420.abstract N2 - 420 Objectives The quantitative link between in vivo Hypoxia PET and tumor microenvironment (TM) is still elusive. Reaction-diffusion models have advantages in considering the interaction between tracer and TM. However, the lack of consistency of the modeling results and real [18F]FMISO PET hampers the application of previous models. This study aims at developing a reaction-diffusion model based on preclinical FMISO PET. Methods Nude mice with xenografted tumors were imaged with dynamic FMISO PET followed by immunofluorescence staining using the hypoxia marker pimonidazole and the endothelium marker CD31. A novel flow-limited oxygen-dependent (FLOD) model was developed. Time activity curves (TAC) were simulated for each mouse by feeding the extracted arterial input function (AIF) into the model. A modeled TAC was counted as matching to a measured TAC with the similarity metric below a certain threshold. Microscopic images were evaluated semi-quantitatively and qualitatively. The optimal threshold of tumor-to-blood (T/B) ratio for hypoxia was determined with FLOD. Results The matching rate between the simulated TACs and mouse PET ranged from 44.7 to 82.5% (avg 65.0%) for FLOD model, which was significantly higher than for the Kelly model [Phys Med Biol. 2006] (0-28.6%; avg 9.1%). The modeled intervessel distance distribution was similar to the results of microscopic images. We found that different AIFs can lead to different optimal T/B thresholds. Overall, this model suggests the optimal hypoxia T/B threshold to be 1.3. Conclusions By proposing a FLOD model, we were able to quantitatively describe the process of FMISO PET and predict the in vivo measurement. This model might improve PET quantification by optimizing the T/B threshold for hypoxia. Application of this model to clinical data is the topic of future studies. Research Support BMBF MobiTUM project (01EZ0826) &amp; DFG Cluster of Excellence: Munich-Centre for Advanced Photonics ER -