Theory and initial validation of a non-invasive hybrid [15O]O2-PET/MRI approach to obtain image-derived input functions

1602 Introduction: The challenges of accurately measuring the image-derived input function (IDIF) in order to avoid invasive arterial sampling for quantitative positron emission tomography (PET) is well recognized [1]. Recent studies focused on imaging the cerebral metabolic rate of oxygen (CMRO2) have derived the IDIF from time-activity curves (TACs); however, these techniques are either computationally intensive or require calibration to a measured arterial input function [2,3]. Here, we propose a PET/magnetic resonance imaging (MRI) method that uses MRI estimates of whole-brain (WB) cerebral blood flow (CBF) and oxygen extraction fraction (OEF) to scale the IDIF derived from WB [15O]O2-TAC (Fig. 1A). The required MRI measurements can be obtained with the OxFlow sequence [4]. In this study, the theory was validated with animal experiments, which incorporated MRI measurements of WB CBF and OEF. Arterial input functions (AIFs) were measured and PET-only estimates of CMRO2 are available for comparison [5-6]. Methods: [15O]H2O and [15O]O2 PET data were acquired in a hybrid PET/MR scanner (3 T Siemens Biograph mMR), together with simultaneous MRI oximetry (OxFlow), from juvenile pigs (n = 8). Animals were anesthetized with 3% isoflurane and 6 mL/kg/h propofol. Recirculating water was modeled by species-specific parameters [7]. Results: WB-CMRO2 estimates obtained with measured AIF and IDIF (Fig. 2) were 1.81 ± 0.10 and 1.70 ± 0.20 mLO2/100g/min, respectively (ns.). Corresponding WB OEF measurements were 0.31 ± 0.09 and 0.29 ± 0.09 (ns.). Finally, WB-CBV estimated using the IDIFs (average = 3.8 ±0.6 mL/100g) were within the expected normal range. Conclusions: The proposed approach offers a non-invasive alternative to measure CMRO2 by incorporating IDIFs obtained by hybrid PET/MRI. The approach is straightforward as it derives the IDIF directly from the WB PET TAC scaled by the corresponding MRI estimates of OEF and CBF. Furthermore, it avoids partial volume errors typically encountered when deriving the IDIF from region-of-interest analysis of the feeding arteries. Further studies are required to fully validate this approach in a PET/MR scanner on healthy volunteers. References: [1] Zanotti-Fregonara et al. (2011-JCBFM) [2] Su et al. (2017-JCBFM) [3] Kudomi et al. (2018-JCBFM) [4] Wehrli et al. (2014- Acad Radiol) [5] Kudomi et al. (2005-JCBFM) [6] Kudomi et al. (2013-JCBFM) [7] Kudomi et al. (2009-JCBFM) Figure 1. [15O]O2-IDIF (Ao(t) , Eq. 1) obtained by combining the WB TAC from [15O]O2-PET dynamic image with MRI measurements of WB CBF and OEF. WB cerebral blood volume (CBV) was estimated by the Grubb relationship. Eq. 1 is similar to the one obtained by Kudomi et al. [3], but with two terms in the Taylor series expansion. Figure 2. (A) Measured AIFs and (B) IDIFs (solid line; ± one standard deviation—dashed lines.) IDIFs internal dispersion was included to obtain comparable curves (dispersion constant of 20 s was used).

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.