Noninvasive Quantification of Cerebral Blood Flow Using Hybrid PET/MR Imaging to Extract the [15 O]H2 O Image-Derived Input Function Free of Partial Volume Errors

Abstract

Background: Quantification of cerebral blood flow (CBF) with [¹⁵ O]H₂ O-positron emission tomography (PET) requires arterial sampling to measure the input function. This invasive procedure can be avoided by extracting an image-derived input function (IDIF); however, IDIFs are sensitive to partial volume errors due to the limited spatial resolution of PET.

Purpose: To present an alternative hybrid PET/MR imaging of CBF (PMRFlow_IDIF ) that uses phase-contrast (PC) MRI measurements of whole-brain (WB) CBF to calibrate an IDIF extracted from a WB [¹⁵ O]H₂ O time-activity curve.

Study type: Technical development and validation.

Animal model: Twelve juvenile Duroc pigs (83% female).

Population: Thirteen healthy individuals (38% female).

Field strength/sequences: 3 T; gradient-echo PC-MRI.

Assessment: PMRFlow_IDIF was validated against PET-only in a porcine model that included arterial sampling. CBF maps were generated by applying PMRFlow_IDIF and two previous PMRFlow methods (PC-PET and double integration method [DIM]) to [¹⁵ O]H₂ O-PET data acquired from healthy individuals.

Statistical tests: PMRFlow and PET CBF measurements were compared with regression and correlation analyses. Paired t-tests were performed to evaluate differences. Potential biases were assessed using one-sample t-tests. Reliability was assessed by intraclass correlation coefficients. Statistical significance: $\alpha$ = 0.05.

Results: In the animal study, strong agreement was observed between PMRFlow_IDIF (average voxel-wise CBF, 58.0 ± 16.9 mL/100 g/min) and PET (63.0 ± 18.9 mL/100 g/min). In the human study, PMRFlow_DIM (y = 1.11x - 5.16, R² = 0.99 ± 0.01) and PMRFlow_PC-PET (y = 0.87x + 3.82, R² = 0.97 ± 0.02) performed similarly to PMRFlow_IDIF, and CBF was within the expected range (eg, 49.7 ± 7.2 mL/100 g/min for gray matter).

Data conclusion: Accuracy of PMRFlow_IDIF was confirmed in the animal study with the primary source of error attributed to differences in WB CBF measured by PC MRI and PET. In the human study, differences in CBF from PMRFlow_IDIF , PMRFlow_DIM , and PMRFlow_PC-PET were due to the latter two not accounting for blood-borne activity.

Level of evidence: 2 TECHNICAL EFFICACY STAGE: 1.

Keywords: PET/MRI; [15O]H2O; cerebral blood flow; image-derived input function; phase-contrast MRI.