Quantification of cerebral blood flow (CBF) is important in the assessment of brain disorders such as stroke or Alzheimer’s disease [1]. ¹⁵O-H₂O PET imaging is an accurate method to measure CBF but it requires the measurement of an arterial input function (AIF), which is usually acquired using invasive arterial blood sampling. A non-invasive method using registered PET and MRI images for AIF estimation was recently introduced [2], which uses MR Angiography (MRA) images for carotid artery segmentation. Here, we propose using PET images along with the co-registered MRA images to segment the carotid artery more robustly and reliably.The time-of-flight enabled Signa PET/MR scanner (GE Healthcare, Waukesha, WI, USA) has enough sensitivity to provide quality PET images over a few seconds of ¹⁵O-H₂O PET imaging. Using an appropriate time window during the ¹⁵O arrival the carotid artery can be identified and segmented. To find the exact arrival time of the tracer into the brain arteries, the PET list file is binned every second after tracer injection and the total of true and scatter coincident events are plotted (Fig. 2). From this plot, the exact arrival time of tracer to the brain arteries is measured and a reconstruction is then performed over a 5-10 s time frame in order to visualize the brain arteries. The study was performed in compliance with regulations of the local Institutional Review Board and all subjects were consented prior to the study. Six subjects were injected with 925 MBq of ¹⁵O-H₂O during PET/MR imaging in order to compare CBF measurements by arterial spin labeling (ASL) and PET [3]. T1w, T2w, MRA, and ASL MRI images were acquired. The PET list file was unlisted for every second and total true and scatter coincident events were plotted to identify tracer arrival into the brain arteries. The carotid artery was then segmented using the MRA and PET images and two masks were generated. Since the PET and MRA images are co-registered, to obtain a more accurate segmentation, the two masks were multiplied by each other. Gray and white matter were segmented using T1w and T2w MR images. The PET images were reconstructed on a 3s frame sliding every 1s for the 1st minute after the injection, using a 5s frame sliding every 1s for the 2nd minute after the injection and using a 10s frame sliding every 5s for the 3rd and 4th minute after the injection. The tracer activity was then calculated over the carotid artery (i.e. AIF), gray matter, white matter, and whole brain by averaging the activities of all voxels in each mask.Figure 1 shows typical PET and MR images of ¹⁵O-H₂O study. Clear gray–white matter contrast is observed for this healthy volunteer. Fig. 2 shows the total of true and scatter coincident events after un-listing the PET list file. For this subject, the tracer enters the brain at about 20 seconds after the injection. A PET reconstruction is performed over a 5s frame between 20-25s after the injection, shown in Figure 3(a-b). A very clear contrast between the carotid artery and brain is observed which allows for segmentation of the carotid artery. Fig 3(c) shows the MRA images, which are used to improve the segmentation of carotid artery. The tracer activity in carotid artery, gray matter, white matter and whole brain is plotted in Figure 4 which shows a higher CBF for gray matter compared to white matter as expected.

The carotid arteries are reliably segmented from the PET images automatically. MRA images are used to further improve the segmentation and avoid partial volume due to higher image resolution.