Artificial intelligence (AI) can generate high-quality amyloid PET images from simultaneously acquired MR images and ultra-low-dose PET data, according to a Dec. 11 study published in Radiology.
In the study, researchers analyzed 40 datasets from 39 patients (16 males, 23 females) who underwent simultaneous amyloid fluorine 18 [18F]–florbetaben PET/MRI exams from March 2016 to October 2017.
One-hundredth of the raw list-mode PET data were chosen randomly to simulate a low-dose scan. A convolutional neural network with either PET-plus-MR images or low-dose PET alone were used to predict full-dose PET images.
PET imaging subjects patients to radiation dose from radiotracers, which also involve long acquisition periods. Patient movement during this process can lead to inaccurate PET radiotracer uptake quantification. Therefore, “reducing collected PET counts either through radiotracer dose reduction (the focus of this work) or shortening scan time (ie, limiting the time for possible motion) while maintaining image quality would be valuable for increased use of PET/MRI,” wrote Kevin T. Chen, of Stanford University in California, and colleagues.
Chen and colleagues found the synthesized images—notably the PET-MR model—achieved “marked” improvement on all quality metrics when compared to the low-dose scan. On a five-point scale measuring image quality (five being excellent), all PET plus-MR images scored a three or higher. Additionally, original full-dose images and the generated PET-MR images had “comparable” proportions of images that scored four or higher. The PET-only model had the fewest high-quality images. Amyloid status was nearly 90 percent accurate across 80 readings.
“Our results demonstrate that the accuracy of the PET-plus-MR synthesized images is comparable to that of readers interpreting the same full-dose images at different time points,” the authors wrote. “Our results can potentially increase the utility of amyloid PET scanning at lower radiotracer dose (radiation exposure similar to that on a transcontinental plane flight) and can inform future Alzheimer disease diagnosis workflows and large-scale clinical trials for amyloid-targeting pharmaceuticals.”