Noninvasive Measurement of [11C]PiB Distribution Volume Using Integrated PET/MRI.
Hidehiko OkazawaMasamichi IkawaTetsuya TsujikawaAkira MakinoTetsuya MoriYasushi KiyonoHirotaka KosakaPublished in: Diagnostics (Basel, Switzerland) (2020)
A noninvasive image-derived input function (IDIF) method using PET/MRI was applied to quantitative measurements of [11C] Pittsburgh compound-B (PiB) distribution volume (DV) and compared with other metrics. Fifty-three patients suspected of early dementia (71 ± 11 y) underwent 70 min [11C]PiB PET/MRI. Nineteen of them (68 ± 11 y) without head motion during the scan were enrolled in this study and compared with 16 age-matched healthy controls (CTL: 68 ± 11 y). The dynamic frames reconstructed from listmode PET data were used for DV calculation. IDIF with metabolite correction was applied to the Logan plot method, and DV was normalized into DV ratio (DVR) images using the cerebellar reference (DVRL). DVR and standardized uptake value ratio (SUVR) images were also calculated using the reference tissue graphical method (DVRr) and the 50-70 min static data with cerebellar reference, respectively. Cortical values were compared using the 3D-T1WI MRI segmentation. All patients were assigned to the early Alzheimer's disease (eAD) group because of positive [11C]PiB accumulation. The correlations of regional values were better for DVRL vs. DVRr (r2 = 0.97) than for SUVR vs. DVRr (r2 = 0.88). However, all metrics clearly differentiated eAD from CTL with appropriate thresholds. Noninvasive quantitative [11C]PiB PET/MRI measurement provided equivalent DVRs with the two methods. SUVR images showed acceptable results despite inferior variability and image quality to DVR images.
Keyphrases
- pet imaging
- computed tomography
- deep learning
- contrast enhanced
- end stage renal disease
- magnetic resonance imaging
- convolutional neural network
- positron emission tomography
- pet ct
- chronic kidney disease
- image quality
- newly diagnosed
- diffusion weighted imaging
- peritoneal dialysis
- optical coherence tomography
- prognostic factors
- high resolution
- big data
- artificial intelligence
- mild cognitive impairment
- patient reported outcomes
- data analysis