Denoising of ultra-low-dose 15O positron emission tomography images using deep image prior with anatomical information extracted through magnetic resonance segmentation

Purpose Recent deep-learning methods can recover standard-dose PET images from low-dose images. However, these methods require a large amount of data. We proposed a novel unsupervised learning approach using conditional deep image prior (DIP) with tissue probability maps extracted by magnetic resonance (MR) segmentation (DIP seg ), to improve denoising performance by more explicit anatomical guidance. Further, we retrospectively applied the proposed DIP seg method to recover standard-dose C 15 O 2 PET images from ultra-low-dose images. Methods Ultra-low-dose C 15 O 2 PET data for 10 patients with cerebrovascular steno-occlusive disease were generated by decimating the list-mode data by 1/128. The reconstructed ultra-low-dose images were denoised using DIP seg , DIP with MR T1-weighted images (DIP mr ), image-guided filtering (IGF), and Gaussian filtering (GF). Biases from the full-dose PET images and the coefficient of variation (CoV) for the ultra-low-dose datasets were calculated to evaluate quantitative accuracy and image noise, respectively. To validate the ability to improve the quantification quality, cerebral blood flow (CBF) maps were estimated using the autoradiographic method. Results DIP seg achieved a significantly lower CoV (∼10%) than the other methods, while maintaining a low bias (∼3%) compared to the full-dose images. Unlike DIP mr , DIP seg suppressed the increase in image noise during the iterations. CBF quantified using DIP seg showed high similarity to that from the full-dose images. Conclusion DIP seg can effectively recover the quality of standard-dose PET images from ultra-low-dose data, enabling accurate CBF quantification. The proposed DIP seg method has the potential to reduce radiation exposure in PET imaging, while maintaining diagnostic quality.