Purpose: Early deep-learning-based image denoising techniques mainly focused on a fully supervised model that learns how to generate a clean image from the noisy input (noise2clean: N2C). The aim of this study is to explore the feasibility of the self-supervised methods (noise2noise: N2N and noiser2noise: Nr2N) for PET image denoising based on the measured PET data sets by comparing their performance with the conventional N2C model. Methods: For training and evaluating the networks, 18F-FDG brain PET/CT scan data of 14 patients was retrospectively used (10 for training and 4 for testing). From the 60-min list-mode data, we generated a total of 100 data bins with 10-s duration. We also generated 40-s-long data by adding four non-overlapping 10-s bins and 300-s-long reference data by adding all list-mode data. We employed U-Net that is widely used for various tasks in biomedical imaging to train and test proposed denoising models. Results: All the N2C, N2N, and Nr2N were effective for improving the noisy inputs. While N2N showed equivalent PSNR to the N2C in all the noise levels, Nr2N yielded higher SSIM than N2N. N2N yielded denoised images similar to reference image with Gaussian filtering regardless of input noise level. Image contrast was better in the N2N results. Conclusion The self-supervised denoising method will be useful for reducing the PET scan time or radiation dose.

BACKGROUND: One-stage repair of aortic arch anomalies and intracardiac defects through median sternotomy has been recently adopted by many institutions since it is known to be safer than the staged repair. The early and midterm results of the one-stage repair of aortic arch anomalies and intracardiac defects were retrospectively evaluated. MATERIAL AND METHOD: 45 patients who underwent one-stage repair of aortic arch anomalies and intracardiac defects performed by one surgeon from January 1996 to July 2003 were included in this retrospective study. The median age of repair was 16 days (range, 3 days~23.7 months) and the mean weight was 3.62+/-1.30 kg. Thirty one (68.9%) had coarctation and 14 (31.1%) had interrupted aortic arch. Associated intracardiac anomalies were VSD in 31 patients (VSD group), TGA or Taussig-Bing anomaly in 10 (TGA group), and others in 4 (ASD in 1, aortopulmonary window 1, truncus arteriosus 1, aortic and mitral stenoses 1, miscellaneous group). The arch obstruction was repaired with end-to-side anastomosis in 23 patients and end-to-end anastomosis in 22. RESULT: Overall postoperative hospital mortality was 22.2% (10/45); 16.1% (5/31) in VSD group, 40% (4/10) in TGA group, and 25% (1/4) in miscellaneous group. There was no mortality in VSD group since 1998, and the mortality in TGA group has remarkably reduced since technical modification for coronary transfer was adopted (75% vs 16.7%). There was no postoperative seizure or other neurological complications. Postoperative aortic restenosis occurred in 5 patients (5/35, 14.3%). Two patients underwent balloon aortoplasty with successful results. There was no reoperation. There was one late death caused by pneumonia 5 months after the operation. Two-year actuarial survival rate including operative death was 72.9%. CONCLUSION: The operative mortality of one-stage repair has been reduced with time and aortic restenosis rate was also acceptable. We concluded that this procedure is a reproducible procedure for aortic arch anomalies associated with intracardiac defects.
Aorta, Thoracic* ; Double Outlet Right Ventricle ; Hospital Mortality ; Humans ; Mitral Valve Stenosis ; Mortality ; Pneumonia ; Reoperation ; Retrospective Studies ; Seizures ; Sternotomy* ; Survival Rate ; Truncus Arteriosus