,PET

Yan-Ran Joyce Wang1, Liangqiong Qu1, Natasha Diba Sheybani1

  • 1From the Departments of Biomedical Data Science (Y.R.J.W., L.Q., N.D.S., K.E.H., X.X., D.R., H.E.D.L.), Radiology (Y.R.J.W., A.J.T.), and Nuclear Medicine (K.E.H.), Stanford University, Stanford, CA 94304; Department of Diagnostic and Interventional Radiology, University Hospital Tübingen, Tübingen, Germany (S.G.); School of Engineering, University of Science and Technology of China, Hefei, China (X.L., J.W.); and Department of Pediatrics, Division of Pediatric Oncology, Lucile Packard Children's Hospital, Stanford University School of Medicine, Stanford, Calif (A.P., D.R., H.E.D.L.).

概括

这项研究介绍了Masked-LMCTrans,这是一种用于儿科癌症成像中超低剂量全身PET重建的深度学习模型. 这种新的方法显著提高了图像质量,减少了噪音,并在低剂量PET扫描中改善了结构细节.