biostudies-literatureTsai CHNational Health Research InstitutesNational Science and Technology CouncilMinistry of Education, Taiwan R.O.C.e70168https://www.ebi.ac.uk/biostudies/studies/S-EPMC12660554biostudies-literatureUnknown52(12)<h4>Background</h4>Amyloid PET imaging has been extensively employed in the noninvasive assessment of amyloid-beta accumulation in Alzheimer's disease. Various amyloid radiotracers are commonly used in clinical settings; however, the limited interchangeability among these radiotracers hinders the feasibility of long-term clinical trials and multicenter comparisons. The Centiloid method was proposed for standardization, though providing a single score per image; voxel-wise translation remains a formidable task.<h4>Purpose</h4>This paper proposes a U-Net model based on a deformable convolution network (DCNv3-based U-Net) for [ 11C$^{11}\textrm {C}$ ]-Pittsburgh compound B-to-[ 18F$^{18}\textrm {F}$ ]-florbetaben image translation to augment existing datasets for large-scale model training anMedical physicsPaired PET-MRI Deep Learning Model for Translating [&lt;sup&gt;11&lt;/sup&gt;C]PiB to [&lt;sup&gt;18&lt;/sup&gt;F]Florbetaben Amyloid Images.PMC12660554114-2628-E-002-019-MY3NHRI-EX114- 11205ECNTU-114L7872NTU-114L900703NTU-114V1015-5NTU-113V1015-4Chen KTTsai CHHuang SYHsiao ITWang HWLin YNfalsePaired PET-MRI Deep Learning Model for Translating [&lt;sup&gt;11&lt;/sup&gt;C]PiB to [&lt;sup&gt;18&lt;/sup&gt;F]Florbetaben Amyloid Images.<h4>Background</h4>Amyloid PET imaging has been extensively employed in the noninvasive assessment of amyloid-beta accumulation in Alzheimer's disease. Various amyloid radiotracers are commonly used in clinical settings; however, the limited interchangeability among these radiotracers hinders the feasibility of long-term clinical trials and multicenter comparisons. The Centiloid method was proposed for standardization, though providing a single score per image; voxel-wise translation remains a formidable task.<h4>Purpose</h4>This paper proposes a U-Net model based on a deformable convolution network (DCNv3-based U-Net) for [ 11C$^{11}\textrm {C}$ ]-Pittsburgh compound B-to-[ 18F$^{18}\textrm {F}$ ]-florbetaben image translation to augment existing datasets for large-scale model training an2025-01-01T00:00:00Z2025 Dec2026-06-06T07:17:17.598Z2026-06-06T03:07:09.006ZS-EPMC126605544131091010.1002/mp.70168