Deep learning for synthetic PET imaging: a systematic mapping review of techniques, metrics, and clinical relevance
Deep learning can create full-dose PET images with less radiation exposure. Neurological applications dominate synthetic PET research, maintaining essential diagnostic detail. Challenges include limited datasets and variability in tracer uptake, necessitating further advancements.
Authors
- A. Guarnera
- Maria Carmen Garganese
- Elisa Placidi (ORCID: https://orcid.org/0000-0003-4613-6870)
- Carlo Gandolfo (ORCID: https://orcid.org/0000-0003-1703-9107)
- Enrico Rosa
- Antonio Napolitano
- Maria Vaccaro
- Aurelio Secinaro
Institutions
- Università degli Studi eCampus (IT)
- Agostino Gemelli University Polyclinic (IT)
- Advanced Radiation Therapy (United States) (US)
- Bambino Gesù Children's Hospital (IT)
Publication Details
- Journal
- European Radiology Experimental
- Published
- 2026-02-09
- DOI
- https://doi.org/10.1186/s41747-025-00651-5
- Citations
- 1
- Primary Topic
- Medical Imaging Techniques and Applications
- Type
- article
- Field-Weighted Citation Impact
- 31.92
Funders
- Ministero della Salute