Moglioni, M. (2024) SYNCT: PET-driven SYNthetic control CT generation for treatment monitoring in proton therapy. Il nuovo cimento C, 47 (5). pp. 1-6. ISSN 1826-9885
|
Text
ncc13066.pdf - Published Version Download (238kB) | Preview |
Abstract
Protontherapy(PT)offerstumourtreatments with highly conformal depth-dose distributions and less damage nearby healthy tissues over photon beam therapy. However, PT is sensitive to patient-specific anatomical variations which may lead to severe dose deviations. A control CT is generally prescribed to check the patient’s morphology. Treatment verification systems such as in-beam Positron Emission Tomography (IB-PET) are desirable to avoid the delayed detection of anatomical variations during PT treatments. However, the interpretation of the PET monitoring data is still a subject of research since PET does not offer a direct representation of the disease progress as a control CT does. The SYNCT project aims to overcome this issue by using Neural Networks (NN) to produce synthetic control CT (sCT) images which can provide a non-invasive and interpretable picture of the anatomical variations in the patients. We studied the feasibility of sCT production with MC simulations. The output of our NN, a Vision Transformer, correctly produced the sCTs, which were compared with the ground truth across multiple similarity metrics. This work can be a highly valuable tool in adaptive PT.
| Item Type: | Article |
|---|---|
| Subjects: | 500 Scienze naturali e Matematica > 530 Fisica |
| Depositing User: | Marina Spanti |
| Date Deposited: | 09 Dec 2024 14:14 |
| Last Modified: | 09 Dec 2024 14:14 |
| URI: | http://eprints.bice.rm.cnr.it/id/eprint/23280 |
Actions (login required)
 |
View Item |
