Synthetic CT supports retrospective upper-abdominal radiotherapy dose validation
Source check: Source checked. Not independent scientific confirmation.
Finding: A CycleGAN trained on scans from 11 patients generated synthetic CT images for six other liver-cancer patients. Dose calculations on those images were compared with deformably registered planning-CT references and generally agreed more closely than calculations using raw cone-beam CT. This is a small retrospective medical-physics validation, with remaining air–tissue interface errors, not demonstrated improved treatment delivery.
Technology: Cycle-consistent generative adversarial network for synthetic CT
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- CycleGAN learns cone-beam/CT image-domain mappings that preserve anatomy while correcting intensity values needed for conventional dose calculations; the resulting generated-image dose estimates answer a defined upper-abdominal planning question.
- Test and result
- Seventeen historical hepatocellular-carcinoma cases supplied 11-patient training and six-patient testing groups. Synthetic-image dose-volume and profile calculations were compared with deformed-planning-CT and raw cone-beam calculations in the six test cases.
- Limitation / still unknown
- The reference is a deformably registered planning image, not an independently measured true delivered patient dose. Model calibration is specific to the studied imaging-system pair; single-slice processing and limited field of view constrain use. Air–tissue interface errors remain. Gamma passing rates alone could miss evident dose underestimation. The study does not establish superior patient outcomes or formal clinical equivalence.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://aapm.onlinelibrary.wiley.com/doi/full/10.1002/acm2.13595
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Exact primary publisher title/DOI/date; relevant indexed primary cohort, learned-model and reference-validation Methods/Results and limitations. Fresh Crossref publication-status metadata, no linked update; direct full-page access limited..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.