Academic Journal
Προβολή σε EDS

Query-based prediction of 3D x-ray scatter fields in an interventional suite using a transformer model.

Λεπτομέρειες βιβλιογραφικής εγγραφής
Τίτλος: Query-based prediction of 3D x-ray scatter fields in an interventional suite using a transformer model.
Συγγραφείς: Rettig N; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany., Schülein P; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany., Bukschat Y; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany., Vetter M; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany.
Πηγή: Journal of radiological protection : official journal of the Society for Radiological Protection [J Radiol Prot] 2026 Oct 06; Vol. 46 (4). Date of Electronic Publication: 2026 Oct 06.
Τύπος έκδοσης: Journal Article
Γλώσσα: English
Στοιχεία περιοδικού: Publisher: IOP Pub. Ltd Country of Publication: England NLM ID: 8809257 Publication Model: Electronic Cited Medium: Internet ISSN: 1361-6498 (Electronic) Linking ISSN: 09524746 NLM ISO Abbreviation: J Radiol Prot Subsets: MEDLINE
Imprint Name(s): Original Publication: [Bristol, UK] : IOP Pub. Ltd., [c1988-
Ιατρικοί όροι (MeSH): Scattering, Radiation* , Imaging, Three-Dimensional*, Monte Carlo Method ; X-Rays ; Fluoroscopy ; Radiation Protection ; Humans ; Radiometry ; Computer Simulation
Περίληψη: Fluoroscopy-guided interventions expose clinical staff to x-ray scatter radiation, motivating accurate three-dimensional scatter-field predictions for radiation protection planning. Monte Carlo (MC) simulations provide reference accuracy but require minutes to hours of computation, while analytical models are faster but often lack spatial fidelity in complex geometries. We present a transformer-based model for scatter dosimetry in a single simulated interventional suite with fixed room geometry and object inventory, supporting near real-time inference on localised subvolumes and interactive runtimes for whole-room evaluation. We train a separate transformer-based model for each target quantity: air kerma per source photon, transported energy, and energy fluence per source photon. Each predictor encodes spectrum and technique parameters and represents room objects and their poses as a set of scene tokens. Predictions are obtained by coordinate-conditioned regression at queried voxel locations, so inference can be restricted to selected subvolumes and runtime scales with the number of queried voxels. The training set comprisessimulated configurations with varied C-arm poses, spectra, patient-table and patient phantom positions. On an independent test set ofconfigurations, the model achieves medianfor log-transformed targets and structural similarity around. In linear units, medianvalues arefor air kerma per source photon,for transported energy, andfor energy fluence. Full-room inference, reconstructing the effective grid of aboutmillion voxels, takeson an NVIDIA L40S GPU compared tofor MC simulation, yieldingaspeed-up. These results represent a step toward interactive, room-scale radiation protection planning, enabling rapid what-if analyses of staff positioning and procedure planning.
(Creative Commons Attribution license.)
Contributed Indexing: Keywords: Monte Carlo simulation; interventional radiology; query-based inference; radiation protection; room-scale dosimetry; transformer; x-ray scatter
Entry Date(s): Date Created: 20261006 Date Completed: 20261006 Latest Revision: 20261006
Update Code: 20261006
DOI: 10.1088/1361-6498/ae9ecf
PMID: 42834616
Βάση Δεδομένων: MEDLINE
FullText Text:
  Availability: 0
Header DbId: cmedm
DbLabel: MEDLINE
An: 42834616
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Query-based prediction of 3D x-ray scatter fields in an interventional suite using a transformer model.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AU" term="%22Rettig+N%22">Rettig N</searchLink>; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany.<br /><searchLink fieldCode="AU" term="%22Schülein+P%22">Schülein P</searchLink>; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany.<br /><searchLink fieldCode="AU" term="%22Bukschat+Y%22">Bukschat Y</searchLink>; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany.<br /><searchLink fieldCode="AU" term="%22Vetter+M%22">Vetter M</searchLink>; Institute for Applied Artificial Intelligence and Robotics (A2IR), Technical University of Applied Sciences Mannheim, Mannheim, Germany.
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%228809257%22">Journal of radiological protection : official journal of the Society for Radiological Protection</searchLink> [J Radiol Prot] 2026 Oct 06; Vol. 46 (4). <i>Date of Electronic Publication: </i>2026 Oct 06.
– Name: TypePub
  Label: Publication Type
  Group: TypPub
  Data: Journal Article
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: TitleSource
  Label: Journal Info
  Group: Src
  Data: <i>Publisher: </i><searchLink fieldCode="PB" term="%22IOP+Pub%2E+Ltd%22">IOP Pub. Ltd </searchLink><i>Country of Publication: </i>England <i>NLM ID: </i>8809257 <i>Publication Model: </i>Electronic <i>Cited Medium: </i>Internet <i>ISSN: </i>1361-6498 (Electronic) <i>Linking ISSN: </i><searchLink fieldCode="IS" term="%2209524746%22">09524746 </searchLink><i>NLM ISO Abbreviation: </i>J Radiol Prot <i>Subsets: </i>MEDLINE
– Name: PublisherInfo
  Label: Imprint Name(s)
  Group: PubInfo
  Data: <i>Original Publication</i>: [Bristol, UK] : IOP Pub. Ltd., [c1988-
– Name: SubjectMESH
  Label: MeSH Terms
  Group: Su
  Data: <searchLink fieldCode="MM" term="%22Scattering%2C+Radiation%22">Scattering, Radiation*</searchLink> <br /><searchLink fieldCode="MM" term="%22Imaging%2C+Three-Dimensional%22">Imaging, Three-Dimensional*</searchLink><br /><searchLink fieldCode="MH" term="%22Monte+Carlo+Method%22">Monte Carlo Method</searchLink> ; <searchLink fieldCode="MH" term="%22X-Rays%22">X-Rays</searchLink> ; <searchLink fieldCode="MH" term="%22Fluoroscopy%22">Fluoroscopy</searchLink> ; <searchLink fieldCode="MH" term="%22Radiation+Protection%22">Radiation Protection</searchLink> ; <searchLink fieldCode="MH" term="%22Humans%22">Humans</searchLink> ; <searchLink fieldCode="MH" term="%22Radiometry%22">Radiometry</searchLink> ; <searchLink fieldCode="MH" term="%22Computer+Simulation%22">Computer Simulation</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Fluoroscopy-guided interventions expose clinical staff to x-ray scatter radiation, motivating accurate three-dimensional scatter-field predictions for radiation protection planning. Monte Carlo (MC) simulations provide reference accuracy but require minutes to hours of computation, while analytical models are faster but often lack spatial fidelity in complex geometries. We present a transformer-based model for scatter dosimetry in a single simulated interventional suite with fixed room geometry and object inventory, supporting near real-time inference on localised subvolumes and interactive runtimes for whole-room evaluation. We train a separate transformer-based model for each target quantity: air kerma per source photon, transported energy, and energy fluence per source photon. Each predictor encodes spectrum and technique parameters and represents room objects and their poses as a set of scene tokens. Predictions are obtained by coordinate-conditioned regression at queried voxel locations, so inference can be restricted to selected subvolumes and runtime scales with the number of queried voxels. The training set comprisessimulated configurations with varied C-arm poses, spectra, patient-table and patient phantom positions. On an independent test set ofconfigurations, the model achieves medianfor log-transformed targets and structural similarity around. In linear units, medianvalues arefor air kerma per source photon,for transported energy, andfor energy fluence. Full-room inference, reconstructing the effective grid of aboutmillion voxels, takeson an NVIDIA L40S GPU compared tofor MC simulation, yieldingaspeed-up. These results represent a step toward interactive, room-scale radiation protection planning, enabling rapid what-if analyses of staff positioning and procedure planning.<br /> (Creative Commons Attribution license.)
– Name: SubjectMinor
  Label: Contributed Indexing
  Group:
  Data: <i>Keywords: </i>Monte Carlo simulation; interventional radiology; query-based inference; radiation protection; room-scale dosimetry; transformer; x-ray scatter
– Name: DateEntry
  Label: Entry Date(s)
  Group: Date
  Data: <i>Date Created: </i>20261006 <i>Date Completed: </i>20261006 <i>Latest Revision: </i>20261006
– Name: DateUpdate
  Label: Update Code
  Group: Date
  Data: 20261006
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1088/1361-6498/ae9ecf
– Name: AN
  Label: PMID
  Group: ID
  Data: 42834616
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=cmedm&AN=42834616
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1088/1361-6498/ae9ecf
    Languages:
      – Code: eng
        Text: English
    Subjects:
      – SubjectFull: Monte Carlo Method
        Type: general
      – SubjectFull: X-Rays
        Type: general
      – SubjectFull: Fluoroscopy
        Type: general
      – SubjectFull: Radiation Protection
        Type: general
      – SubjectFull: Humans
        Type: general
      – SubjectFull: Radiometry
        Type: general
      – SubjectFull: Computer Simulation
        Type: general
      – SubjectFull: Scattering, Radiation
        Type: general
      – SubjectFull: Imaging, Three-Dimensional
        Type: general
    Titles:
      – TitleFull: Query-based prediction of 3D x-ray scatter fields in an interventional suite using a transformer model.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Rettig N
      – PersonEntity:
          Name:
            NameFull: Schülein P
      – PersonEntity:
          Name:
            NameFull: Bukschat Y
      – PersonEntity:
          Name:
            NameFull: Vetter M
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 06
              M: 10
              Text: 2026 Oct 06
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-electronic
              Value: 1361-6498
          Numbering:
            – Type: volume
              Value: 46
            – Type: issue
              Value: 4
          Titles:
            – TitleFull: Journal of radiological protection : official journal of the Society for Radiological Protection
              Type: main
ResultId 1