SGRT in Sim Usage of Surface Monitoring for Breathing-adapted Four-dimensional Computed Tomography (4DCT)

SGRT US 2024
Niklas Lackner,
MS PhD Student, Universitätsklinikum Erlangen Strahlenklinik

1. 1/22

2. Surface Monitoring for Breathing-Adapted
Four-Dimensional Computed Tomography
Niklas Lackner
Strahlenklinik, Universitätsklinikum Erlangen

3. 3/22
Disclaimer
The Universitätsklinikum Erlangen and the Department of
Radiation Oncology have institutional research grants with
àSiemens Healthineers AG
àVisionRT
https://ontheworldmap.com/germany/city/erlangen/

4. 4/22
The Role of Four-Dimensional Computed Tomography (4DCT) in Radiotherapy
nBreathing as a special challenge
nStereotactic- / hadron-radiotherapy
requires precise imaging
n4DCT as an essential component
àcan be used to estimate movement
with a series of time resolved images
Siemens Somatom Go.Open Pro // UICCI https://www.uicc.org (2024)

The futureisin motion
*CourtesyofSiemens HealthineersAG

5. 5/22
Conventional 4DCT
Functional Principle
nSynchronous acquisition of respiratory signal
and CT projection data
nIdentification of specific respiratory phases
nRetrospective assignment of
projection data to breathing curve
nAmplitude / phase-based reconstruction of
3D data sets of defined breathing phases
(often 10 3DCTs)

6. 6/22
Conventional 4DCT algorithms
Surrogate Systems: Breathing Curve Acquisition
Surface system/ BreathingbeltSpirometry
IR camerasystems
Images-copyright: VisionRT / Varian / Siemens Healthineers / Anzai* - Dr. Nassir Navab / Radiation Oncology for dentists - Kanhu Charan

7. 7/22
Conventional 4DCT algorithms
CT data acquisition
7
Spiral ModeSequenceScanning/ CineMode
…
Continuousbeam-on and tablemovementBeam-on
Table movement
Beam-on
Table movement
Beam-on
Table positionTable position
Directionof
patientmovement
Directionof
patientmovement
https://www.lecturio.com/concepts/computed-tomography-ct

8. 8/22
Conventional 4DCT Algorithms
Assumptions
I.Regular breathing of the patient during data acquisition
II.Constant relationship between respiratory signal and internal anatomical movement
III.Sufficient data to reconstruct defined breathing phases at each table position
àViolation leads to movement artifacts
Werner et al. – Med Phys (2019)

9. 9/22
Conventional 4DCT Algorithms
Assumptions
I.Regular breathing of the patient during data acquisition
II.Constant relationship between respiratory signal and internal anatomical movement
III.Sufficient data to reconstruct defined breathing phases at each table position
àViolation leads to movement artifacts
*Werner et al. – Med Phys (2017, 2019)
àConventional4DCT longoverduetoInnovation!
75% ofthetime patientsbreath
irregular, whichleadstoartifacts*

10. 10/22
Innovation! Intelligent 4DCT (i4DCT) Sequence Scanning
Functional Principle
nFAST 4D
§Initial learning period for the acquisition
of the patient-specific reference breathing cycle
§Automatic adaptation of scan parameters
nDirect i4D
§Online analysis of breathing curve
§Automatic x-ray trigger selection
§Optimized reconstruction algorithm
à„breathing-adapted 4DCT scanning“
Siemens 4DCT Cookbook (2019)
*CourtesyofSiemens HealthineersAG

11. 11/22
Innovation! Intelligent 4DCT (i4DCT) Sequence Scanning
Clinical Evidence for Artifact Free Images
Sentker et. al. - Radiotherapy and Oncology (2020)
nClinical study correlates the 4DCT image quality with
treatment success for conventional 4DCT
nRetrospective study: 62 patients with 102 lung/liver
metastases treated with SBRT
n4DCT artifacts have a negative impact on treatment
success (multivariate analysis p<0.05)

12. 12/22
Purpose
ni4DCT limitation:
àCurrentlyonlysupportedsurrogates: RGSC/Anzai
Purpose
nAssess the viability of two camera systems, an infrared system: Polaris Spectra and a SGRT system: SimRT in
comparison to our clinically used system RGSC in generating breathing-adapted 4DCT scans
àEvaluate breathing curve accuracy's impact on 4DCT quality
àSimulate latency's effect on curve reliability and image quality
àEvaluation utilizes experimental phantom data & patient data
Images-copyright: Varian / Anzai / NDI / VisionRT

13. 13/22
https://www.visionrt.com/de/products/
Why SimRT?
nMarkerlesstracking
nMightyieldmoreinformation
nLongterm-betterworkflowintegration:
àSimRTon imagingsite
àAlignRTon theacceleratorsite

14. 14/22
Materials and Methods
Phantom Measurements
CT scanner
nSOMATOM go.OpenPro
–Siemens Healthineers
Camerasystems
nPolarisSpectra(NDI)
nRGSC (Varian)
nSimRT(VisionRT)
Phantom
nDynamic Thorax Phantom 008A - CIRS
Reconstructions
nNon-clinical reconstruction software
– Siemens Healthineers

15. 15/22
Materials and Methods
Phantom Measurements
CT scanner
nSOMATOM go.OpenPro
–Siemens Healthineers
Camerasystems
nPolarisSpectra(NDI)
nRGSC (Varian)
nSimRT(VisionRT)
Phantom
nDynamic Thorax Phantom 008A - CIRS
Reconstructions
nNon-clinical reconstruction software
– Siemens Healthineers
https://www.cirsinc.com/products/radiation-therapy/dynamic-thorax-motion-phantom

16. 16/22
Results
Phantom Measurements - Breathing Curve Fix
Corrected curvesin blue
Integration of
table motion profile
Initially recorded curve
Removal of
SimRT’s
corrections
* projection dependent
scaling factor

17. 17/22
Results
Phantom Measurements

18. 18/22
Results
Patient Measurements

19. 19/22
Results
Image Quality in Phantom Measurements

20. 20/22
Results
LatencyMeasurements
Absolute latency:
nSimRT: ~ 60 ms
nRGSC: ~ 23 ms
nPolaris: ~ 16 ms
Latency
delay between the occurrence of an event and
measurement of that event through a surrogate signal

21. 21/22
Conclusion
In our retrospective study using three different optical surrogate systems in breathing adapted 4DCT:
nWe found excellent agreement* in phantom and patient breathing curves for the different systems
nImage quality generated with different surrogate input was comparable
nIn latency measurements, the two IR-marker systems were slightly faster than the SGRT system
nA larger set of experiments in a prototype environment is planned
à To investigate the direct impact of breathing curve quality and latency on image quality
*after table motion correction

22. 22/22
Veryspecialcontributionstoacknowledge
Erlangen UK
nJuliane Szkitsak, Andre Karius, Christoph Bert
VisionRT
nTorsten Moser, Julian Young, MananAdhvaryu
Siemens Healthineers
nChristian Hofmann, Jannis Dickmann

23. 23/22
Thankyou!
Contact:
Niklas Lackner
[email protected]