7/31/2011. Learning Objective. Video Positioning. 3D Surface Imaging by VisionRT

Transcription

1 CLINICAL COMMISSIONING AND ACCEPTANCE TESTING OF A 3D SURFACE MATCHING SYSTEM Hania Al-Hallaq, Ph.D. Assistant Professor Radiation Oncology The University of Chicago Learning Objective Describe acceptance testing procedures for surface imaging systems Understand how surface imaging isocenter congruence testing must be incorporated into ongoing quality assurance of IGRT Describe the commissioning process of surface imaging systems for whole breast radiotherapy Video Positioning Video Positioning Re-record after each MV verification Video Positioning 3D Surface Imaging by VisionRT Advantages: Reproducing patient position from last filmed fraction thereby reducing filming frequency Real-time monitoring during treatment Disadvantages: Qualitative not quantitative feedback No interface with LINAC for beam-hold 1

2 Learning Objective Describe acceptance testing procedures for surface imaging systems Understand how surface imaging isocenter congruence testing must be incorporated into ongoing quality assurance of IGRT Describe the commissioning process of surface imaging systems for whole breast radiotherapy Acceptance Testing Acceptance test procedures provided by the vendor are typically designed to verify contractual system specifications for performance characteristics of the system. (TG-101) Acceptance Tests Note: on importing DICOM RT data as reference surface, VisionRT accepts no responsibility for the accuracy of such data which may be affected by data resolution and breathing artefacts. Such errors could influence positioning accuracy. Optimize camera positions Verify DICOM transfer from TPS Verify translational/rotational accuracy of registration algorithm Verify accuracy of retrospective gating Verify accuracy of prospective gating For both surface imaging modes: 1) Static capture 2) Real-time monitoring Additional Acceptance Tests? Acceptance: DICOM transfers Consistency of real-time data With static capture mode Over time Verify coincidence with treatment isocenter Verify beam hold-off capabilities 2

3 Acceptance: DICOM transfers Reference and captured surfaces are from 2 different patients! Acceptance: DICOM transfers Prone Orientation TPS AlignRT Acceptance: Translational & Rotational Accuracy Acceptance: Gating LAT Shift = 2.0cm VRT, LNG, LAT Shift = 2.0cm According to TG-76, QA of gating systems must address two fundamental sources of potential error in dose delivery: Determination of the tumor position as a function of time Calibration of the spatial relationship between the tracking coordinate system and the beam-delivery coordinate system 4D MOTION PHANTOM All Shifts = 0.0cm Rotation = 10 o Rotation = 45o Acceptance: Prospective Gating Acceptance: Real-Time Data Stability AlignRT Innacurate Tracking of Tumor Motion Accurate Tracking of Tumor Motion Philips BigBore: Table motion during scanning is out of bore Delay between trigger and acquisition Polaris Lights Off Lights On Data courtesy of Rodney Wiersma, Ph.D. at Univ. of Chicago 3

4 Acceptance: Isocenter Coincidence TG-142: Imaging & Treatment Isocenter Coincidence If individual errors are small by themselves, cumulative system accuracy for the procedure can be significant and needs to be characterized through an end-to-end test using phantoms with measurement detectors and imaging (TG-101) Learning Objective Describe acceptance testing procedures for surface imaging systems Understand how surface imaging isocenter congruence testing must be incorporated into ongoing quality assurance of IGRT Describe the commissioning process of surface imaging systems for whole breast radiotherapy IGRT QA Goals To develop an end-to-end test to characterize isocenter coincidence: Planning CT MV kv CBCT AlignRT To incorporate additional tests: DICOM transfer Translational & rotation accuracy Multiple Imaging Modality Isocentricity (MiMi) Phantom from Standard Imaging MiMi Phantom Makeover Easy Alignment due to Unique Design: The MIMI Phantom incorporates five bone equivalent rods uniquely set so that four of them intersect at 90 o angles when viewed in DRRs or a 2D projection image. The rods traverse the entire phantom making them visible in any image or slice allowing for easy 2D/2D and 3D/3D matching for fast verification of isocenter position. Test Integrated System Accuracy of: 3D Cone Beam MV/kV Lasers and Couch Table Adjustments Optical Guidance Systems Test Automatic Table Adjustments: Additional cross-hair markers that are offset known distances from the true isocenter allow for verification of the shifts prescribed by automatic table positioning systems. 4

5 MiMi Phantom Makeover Al 2 O 3 Al 2 O 3 Isocenter Coincidence Testing Axial CT Scanning (0.75mm) DICOM Transfer to TPS (Verify Geometric accuracy) Align Phantom to Lasers DICOM Transfer from TPS with DRRs and RT Structures MV 1MU Al 2 O 3 kv ZrO 2 Center Phantom in Radiation Isocenter by MV imaging at 4 orthogonal angles Measure Offset to kv Isocenter by imaging at 4 orthogonal angles Measure Offset to CBCT Isocenter Introduce Known Physical Shift & Measure Accuracy Measure Offset to AlignRT Isocenter Measure Offset to Laser Isocenter MiMi Phantom Axial CT Scans Center Phantom in MV Isocenter by Imaging at 4 Gantry Angles Measure Offset to kv Isocenter by 2D/2D Match at 4 Angles Measure Offset to CBCT Isocenter Dependent upon CBCT Technique! 5

6 Measure Offset to Laser Isocenter Root Mean Square Distances of IGRT Isocenter Offsets IGRT isocenters aligned to SRS isocenter at our institution! TG-142: Imaging & Treatment Isocenter Coincidence Learning Objective Describe acceptance testing procedures for surface imaging systems Understand how surface imaging isocenter congruence testing must be incorporated into ongoing quality assurance of IGRT Describe the commissioning process of surface imaging systems for whole breast radiotherapy Commissioning Commissioning tests should be developed by the institution s physics team to explore in detail every aspect of the system with the goal of developing a comprehensive baseline characterization of the performance of the system. (TG-101) Commissioning: Whole-Breast RT While the patient s surface is a good surrogate for the target, the quality of 3D surface registration could be compromised by deformation We investigated the reliability of 3D surface matching using AlignRT compared to positioning using skin marks followed by MV portal imaging for whole-breast radiotherapy (WBRT) 6

7 WBRT: Methods Supine Immobilization with both arms above the head: Breastboard (n=2) Alphacradle (n=3) No respiratory gating Alignment for breast positioning includes: Setup to skin marks using lasers daily MV verification & alignment weekly Orthogonal pair (AP & LAT) Tangential portal images Commissioning: WBRT 3D surface generated from their free-breathing CT scan was used as reference Absolute or Relative positioning? Absolute: Inter-fraction reproducibility? Could reduce frequency of filming? Relative: Intra-fraction reproducibility? Relies on use of other IGRT modality WBRT: Methods Commissioning: WBRT Average 3D distance: 3.8 ± 5.1 mm: table shifts after MV imaging 5.9 ± 4.1mm: AlignRT registration of all 5.6 ± 3.3mm: AlignRT registration of Breast 3D Surface from CT data All ROI Breast ROI Regression analysis indicated that only all registration correlated with table shifts Table shifts were detected accurately with AlignR for both ROIs (r 0.9) WBRT: Correlation with MV WBRT: Correlation with Table Shifts Results presented in Poster SU-E-T-228 7

8 Commissioning: WBRT Registration of Breast ROI Absolute table rotations calculated by AlignRT for consecutively acquired surfaces exhibited fluctuations that were significantly larger for registrations of: Breast : 0.62 o ±0.58 o All : 0.27 o ±0.18 o Registration of All yielded rotations of 1.32 o and 1.37 o. WBRT: Conclusions Breast surface registration depends upon the ROI A large ROI showed: Higher correlation with MV shifts Increased stability when calculating table rotations Before clinical use, we must investigate the sources of error that prevent a direct one-to-one correlation between AlignRT and MV 3D shifts including: Respiratory motion Misalignment between the MV and AlignRT isocenters Deformation of breast surface Misinterpretation errors due to MV imaging Acceptance & Commissioning: Conclusions Surface imaging systems must undergo extensive acceptance testing & commissioning Isocenter congruence must be verified for all imaging modalities simultaneously Comprehensive testing of VisionRT product revealed in-operational components Commissioning of AlignRT indicates that discrepancies are larger for: Registration to CT reference surface Registration of smaller ROIs Utility of Surface Imaging? Reduce filming frequency Requires absolute positioning? Improve patient safety Particularly for multiple isocenter treatments Improve intra-fraction positioning Real-time monitoring throughout treatment Improve throughput Utility of Surface Imaging? Variability in repositioning is dominated by the ability of therapists to make small, controlled changes in the position of the patient. (Milliken et al., Int J Rad Onc Biol Phys, 38(4): , 1997) Surface imaging does not preclude need for: Good immobilization Adequate PTV margins Common sense! 8

9 Acknowledgements Charles Pelizzar, Ph.D., Director of Physics Kamil Yenice, Ph.D., Clinical Physics Director Karl Farrey, M.S., Dosimetry/Physics Manager Rodney Wiersma, Ph.D., Assistant Professor y,, Hyejoo Kang, Ph.D., Research Associate Ji Li, Ph.D., Physics Resident Dimple Modgil, Ph.D., Physics Resident 9