Dynamic Couch Motion Many vendors are beginning to allow couch motion during radiation delivery. Varian developer mode allows institutions to perform research using these types of treatments. Tomotherapy uses one directional dynamic couch motion to produce helical treatments. Can the same thing be done with a conventional linac?
Helical Delivery Goals: Treat long targets with a single setup by moving the couch longitudinally during gantry rotation. Plan rapid-arc helical delivery using modified VMAT optimization. Produce plans comparable to those achievable with Tomotherapy. Measure a treatment with arccheck as a proof of concept 2007 RUSH University Medical Center
Challenges A. Developer mode is expensive and has a significant learning curve. B. Eclipse VMAT optimization does not allow dynamic couch motion. C. Modern linac gantries do not allow continual rotation, i.e. can t pass 185 E. D. Arccheck can not measure treatment fields longer than 20cm.
Modeling Helical Delivery VMAT optimization with Eclipse works even if the arcs have different isocenters. 120 Arcs 5cm shifts Helical delivery was modeled by splitting the helical revolution into a series of transverse arcs The dynamic couch can then be modeled by a longitudinal shift between the isocenter of each arc. Eclipse does have a limit of 10 arcs per treatment plan.
Minimum Modulation per Slice Typical spine SBRT treatments use two full arcs, so each transverse slice gets 720 of modulation. Both true helical delivery and the arc-shift approximation mean each transverse slice receives less modulation. Amount of modulation depends heavily on helical pitch How much modulation is sufficient? Typical spine SBRT 720 120 240 360 240 120 Arc-shift approximation FOV 2 Arcs 360 Arc 6 120 Arc 5 120 Arc 4 120 Arc 3 120 Arc 2 120 Arc 1 120
Minimum Modulation per Slice 2D optimization tests were run for both brain and spine. A single arc was used length between 90 and 720 Minimal gains were seen for arc length 180-270 Brain PTV Optic nerves Spine PTV bone marrow Maximum shift must allow at least 180 of modulation per slice.
Planning Tests Arc-shift approximation planning was done on data from 5 CSI patients. Arc length was 120 Shift was varied between 5, 10, 15, and 20cm. (20cm has ~360 of modulation per slice) PTV height varied from 71 to 77cm; this meant more than 10 arcs were needed for all plans. Plans were optimized in 10 arc portions, with later portions being optimized on the partial dose.
Plan A 7 Arcs Plan B 8 Arcs Couch shift between arcs Plan C 7 Arcs 120 Transverse arcs 2007 RUSH University Medical Center
Dose (Gy) Dose (Gy) Individual Case Results Results from two cases comparing effect of shift. 50 45 Case 1 5 cm/rev 10 cm/rev 50 45 Case 2 40 15 cm/rev 40 35 20 cm/rev 35 30 30 25 25 20 20 15 15 10 10 5 5 0 PTV Max PTV 95% PTV Min Lenses Max Bone Marrow Mean Esophagus Max 0 PTV Max PTV 95% PTV Min Lenses Max Shift has little effect on dosimetric results. Bone Marrow Mean Esophagus Max 2007 RUSH University Medical Center
Helical SBRT and Tomotherapy Tomotherapy Comparsion One case was optimized using the same criteria as a clinical tomotherapy treatment. The clinical prescription involved three dose targets of 45, 50, and 55 Gy, as well as OAR restrictions to critical structures. A shift of 20cm was used for the helical arc-shift plan.
Volume (%) Helical SBRT and Tomotherapy 100 90 80 70 60 50 40 PTV 45 (Helical) PTV45 (Tomo) PTV 50 (Helical) PTV50 (Tomo) PTV 55 (Helical) PTV55 (Tomo) PTV Comparison 30 20 10 0 30 35 40 45 50 55 60 65 Dose (Gy) 2007 RUSH University Medical Center
Volume (%) Helical SBRT and Tomotherapy 100 90 80 70 60 50 40 30 20 10 0 Eye Doses 0 10 20 30 40 50 Dose (Gy) 2007 RUSH University Medical Center Lens Lt (Helical) Lens Lt (Tomo) Lens Rt (Helical) Lens Rt (Tomo) Eye Lt (Helical) Eye Lt (Tomo) Eye Rt (Helical) Eye Rt (Tomo)
Helical SBRT and Tomotherapy 100 90 80 70 60 50 OAR Doses Both Lungs (Helical) Both Lungs (Tomo) Bone Marrow (Helical) Bone Marrow (Tomo) Cardiac Ventricle (Helical) Cardiac Ventricles (Tomo) Esophagus (Helical) Esophagus (Tomo) 40 30 20 10 0 0 10 20 30 40 50 60 2007 RUSH University Medical Center
Deviations from True Helical MLC positions at the end of one 120 arc do not impose any limitations on leaf positions at the start of the next arc. Arcs in which the gantry passes 180 are currently invalid, though slip-ring technology may be in the next generation of linacs. Optimization using multiple plans can produce boundary effects. 2007 RUSH University Medical Center
Proof of Concept One plan was delivered to arccheck as a proof of concept. Each of the 12 arcs were delivered with arccheck at isocenter and no couch shifts. Using a gamma criterion of 3%/2mm, individual arcs passed with an average rate of 98%. Data was extracted, shifted and summed to produce a full treatment measurement.
Proof of Concept 90-80- Sum of individual 120 arcs. 70-60- 50-40- 30-20- 10-0- 180 270 0 90 180 Brain region is clearly evident, as are portions of spine. Inferior hot spot due to end effect. Transverse discontinuities are due to lack of MLC matching and optimizing across multiple plans.
Summary Helical CSI treatment using a conventional linac was modelled using available technology. Optimized plans were found to have comparable results to Tomotherapy treatments. One modelled plan was delivered and measurements combined as proof of concept.