How to Achieve 0 2 mm Isocenter Alignment

How to Achieve 0. 2 mm Isocenter Alignment Accuracy with XKnife. TM Jimm Grimm 1, Ph. D. , Shu-Ya Lisa Grimm 2, Ph. D. , Larry Simpson 1, Ph. D. , Dayee Jacob 1, M. S. , Abhirup Sarkar 1, M. S. , Hansen Chen 1, M. S. 1 Christiana Care Health Services, Newark, De 2 Albert Einstein Medical Center, Philadelphia, Pa AAPM - Radionics User Meeting - 7/22/2007 JGrimm@Christiana. Care. org 20070722 Currently under review by the FDA - 510 k Pending 1

Disclosure • Member of Speakers Bureau of Integra Life. Sciences. TM 20070722 Currently under review by the FDA - 510 k Pending 2

Isocenter Accuracy, mm Isocentric Accuracy of Stereotactic Systems 20070722 Currently under review by the FDA - 510 k Pending 3

• The SAlinac system uses the same XKnife. TM hardware you are accustomed to, with a few modifications w Cameras on the walls and ceiling for laser guidance w The laser target is a cube instead of single-sided 20070722 review by the FDA - 510 k Pending w Add a film scanner. Currently and under a PDA… 4

SAlinac system can enable the physicist to improve alignment from current 1. 5 mm isocenter accuracy to 0. 2 mm accuracy • SAlinac is an alternative system to MIS • It is an improved accuracy Linac-Based Stereotaxy System • It uses a computer program to analyze the film test. Based on the analyzed result, computer calculated/identify the position of the isocenter and gives advice to the physicist to move the test ball to the center • The accuracy of final film result is within 0. 2 mm 20070722 Currently under review by the FDA - 510 k Pending 5

Align. LTC Mode: Computer Assisted Laser Guidance Laser Target Cube (LTC) – targets on 3 sides • The PDA wirelessly gets live repositioning advice, based on w the most recent couch mount film test w and on near-real-time laser target images Red LCMA (y-axis) microadjust knob 20070722 Currently under review by the FDA - 510 k Pending 6

What if the lasers are out of alignment? 20070722 Currently under review by the FDA - 510 k Pending 7

Align. Laser Mode: Computer Assisted Laser Alignment • The PDA receives live laser adjustment advice over a wireless network • Absolute laser position is calculated from laser target images and the most recent film analysis 20070722 Currently under review by the FDA - 510 k Pending 8

Align the ball to the laser and do a couch mount QA film test Right Laser Film Right Digital Camera Light Field, Ball Shadow Radiopaque QA ball Stereotactic treatment cone 20070722 Currently under review by the FDA - 510 k Pending 9

Couch Mount QA Film Test, AP Shot 20070722 Currently under review by the FDA - 510 k Pending 10

Film. Analysis Mode: Measure Absolute Alignment x-coordinate beam centroid x-coordinate ball centroid Ball Contour FWHM Beam Contour Window/Level has been adjusted to show the ball shadow; so the beam near FWHM and below is not visible 20070722 Currently under review by the FDA - 510 k Pending 11

Objection! Cart Before the Horse Problems! • Every mode depends on every other mode w How can a couch mount film test be good if the lasers are misaligned? w How can you align the lasers if the target is not at the isocenter? w How can you put the target at isocenter if the lasers are not aligned? • It still seems hopelessly convoluted. . . • Unless we can compensate for and correct any type of error! 20070722 Currently under review by the FDA - 510 k Pending 12

Setting the Errors Straight • Film. Analysis mode can accurately measure ball position offsets up to 3 mm w It's easy to put the ball within 3 mm of isocenter even with poorly aligned lasers or just the light field • The program gives advice how far to adjust the x, y, and z coordinates of the ball w It even advises you which way to turn each knob • Laser. Analysis modes can tolerate laser offsets up to 3 mm w Even if the lasers are initially way off, we can use them as a relative reference to guide the repositioning of the laser target 20070722 Currently under review by the FDA - 510 k Pending 13

Daily Isocenter Alignment Procedure • The system has three modes which can be run in any order the physicist desires: w Align. LTC mode: Align the LTC to isocenter based on the lasers w Align. Laser mode: Align the lasers to the isocenter (if needed) w Film. Analysis mode: Determine how close the ball is to isocenter • A final film test must always end the session • The Winston-Lutz film test is still the authority on alignment accuracy • As long as the system is used every week or so, and as long as the lasers are mounted well on a stable surface, usually one Align. LTC and one Film. Analysis get the QA ball within 0. 2 mm of isocenter 20070722 Currently under review by the FDA - 510 k Pending 14

Clinical Results With respect to the centroid of the linac's isocenter 20070722 Currently under review by the FDA - 510 k Pending 15

SAlinac system algorithm • • • The program uses full width at half maximum (FWHM) algorithm to find the center of the radiation field and the center of the ball Linear and nonlinear filtering is employed to make the contours more accurate The alignment error in each shot is ball centroid minus beam centroid Three-dimensional reconstruction from all shots is calculated Program will calculate the distance between these two isocenters and suggest the direction of the adjustment Orientation, with respect to a supine, headfirst patient: +x: patient's left +y: superior (towards gantry) +z: anterior (towards ceiling) 20070722 Currently under review by the FDA - 510 k Pending 16

Sag • The largest remaining source of error is gantry sag w The weight from gravity on the gantry causes it to sag • Most linacs have 0. 2 – 0. 4 mm sag • Our program can help you center the AP sag and the PA sag to minimize the effects • Compared to end-to-end performance (including CT and physician contouring uncertainty), the sag should not be a problem 20070722 Currently under review by the FDA - 510 k Pending 17

Steel on Steel • Radionics frames have amazingly good precision • I believe we are limited more by w ability to precisely quantify alignment errors w ability to accurately correct alignment errors • Than we are limited by mechanical precision 20070722 Currently under review by the FDA - 510 k Pending 18

At a scale of 0. 1 mm, everything is moving! • Gantry may drift slightly • Cone offset may change • Couch axis may shift slightly • The MIS can't adapt to any of these w It may drift slightly over time as well • We need an adaptive system that automatically re-calibrates itself! Mechanical Isocenter Stand (MIS) 20070722 Currently under review by the FDA - 510 k Pending 19

System Benefits • Accurate and fast – enables physicist to usually get within 0. 2 mm of isocenter using a single couch mount film test w If the goal is not met on the first film test, the program updates the position estimates and provides repositioning advice w A final film test is taken whenever the couch mount or lasers need to be repositioned • The system can be used to troubleshoot and correct gantry skew and collimator offset problems • The new system tracks slight gantry, couch, and collimator movements every time you use it, and always adapts to minimize the errors 20070722 Currently under review by the FDA - 510 k Pending 20

Claim: • The SAlinac system can enable the physicist w w To position the room lasers And the QA ball Within 0. 2 mm of the centroid of the isocenter Of most linear accelerators equipped with stereotactic cones • We don't want to abuse this definition though! w Even with a 1 mm gantry skew you could put the QA ball within 0. 2 mm of the centroid of the isocenter, but that would not be ideal for patients w The SAlinac system can help you fix gantry skew on many linacs 20070722 Currently under review by the FDA - 510 k Pending 21

Conclusion • With the addition of: w state of the art electronics and w advanced image processing algorithms • SAlinac enables the physicist to achieve 0. 2 mm isocenter alignment • This should serve as a good foundation for future research to improve end-to-end accuracy Isocentric Accuracy Specification 20070722 Gamma Knife 0. 07 mm X-Knife 1. 5 mm Brain. Lab 0. 75 mm Linac Scalpel 0. 2 mm Cyber. Knife 0. 2 mm SAlinac 0. 2 mm Currently under review by the FDA - 510 k Pending 22

Thank you 20070722 Currently under review by the FDA - 510 k Pending 23