Program Information

Near Real-Time Analysis of Radiation Delivery and Imaging, Accuracy to Ensure Patient Safety

K Wijesooriya

K Wijesooriya*, K Seitter , V Desai , P Read , J Larner , University of Virginia Health Systems, Charlottesville, VA

Presentations

SU-E-T-248 Sunday 3:00PM - 6:00PM Room: Exhibit Hall

Purpose:To develop and optimize an effective software method for comparing planned to delivered control point machine parameters for all VARIAN TrueBeam treatments so as to permit (1) assessment of a large patient pool throughout their treatment course to quantify treatment technique specific systematic and random uncertainty of observables, (2) quantify the site specific daily imaging shifts required for target alignment, and (3) define tolerance levels for mechanical parameters and imaging parameters based on statistical analysis data gathered, and the dosimetric impact of variations.

Methods:Treatment and imaging log files were directly compared to plan parameters for Eclipse and Pinnacle planned treatments via 3D, IMRT, control point, RapidArc, and electrons. Each control point from all beams/arcs (7984) for all fractions (1940) of all patients treated over six months were analyzed. At each control point gantry angle, collimator angle, couch angle, jaw positions, MLC positions, MU were compared. Additionally per-treatment isocenter shifts were calculated. Results were analyzed as a whole in treatment type subsets: IMRT, 3D, RapidArc; and in treatment site subsets: brain, chest/mediastinum, esophagus, H&N, lung, pelvis, prostate.

Results:Daily imaging isocenter shifts from initial external tattoo alignment were dependent on the treatment site with <0.5 cm translational shifts for H&N, Brain, and lung SBRT, while pelvis, esophagus shifts were ~1 cm. Mechanical delivery parameters were within tolerance levels for all sub-beams. The largest variations were for RapidArc plans: gantry angle 0.11±0.12,collimator angle 0.00±0.00, jaw positions 0.48±0.26, MLC leaf positions 0.66±0.08, MU 0.14±0.34.

Conclusion:Per-control point validation reveals deviations between planned and delivered parameters. If used in a near real-time error checking system, patient safety can be improved by equipping the treatment delivery system with additional forcing functions which by-pass human error avenues.


Contact Email: