Quality assurance of a robotic, image guided radiosurgery system

Abstract

The CyberKnife^® (Accuray Inc., Sunnyvale, CA, USA) technology constitutes a new concept in the precise delivery of therapeutic doses for stereotactic radiosurgery (SRS)[1]. Image guidance and highly accurate robotics replace the sharp fixation which characterized SRS systems in the past. The overall system targeting precision is specified to be below 1 mm [2]. Evolving from our growing experience with the system we have developed a QA schedule which exceeds manufacturer recommendations and includes specific tests of all involved system components: robot, imaging and the target location system (TLS), as well as the motion tracking system (MTS) for moving targets. Overall system performance is evaluated with a phantom-based end-to-end (E2E) test for various tracking modalities. The results are confirming the manufacturers specifications as well as the system stability: the ‘Manipulator Positioning Test’ revealed a mean deviation 0.19 mm (max. deviation for a single robot position: 0.7 mm) from the initial system calibration. Target tracking results showed differences of 0.30 ± 0.05 mm for translations and 0.33° ± 0.10° for rotations between nominal and measured phantom position. The daily AQA test resulted in a mean radial error of 0.46 ± 0.17 mm (494 tests), This corresponds well with the results of our E2E tests: 0.47 ± 0.24 mm. Our QA measures have proven to be suitable for verification of CK system performance. We are able to trace back a general decrease in precision to individual system components, allowing for specific recalibration measures.