Teboh F. Roland MS, Presenter: Nothing to Disclose

Nikos Papanikolaou PhD, Abstract Co-Author: Nothing to Disclose

In the course of dynamic radiotherapy to compensate for motion, system latency, response time or residual organ motion effects can be minimized but not eliminated. The dose prediction step should therefore take this delivery limitation into account. In this study, we use in-phantom measurements to derive a geometric error distribution associated with system latency in the course of dynamic radiotherapy of tumors affected by organ motion.

1-D sinusoidal tumor motion in the superior-inferior axis with 1cm amplitude was replicated on a respiratory motion phantom (QUASAR). A simple AP field from a LINAC was irradiated on the phantom with radiochromic film placed in a coronal plane to measure the dose distribution for three cases; (a) involving a static target and static field, (b) involving a moving target and moving synchronized field (c) involving a moving target and a static field. The expected dose distribution obtained in case (b) will be the convolution of the dose distribution obtained in (a) and the geomtric error distribution due to system latency. Similarly, the expected dose distribution in (c) will be the convolution of (a) with the geometric error distribution associated with respiratory motion. Therefore the error distribution can be derived from the measured dose distribution via an inverse convolution approach. This was implemented in MATLAB

Gaussian - like geometric error distributions associated with unaccounted organ motion (case c) and distributions associated with system latency (case b) were derived in this study that accurately described our system. Quantifiers similar to FWHM will be considered for future comparison studies involving system latency effects.

With dynamic radiotherapy delivery capability to compensate for organ motion, we can derive geometric error distributions associated with system latency effects directly from in-phantom measurements by application of the inverse process of convolution. Once the error distribution is known, the system latency effect can be properly accounted for in the dose prediction step by convolving the ideal case predicted dose with the geometric error distribution.

This is a measurement based technique to properly account for dose perturbations resulting from unavoidable system latency associated with dynamic dose delivery for tumors affected by organ motion.

Roland, T, Papanikolaou, N, Incorporating System Latency Effects Associated with Dynamic Radiotherapy Delivery in the Dose Prediction Step: Preliminary Results Based on Phantom Measurements.  Radiological Society of North America 2008 Scientific Assembly and Annual Meeting, February 18 - February 20, 2008 ,Chicago IL. http://archive.rsna.org/2008/6011936.html