Dan Ruan PhD, Presenter: Nothing to Disclose

Stephen E Tenn, Abstract Co-Author: Nothing to Disclose

Daniel A. Low PhD, Abstract Co-Author: Nothing to Disclose

We apply the ALARA principles to IGRT, where dose-dependent image quality translates to motion estimates via image registration. The purpose of this study is to examine the impact of imaging technique on intensity-based registration results, and to determine the feasibility of using low dose image guidance in radiographic IGRT.

This study focuses on estimating rigid (translation + rotation) motion from intensity-based 2D-3D registration between radiographic images and a reference CT volume. IGRT motion estimation benefits from precedent estimates, so we use initialization in our registration. An anthropomorphic thorax phantom was scanned using a technique typical for thorax setup (high-dose) and registered to its CT to provide a baseline motion estimate. The same procedure was repeated, varying the technique ( 40~120 kvp, 10~160 mA and 4~160ms). The registration results were compared with the baseline and registration error was evaluated in terms of 3D translation error, maximum 1D translation error, and maximum angular error. Dose corresponding to each technique was quantified using kv^2*mAs, normalized to the baseline dose.

In general, motion estimation accuracy degrades with decreased imaging dose. However, this trend is mild, and a dose reduction of O(10^-3) w.r.t baseline technique was achieved with translational error within 0.5mm and rotational error within 1deg, despite noisy images. Among the technique factors, reduction of kvp deteriorated the registration accuracy the most, under equivalent dose condition with higher mAs. We will continue to investigate other body sites.

This study quantitatively studies the dependence of motion estimation accuracy on imaging techniques used in radiographic IGRT. There are 3 major observations: (1) Image guidance dose can be reduced significantly without much sacrifice in motion monitoring capability. The specific threshold for acceptable motion error is application-dependent. (2) Under equal imaging dose, techniques with higher kvp and lower mAs yields better registration results. (3) Intensity based registration is beneficial under low dose techniques, when landmark-based registration is prone to error due to noise.

This study provides a reference to establish imaging protocols for radiographic/fluoroscopic IGRT. It demonstrates the potential for using low-dose techniques for intrafractional motion monitoring.

Ruan, D, Tenn, S, Low, D, ALARA for IGRT: Low Dose Image Guidance for Markerless Motion Monitoring.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11008178.html