THE CUMULATIVE VERIFICATION IMAGE-ANALYSIS TOOL FOR OFFLINE EVALUATION OF PORTAL IMAGES

Purpose: Daily portal images acquired using electronic portal imaging devices contain important information about the setup variation of the individual patient, The data can be used to evaluate the treatment an d to derive correction for the individual patient. The large volume of images also require software tools for efficient analysis, This artic le describes the approach of cumulative verification image analysis (C VIA) specifically designed as an offline tool to extract quantitative information from daily portal images. Methods and Materials: The user interface, image and graphics display, and algorithms of the CVIA tool have been implemented in ANSCI C using the X Window graphics standard s. The tool consists of three major components: (a) definition of trea tment geometry and anatomical information; (b) registration of portal images with a reference image to determine setup variation; and (c) qu antitative analysis of all setup variation measurements, The CVIA tool is not automated, User interaction is required and preferred, Success ful alignment of anatomies on portal images at present remains mostly dependent on clinical judgment, Predefined templates of block shapes a nd anatomies are used for image registration to enhance efficiency, ta king advantage of the fact that much of the tool's operation is repeat ed in the analysis of daily portal images. Results: The CVIA tool is p ortable and has been implemented on workstations with different operat ing systems, Analysis of 20 sequential daily portal images can be comp leted in less than 1 h, The temporal information is used to characteri ze setup variation in terms of its systematic, random and time-depende nt components, The cumulative information is used to derive block over lap isofrequency distributions (BOIDs), which quantify the effective c overage of the prescribed treatment area throughout the course of trea tment, Finally, a set of software utilities is available to facilitate feedback of the information for treatment plan recalculation and to t est various decision strategies for treatment adjustment. Conclusions: The CVIA tool provides comprehensive analysis of daily images acquire d with electronic portal imaging devices, Its offline approach allows characterization of the nature of setup variation for the individual p atient that would have been difficult to deduce using only a few daily or weekly portal images, Distribution of the tool will help establish an important database of setup variation from many clinics, The infor mation derived from CVIA can also serve as the foundation to integrate treatment verification, treatment planning, and treatment delivery.