Quantitative coronary arteriography: efficient correction of catheter calibrated vessel measurement

Abstract

1. BACKGROUND. To quantify coronary dimensions from digitized angiograms, the coronary catheter is commonly used as a scaling device. However, significant errors may result due to different angiographic magnification (DM) of the vessel and of the catheter resulting from deviating locations in the X-ray field. These errors can be corrected by biplane angiography. Since this correction needs a gantry measurement system, DM correction is less widely used. We analyzed the magnitude of DM that affects the accuracy of catheter calibrated vessel dimensions and developed DM corrections requiring low computational and measurement effort. 2. MATERIAL AND METHODS. We filmed biplane a perspex bloc of vessel phantoms (0.3 - 4.5 mm) in the isocenter of the x-ray system to get the true pixel sizes before any calibration. The phantoms were detected in the digitized images using a cardiac workstation (Kontron Cardio 500) and calibrated with differently located catheters. The calibration error was then calculated for increasing distances phantom-catheter. For DM correction, we developed procedures decreasing: (i) the computational effort (E) by approximated (A) formulas, as well as (ii) the measurement effort; this decrease approximates measured gantry settings (M) through fixed pre-settings for distances (D) and videochain (V). The several DM corrections were applied on the phantom images to analyze their correction performance and remaining calibration error. 3. RESULTS. The calibration of correctly detected pixel dimensions by a differentially magnified catheter causes a deviation of the absolute vessel dimensions, increasing with the vessel size. However, as shown in Table 1, DM correction reduces this calibration error substantially. Table 1: Calibration errors of catheter calibration and performance of DM corrections with decreasing effort (50 mm location distance phantom-catheter) 4. CONCLUSIONS. DM causes errors of vessel diameters of up to 15%. Even a measurement-free procedure (fixed presettings) corrects 71% of DM. Thus, the only additional effort of selecting an angiogram from the opposite view and marking the vessel and catheter locations in both images should be accepted.