Effect of data compression on quantitative coronary measurements

Abstract

Digital coronary and left ventricular angiography demand high transfer rates and very large data storage if all the clinical data are to be archived. If appropriate compression schemes were available without compromising the quality and resolution of the image data, such demands could be lessened. In this study we compared the influence of different compression factors of the Adaptive Real Time Image Compression (ARTIC) scheme used on the Philips DCI‐SX systems on coronary measurements assessed with the Automated Coronary Analysis (ACA) package. Loss‐free acquired images of size 512² × 8 bits, which had been stored digitally on tape, were reloaded into the DCI with compression factors of 2, 3, and 4; only the factor 2 is loss free. To evaluate the effect of the different data compressions on the accuracy of the measurements, the diameters of a vessel phantom (tube sizes ranging from 0.687 to 5.062 mm) were determined. To evaluate the reproducibility of the results, the intraobserver variability was determined for the different compression factors from 40 coronary obstructions. The differences in the reference diameter measurements of the vessel phantom were ‐0.03 ± 0.06 mm, 0.01 ± 0.07 mm, and 0.04 ± 0.08 mm for the compression factors 2 (loss free), 3, and 4, respectively. The results were not statistically significantly different. The intraobserver variabilities in the obstruction diameter measurements of the coronary obstructions were ‐0.04 ± 0.13 mm, 0.00 ± 0.14 mm, and 0.02 ± 0.13 mm for the compression factors 2, 3 and 4, respectively. The intraobserver variabilities in the reference diameter measurements were ‐0.02 ± 0.12 mm, 0.01 ± 0.09 mm, and 0.03 ± 0.09 mm for the compression factors 2,3, and 4, respectively. The intraobserver variabilities of the percent diameter stenosis were 0.96 ± 4.19%, ‐0.01 ± 4.88%, and ‐0.04 ± 4.68% for the compression factors 2,3, and 4, respectively. None of these differences were statistically significant. Both from a qualitative and quantitative point of view, data compression factors 3 and 4 are acceptable in digital coronary arteriography.