On the accuracy of densitometric measurements of coronary artery stenosis based on Lambert-Beer’s absorption law

Summary

Densitometry of coronary arteries is often thought to be independent of the size and shape of the lumens involved. This belief results from the postulation that Lambert-Beer’s absorption law (LB’s law) is the appropriate theoretical basis for densitometric measurements on coronary arteries. Experiments with certain phantoms yield, however, results incompatible with the LB-model. For instance, a dependence of the obtained degrees of stenosis on the respective size, shape and rotational orientation (i.e. around the vessel axis and with respect to the incident X-rays) of the lumens can be observed. Since this dependence is not removed by compensation for the nonlinearity introduced by the cinefilm, one must admit that LB’s law is not the adequate basis for densitometric measurements of coronary artery stenosis, and that conventional densitometry (i.e. based on LB’s law) is likely to produce systematic errors. Using a previously designed, more adequate mathematical-physical model which takes X-ray polychromasy and scattered radiation into account, the magnitude of errors of conventionally obtained degrees of stenosis was assessed, as well as the magnitude of errors affecting the residual luminal area, this latter one being obtained from the erroneously assessed degree of stenosis and the area of the intact lumen (determined itself by calibrated diameter measurement).