Quality enhancement of arterial measurements based on a more accurate stochastic model of X-ray cine-angiography

The problem of estimating the diameter of narrow coronary arteries with low measurement error and variability remains unresolved. Current methods for improving measurements increase accuracy at the expense of reproducibility and vice-versa. To address this problem, we have constructed a parametric model for X-ray cine-angiography that accurately captures the non-linearity of the process. In addition, we have derived a practical method for the estimation of the image model parameters directly from the angiographic data. These developments make it possible to use absolute image intensity values to improve the diameter estimation quality. We have characterized the performance and sensitivity of this technique by using Monte Carlo simulation and by using actual X-ray image data. Our results show that this estimator is robust to the expected range of encountered imaging conditions and produces lower estimation error and variability than existing methods. This quality enhancement is greatest for narrow vessel widths associated with severe coronary disease.