A Study On The Airway Geometry Measurement Of The Parameter Fitting Algorithm Based On The Mathematical Model Of The CT Scanning Process

Chronic Obstructive Pulmonary Disease is a kind of disease that seriously destruct human health. Geometry measurement of intrathoracic airways with high accuracy based on2-D CT images is significant for the dignosis and treatment of COPD patients.The most commonly used method nowadays is the technique of FWHM-full width at half maximum, which can only obtain comparativly ideal measurement results for thick-walled airways with the airway axis perpendicular to the CT image plane, yet is not suitable for airways with different sizes.In order to improve the accuracy of geometry measurement for small airways, a parameter fitting algorithm based on the mathematical model of the CT scanning process was adopted in this paper. In this method, Bresenham algorithm was firstly used to examine the gray levels of the rays that cast through the airway wall to obtain the true gray level profiles of the rays. The mathematical model was then established in which the blurring introduced during CT scanning was modeled. Finally, the algorithm of Levenberg-Marquardt was employed to match the true gray level profiles and the predicted gray level profiles that was obtained from the mathematical model to get estimates of the parameters of the model, in which the parameters-the estimates of the inner and outer diameters were of the least difference with the true gray level profiles. Since the blurring introduced during the scanning process was explicitly modeled, the model based method in this paper can be accurately applied to airways of different sizes.To verify the validity of the model based method in this paper, comparative experiment were carried out between FWHM method and the model based method using a phantom, in which7ideal2D ring airway phantoms with different diameters were simulated using Monte Carlo EGSnrc(V4r2.3.2). Results showed that the model based method gave measurements with higher accuracy and greater stability when measuring thin-walled airway models. FWHM method and the model based method were then used in the inspiratory CT images of the intrathoracic airways of RB1from the3rd to the5th generation, respectively. Results showed that both of the methods can yield stable measurements, the main difference occurred when measuring the5th generation, the model based method can give more stable results than FWHM method.