ABSTRACT
The detection and delineation of vascular structures from computed tomography (CT) angiography (CTA) data is often a prerequisite for diagnosis, treatment planning, and follow-up studies in clinical applications. To this end, a (iodine-based) contrast agent is typically injected to give blood a higher attenuation, allowing its better distinction from surrounding tissues. Such examinations using a CT device are then referred to as CTA. With recent advances in image acquisition technologies, the spatial resolution of image data has been signiŽcantly increased. Consequently, the number of slices that need to
CONTENTS
9.1 Introduction ........................................................................................................................ 189 9.1.1 Anatomy .................................................................................................................. 190 9.1.2 Vascular Diseases .................................................................................................. 190
9.1.2.1 Pulmonary Embolism ............................................................................ 191 9.1.2.2 Pulmonary Hypertension ...................................................................... 193
9.2 Vessel Segmentation .......................................................................................................... 193 9.2.1 Intensity-Based Approaches ................................................................................. 194 9.2.2 Vessel Enhancement .............................................................................................. 197 9.2.3 Vesselness-Based Approaches ............................................................................. 199 9.2.4 Fuzzy Segmentation .............................................................................................. 201
9.2.4.1 Core Component IdentiŽcation ............................................................. 202 9.2.4.2 Fuzzy Vessel Segmentation ................................................................... 204 9.2.4.3 Probability Function ............................................................................... 205 9.2.4.4 Centerline Extraction .............................................................................. 208
9.2.5 Artery-Vein Separation ......................................................................................... 208 9.3 Applications ........................................................................................................................ 210
9.3.1 PE-CAD ................................................................................................................... 210 9.3.2 Lung Nodule-CAD ................................................................................................ 212 9.3.3 Airway Segmentation and Virtual Bronchoscopy ............................................ 213 9.3.4 Lung Fissures and Lung Registration ................................................................. 214
9.4 Summary and Conclusions .............................................................................................. 215 References ..................................................................................................................................... 215
be read by the physician during a typical examination has been constantly increasing, whereas CT images reveal increasingly more details, allowing an improved diagnosis of diseases that had been previously difŽ cult to detect. Considering the limited time a radiologist may typically spend on a case in clinical routine, imaging applications that support the examiner's tasks and optimize the clinical work¸ow have become more and more important.
