Imaging of Traumatic Injury to Neurovasculature.

Trauma involving the vessels of the head and neck portends risk of hemorrhage, cerebral infarction, and death. Imaging is essential for the timely diagnosis of these injuries. This article will review diagnostic imaging findings and pitfalls as well as related pathophysiology, screening risk factors, management, and follow-up recommendations.

Neurointerventional management of low-flow vascular malformations of the head and neck.

Mulliken and Glowacki's seminal classification of vascular anomalies into vascular tumors (with infantile hemangiomas being paradigmatic) versus nontumorous vascular malformations has been as important in the head and neck region as elsewhere. These latter are congenital, have an equal gender incidence, virtually always grow in size with the patient during childhood, and virtually never involute spontaneously. The vascular malformations can in turn be subclassified into high-flow and low-flow. Our focus is on the low-flow malformations, which include those with venous, lymphatic, and, to a lesser extent, capillary components. We address diagnostic and clinical characteristics, particularly insofar as they relate to the structures of the head and neck, and discuss neurointerventional management in some detail.

Estimation of Cardiac Respiratory-Motion by Semi-Automatic Segmentation and Registration of Non-Contrast-Enhanced 4D-CT Cardiac Datasets.

The goal of this work is to investigate, for a large set of patients, the motion of the heart with respiration during free-breathing supine medical imaging. For this purpose we analyzed the motion of the heart in 32 non-contrast enhanced respiratory-gated 4D-CT datasets acquired during quiet unconstrained breathing. The respiratory-gated CT images covered the cardiac region and were acquired at each of 10 stages of the respiratory cycle, with the first stage being end-inspiration. We devised a 3-D semi-automated segmentation algorithm that segments the heart in the 4D-CT datasets acquired without contrast enhancement for use in estimating respiratory motion of the heart. Our semi-automated segmentation results were compared against interactive hand segmentations of the coronal slices by a cardiologist and a radiologist. The pairwise difference in segmentation among the algorithm and the physicians was on the average 11% and 10% of the total average segmented volume across the patient, with a couple of patients as outliers above the 95% agreement limit. The mean difference among the two physicians was 8% with an outlier above the 95% agreement limit. The 3-D segmentation was an order of magnitude faster than the Physicians' manual segmentation and represents significant reduction of Physicians' time. The segmented first stages of respiration were used in 12 degree-of-freedom (DOF) affine registration to estimate the motion at each subsequent stage of respiration. The registration results from the 32 patients indicate that the translation in the superior-inferior direction was the largest component motion, with a maximum of 10.7 mm, mean of 6.4 mm, and standard deviation of 2.2 mm. Translation in the anterior-posterior direction was the next largest component of motion, with a maximum of 4.0 mm, mean of 1.7 mm, and standard deviation of 1.0 mm. Rotation about the right-left axis was on average the largest component of rotation observed, with a maximum of 4.6 degrees, mean of 1.6 degrees, and standard deviation of 2.1 degrees. The other rotation and shear parameters were all close to zero on average indicting the motion could be reasonably well approximated by rigid-body motion. However, the product of the three scale factors averaged about 0.97 indicating the possibility of a small decrease in heart volume with expiration. The motion results were similar whether we used the Physician's segmentation or the 3-D algorithm.