We propose a Parallel Banding Algorithm (PBA) on the GPU to compute the exact Euclidean Distance Transform (EDT) for a binary image in 2D and higher dimensions. Partitioning the image into small bands to process and then merging them concurrently, PBA computes the exact EDT with optimal linear total work, high level of parallelism and a good memory access pattern. This work is the first attempt to exploit the enormous power of the GPU in computing the exact EDT, while prior works are only on approximation. Compared to these other algorithms in our experiments, our exact algorithm is still a few times faster in 2D and 3D for most input sizes. We illustrate the use of our algorithm in applications such as computing the Euclidean skeleton using the integer medial axis transform, performing morphological operations of 3D volumetric data, and constructing 2D weighted centroidal Voronoi diagrams.

Purpose To validate a computed tomographic (CT) glomerular filtration rate (GFR) measurement and compare it with renal dynamic imaging GFR obtained by using the "Gates" method, with dual plasma sampling technetium 99m (Tc) diethylenetriaminepenta-acetic acid (DTPA) clearance ("true GFR") as the reference standard. Materials and Methods This prospective study was approved by the institutional review board, and written informed consent was obtained from all patients. Forty-two patients with unilateral renal disease were included. Single-kidney CT GFR was calculated as excretory phase whole-kidney CT number enhancement divided by the area under the time-attenuation curve for the aorta, multiplied by (1 - hematocrit level). The CT GFR was then obtained by summing the result of the two sides. The true GFR and the Gates GFR were measured by using a single injection of Tc-DTPA. The CT GFR and Gates GFR were respectively compared with the true GFR by using a paired t test and linear regression analysis. Results The difference between CT GFR (mean ± standard deviation, 96.02 mL/min ± 23.11) and true GFR (90.50 mL/min ± 21.46) was 5.51 mL/min ± 6.96 (P< .001), demonstrating 6.09% systemic overestimation. The difference between Gates GFR (93.93 mL/min ± 26.97) and true GFR was 3.42 mL/min ± 16.10 (P = .176). Linear regression findings confirmed the association between CT GFR (y-axis) and true GFR (x-axis) and between Gates GFR (y-axis) and true GFR (x-axis) (P < .001 for both). Both regression lines paralleled the diagonal (intercept = 0 and slope = 1) (P = .599 and P = .945, respectively). The 95% confidence interval of the former was above the diagonal, confirming the systemic overestimation. The standard deviations of residuals of both linear regressions were 7.02 mL/min and 16.30 mL/min, respectively, demonstrating smaller deviation of the CT GFR (P < .001). Conclusion The proposed CT GFR measurement was validated in this study and was proved to be more accurate than the Gates method despite slight (6.09%) systemic overestimation. RSNA, 2016 Online supplemental material is available for this article.

Background: Measurement of carotid intima-media thickness (IMT) has been shown to be beneficial in patients with increased CV risk. Patients with low CV risk have few events and further risk stratification has little benefit. We therefore we sought whether patients with a family history of premature CVD would benefit from measurement of IMT.Methods: We studied 226 patients (90 men; age 52 ± 6) with a first degree relative with premature atherosclerosis. Framingham 10-year risk was calculated from risk data. Patients with diabetes mellitus were excluded from this study. Intima-media thickness was acquired by a trained study nurse and measured offline by an experienced sonographer in the far wall of the common carotid artery within 2 cm of the bifurcation. Age-corrected values for the 75% CI from normal were used as a cut-off for analysis.Results: Two hundred and two patients were low risk, 12 intermediate and two high. BMI was increased (26 ± 4) and there was a mod incidence of smoking (43%) and HTN (48%). Total cholesterol (TC) was 5.3 ± 1.3 mmol/l and HDL was 1.47 ± 0.04 mmol/l; only 38% of patients exercised regularly. Intima-media thickness in the patient cohort was 0.68 ± 0.13 mm and there were significant differences in IMT between the low and intermediate and low and high risk groups but not between the intermediate and high groups. In the patients with low CV risk, 118 (52%) had increased age-corrected IMT. There were significant differences in smoking (61 vs 37; p < 0.0001) TC (4.94 vs 5.60; p < 0.0001) and HTN (62 vs. 48; p = 0.01) in the low risk group with normal and those with increased IMT.