Mani Vembar

Automatic Model-Based Segmentation of the Heart in CT Images

Automatic image processing methods are a pre-requisite to efficiently analyze the large amount of image data produced by computed tomography (CT) scanners during cardiac exams. This paper introduces a model-based approach for the fully automatic segmentation of the whole heart (four chambers, myocardium, and great vessels) from 3-D CT images. Model adaptation is done by progressively increasing the degrees-of-freedom of the allowed deformations. This improves convergence as well as segmentation accuracy. The heart is first localized in the image using a 3-D implementation of the generalized Hough transform. Pose misalignment is corrected by matching the model to the image making use of a global similarity transformation. The complex initialization of the multicompartment mesh is then addressed by assigning an affine transformation to each anatomical region of the model. Finally, a deformable adaptation is performed to accurately match the boundaries of the patients anatomy. A mean surface-to-surface error of 0.82 mm was measured in a leave-one-out quantitative validation carried out on 28 images. Moreover, the piecewise affine transformation introduced for mesh initialization and adaptation shows better interphase and interpatient shape variability characterization than commonly used principal component analysis.

Myocardium: Dynamic versus Single-Shot CT Perfusion Imaging

PURPOSE To determine the diagnostic accuracy of dynamic computed tomographic (CT) perfusion imaging of the myocardium for the detection of hemodynamically relevant coronary artery stenosis compared with the accuracy of coronary angiography and fractional flow reserve (FFR) measurement. MATERIALS AND METHODS This study was approved by the institutional review board and the Federal Radiation Safety Council (Bundesamt für Strahlenschutz). All patients provided written informed consent. Thirty-two consecutive patients in adenosine stress conditions underwent dynamic CT perfusion imaging (14 consecutive data sets) performed by using a 256-section scanner with an 8-cm detector and without table movement. Time to peak, area under the curve, upslope, and peak enhancement were determined after calculation of time-attenuation curves. In addition, myocardial blood flow (MBF) was determined quantitatively. Results were compared with those of coronary angiography and FFR measurement by using a receiver operating characteristic (ROC) analysis. In addition, threshold values based on the Youden index and sensitivity and specificity were calculated. RESULTS Area under the ROC curve, sensitivity, and specificity, respectively, were 0.67, 41.4% (95% confidence interval [CI]: 23.5%, 61.1%), and 86.6% (95% CI: 76.0%, 93.7%) for time to peak; 0.74, 58.6% (95% CI: 38.9%, 76.5%), and 83.6% (95% CI: 72.5%, 91.5%) for area under the curve; 0.87, 82.8% (95% CI: 64.2%, 94.1%), and 88.1% (95% CI: 77.8%, 94.7%) for upslope; 0.83, 82.8% (95% CI: 64.2%, 94.1%), and 89.6% (95% CI: 79.6%, 95.7%) for peak enhancement; and 0.86, 75.9% (95% CI: 56.5%, 89.7%), and 100% (95% CI: 94.6%, 100%) for MBF. The thresholds determined by using the Youden index were 148.5 HU · sec for area under the curve, 12 seconds for time to peak, 2.5 HU/sec for upslope, 34 HU for peak enhancement, and 1.64 mL/g/min for MBF. CONCLUSION The semiquantitative parameters upslope and peak enhancement and the quantitative parameter MBF showed similar high diagnostic accuracy. SUPPLEMENTAL MATERIAL http://radiology.rsna.org/lookup/suppl/doi:10.1148/radiol.13121441/-/DC1.

Segmentation of the heart and great vessels in CT images using a model-based adaptation framework

Recently, model-based methods for the automatic segmentation of the heart chambers have been proposed. An important application of these methods is the characterization of the heart function. Heart models are, however, increasingly used for interventional guidance making it necessary to also extract the attached great vessels. It is, for instance, important to extract the left atrium and the proximal part of the pulmonary veins to support guidance of ablation procedures for atrial fibrillation treatment. For cardiac resynchronization therapy, a heart model including the coronary sinus is needed. We present a heart model comprising the four heart chambers and the attached great vessels. By assigning individual linear transformations to the heart chambers and to short tubular segments building the great vessels, variable sizes of the heart chambers and bending of the vessels can be described in a consistent way. A configurable algorithmic framework that we call adaptation engine matches the heart model automatically to cardiac CT angiography images in a multi-stage process. First, the heart is detected using a Generalized Hough Transformation. Subsequently, the heart chambers are adapted. This stage uses parametric as well as deformable mesh adaptation techniques. In the final stage, segments of the large vascular structures are successively activated and adapted. To optimize the computational performance, the adaptation engine can vary the mesh resolution and freeze already adapted mesh parts. The data used for validation were independent from the data used for model-building. Ground truth segmentations were generated for 37 CT data sets reconstructed at several cardiac phases from 17 patients. Segmentation errors were assessed for anatomical sub-structures resulting in a mean surface-to-surface error ranging 0.50-0.82mm for the heart chambers and 0.60-1.32mm for the parts of the great vessels visible in the images.

Impact of iterative reconstruction on CNR and SNR in dynamic myocardial perfusion imaging in an animal model

AbstractObjectivesTo evaluate a new iterative reconstruction (IR) algorithm for radiation dose, image quality (IQ), signal-to-noise-ratio (SNR), and contrast-to-noise-ratio (CNR) in multidetector computed tomography (MDCT) dynamic myocardial perfusion imaging (MPI).MethodsECG-gated 256-slice MDCT dynamic MPI was performed in six pigs after subtotal balloon occlusion of one artery. Two 100 kVp protocols were compared: high dose (HD): 150 mAs; low dose (LD): 100 mAs. HD images were reconstructed with filtered back projection (FBP), LD images with FBP and different strengths of IR (L1, L4, and L7). IQ (5-point scale), SNR, and CNR (ischemic vs. normal myocardium) values derived from the HD (FBP) images and the different LD images were compared.ResultsMean SNR values for myocardium were 16.3, 11.3, 13.1, 17.1, and 28.9 for the HD, LD (FBP), LD (L1), LD (L4), and LD (L7) reconstructions, respectively. Mean CNR values were 8.9, 6.3, 7.8, 9.3, and 12.8. IQ was scored as 4.6, 3.3, 4.4, 4.7, and 3.4, respectively. A significant loss of IQ was observed for the LD (L7) images compared to the HD (FBP) images (P < 0.05).ConclusionAppropriate levels of iterative reconstruction can improve SNR and CNR, facilitating radiation dose savings in CT-MPI without influencing diagnostic quality.Key Points• Iterative reconstruction (IR) can reduce radiation dose in myocardial perfusion CT. • Our study also demonstrated improvements in image quality (noise, SNR, and CNR). • Dynamic CT-MPI could help determine the hemodynamic significance of coronary artery disease.• With dynamic CT MPI, myocardial blood flow can be determined quantitatively.

Iterative model reconstruction: improved image quality of low-tube-voltage prospective ECG-gated coronary CT angiography images at 256-slice CT.

OBJECTIVES To investigate the effects of a new model-based type of iterative reconstruction (M-IR) technique, the iterative model reconstruction, on image quality of prospectively gated coronary CT angiography (CTA) acquired at low-tube-voltage. METHODS Thirty patients (16 men, 14 women; mean age 52.2 ± 13.2 years) underwent coronary CTA at 100-kVp on a 256-slice CT. Paired image sets were created using 3 types of reconstruction, i.e. filtered back projection (FBP), a hybrid type of iterative reconstruction (H-IR), and M-IR. Quantitative parameters including CT-attenuation, image noise, and contrast-to-noise ratio (CNR) were measured. The visual image quality, i.e. graininess, beam-hardening, vessel sharpness, and overall image quality, was scored on a 5-point scale. Lastly, coronary artery segments were evaluated using a 4-point scale to investigate the assessability of each segment. RESULTS There was no significant difference in coronary arterial CT attenuation among the 3 reconstruction methods. The mean image noise of FBP, H-IR, and M-IR images was 29.3 ± 9.6, 19.3 ± 6.9, and 12.9 ± 3.3 HU, respectively, there were significant differences for all comparison combinations among the 3 methods (p<0.01). The CNR of M-IR was significantly better than of FBP and H-IR images (13.5 ± 5.0 [FBP], 20.9 ± 8.9 [H-IR] and 39.3 ± 13.9 [M-IR]; p<0.01). The visual scores were significantly higher for M-IR than the other images (p<0.01), and 95.3% of the coronary segments imaged with M-IR were of assessable quality compared with 76.7% of FBP- and 86.9% of H-IR images. CONCLUSIONS M-IR can provide significantly improved qualitative and quantitative image quality in prospectively gated coronary CTA using a low-tube-voltage.

Qualitative and Quantitative Assessment of Adenosine Triphosphate Stress Whole-Heart Dynamic Myocardial Perfusion Imaging Using 256-Slice Computed Tomography

Background The aim of this study was to investigate the correlation of the qualitative transmural extent of hypoperfusion areas (HPA) using stress dynamic whole-heart computed tomography perfusion (CTP) imaging by 256-slice CT with CTP-derived myocardial blood flow (MBF) for the estimation of the severity of coronary artery stenosis. Methods and Results Eleven patients underwent adenosine triphosphate (0.16 mg/kg/min, 5 min) stress dynamic CTP by 256-slice CT (coverage: 8 cm, 0.27 s/rotation), and 9 of the 11 patients underwent coronary angiography (CAG). Stress dynamic CTP (whole–heart datasets over 30 consecutive heart beats in systole without spatial and temporal gaps) was acquired with prospective ECG gating (effective radiation dose: 10.4 mSv). The extent of HPAs was visually graded using a 3-point score (normal, subendocardial, transmural). MBF (ml/100g/min) was measured by deconvolution. Differences in MBF (mean ± standard error) according to HPA and CAG results were evaluated. In 27 regions (3 major coronary territories in 9 patients), 11 coronary stenoses (> 50% reduction in diameter) were observed. In 353 myocardial segments, HPA was significantly related to MBF (P < 0.05; normal 295 ± 94; subendocardial 186 ± 67; and transmural 80 ± 53). Coronary territory analysis revealed a significant relationship between coronary stenosis severity and MBF (P < 0.05; non-significant stenosis [< 50%], 284 ± 97; moderate stenosis [50–70%], 184 ± 74; and severe stenosis [> 70%], 119 ± 69). Conclusion The qualitative transmural extent of HPA using stress whole-heart dynamic CTP imaging by 256-slice CT exhibits a good correlation with quantitative CTP-derived MBF and may aid in assessing the hemodynamic significance of coronary artery disease.

Optimizing Radiation Dose Levels in Prospectively Electrocardiogram-Triggered Coronary Computed Tomography Angiography Using Iterative Reconstruction Techniques: A Phantom and Patient Study

Aim To investigate the potential of reducing the radiation dose in prospectively electrocardiogram-triggered coronary computed tomography angiography (CCTA) while maintaining diagnostic image quality using an iterative reconstruction technique (IRT). Methods and Materials Prospectively-gated CCTA were first performed on a phantom using 256-slice multi-detector CT scanner at 120 kVp, with the tube output gradually reduced from 210 mAs (Group A) to 125, 105, 84, and 63 mAs (Group B–E). All scans were reconstructed using filtered back projection (FBP) algorithm and five IRT levels (L2-6), image quality (IQ) assessment was performed. Based on the IQ assessment, Group D(120 kVp, 84 mAs) reconstructed with L5 was found to provide IQ comparable to that of Group A with FBP. In the patient study, 21 patients underwent CCTA using 120 kV, 210 mAs with FBP reconstruction (Group 1) followed by 36 patients scanned with 120 kV, 84 mAs with IRT L5 (Group 2). Subjective and objective IQ and effective radiation dose were compared between two groups. Results In the phantom scans, there were no significant differences in image noise, contrast-to-noise ratio (CNR) and modulation transfer function (MTF) curves between Group A and the 84 mAs, 63 mAs groups (Groups D and E). Group D (120 kV, 84 mAs and L5) provided an optimum balance, producing equivalent image quality to Group A, at the lowest possible radiation dose. In the patient study, there were no significant difference in image noise, signal-to-noise ratio (SNR) and CNR between Group 1 and Group 2 (p = 0.71, 0.31, 0.5, respectively). The effective radiation dose in Group 2 was 1.21±0.14 mSv compared to 3.20±0.58 mSv (Group 1), reflecting dose savings of 62.5% (p<0.05). Conclusion iterative reconstruction technique used in prospectively ECG-triggered 256-slice coronary CTA can provide radiation dose reductions of up to 62.5% with acceptable image quality.

Dynamic CT Perfusion Imaging of the Myocardium: A Technical Note on Improvement of Image Quality

Objective To improve image and diagnostic quality in dynamic CT myocardial perfusion imaging (MPI) by using motion compensation and a spatio-temporal filter. Methods Dynamic CT MPI was performed using a 256-slice multidetector computed tomography scanner (MDCT). Data from two different patients–with and without myocardial perfusion defects–were evaluated to illustrate potential improvements for MPI (institutional review board approved). Three datasets for each patient were generated: (i) original data (ii) motion compensated data and (iii) motion compensated data with spatio-temporal filtering performed. In addition to the visual assessment of the tomographic slices, noise and contrast-to-noise-ratio (CNR) were measured for all data. Perfusion analysis was performed using time-density curves with regions-of-interest (ROI) placed in normal and hypoperfused myocardium. Precision in definition of normal and hypoperfused areas was determined in corresponding coloured perfusion maps. Results The use of motion compensation followed by spatio-temporal filtering resulted in better alignment of the cardiac volumes over time leading to a more consistent perfusion quantification and improved detection of the extend of perfusion defects. Additionally image noise was reduced by 78.5%, with CNR improvements by a factor of 4.7. The average effective radiation dose estimate was 7.1±1.1 mSv. Conclusion The use of motion compensation and spatio-temporal smoothing will result in improved quantification of dynamic CT MPI using a latest generation CT scanner.

Cardiac imaging using multislice computed tomography scanners: technical considerations.

Conventional coronary angiography is currently the gold standard in the detection and diagnosis of coronary artery disease. This modality, however, is invasive in nature. Hence, there is a need for noninvasive imaging techniques to provide comprehensive assessment of coronary artery disease, especially in stable patients at low to moderate risk of disease. In recent years, a number of noninvasive modalities have found wide applications in cardiac imaging. Most recent investigations have used magnetic resonance imaging, multislice computed tomography and electron-beam computed tomography scanners. This review discusses the clinical challenges existing in the field of cardiac imaging and focuses on the technical advancements of multislice computed tomography scanners that have made them a very attractive noninvasive option for the detection and diagnosis of coronary artery disease.

Quantitative myocardial perfusion imaging in a porcine ischemia model using a prototype spectral detector CT system.

We optimized and evaluated dynamic myocardial CT perfusion (CTP) imaging on a prototype spectral detector CT (SDCT) scanner. Simultaneous acquisition of energy sensitive projections on the SDCT system enabled projection-based material decomposition, which typically performs better than image-based decomposition required by some other system designs. In addition to virtual monoenergetic, or keV images, the SDCT provided conventional (kVp) images, allowing us to compare and contrast results. Physical phantom measurements demonstrated linearity of keV images, a requirement for quantitative perfusion. Comparisons of kVp to keV images demonstrated very significant reductions in tell-tale beam hardening (BH) artifacts in both phantom and pig images. In phantom images, consideration of iodine contrast to noise ratio and small residual BH artifacts suggested optimum processing at 70 keV. The processing pipeline for dynamic CTP measurements included 4D image registration, spatio-temporal noise filtering, and model-independent singular value decomposition deconvolution, automatically regularized using the L-curve criterion. In normal pig CTP, 70 keV perfusion estimates were homogeneous throughout the myocardium. At 120 kVp, flow was reduced by more than 20% on the BH-hypo-enhanced myocardium, a range that might falsely indicate actionable ischemia, considering the 0.8 threshold for actionable FFR. With partial occlusion of the left anterior descending (LAD) artery (FFR < 0.8), perfusion defects at 70 keV were correctly identified in the LAD territory. At 120 kVp, BH affected the size and flow in the ischemic area; e.g. with FFR ≈ 0.65, the anterior-to-lateral flow ratio was 0.29 ± 0.01, over-estimating stenosis severity as compared to 0.42 ± 0.01 (p < 0.05) at 70 keV. On the non-ischemic inferior wall (not a LAD territory), the flow ratio was 0.50 ± 0.04 falsely indicating an actionable ischemic condition in a healthy territory. This ratio was 1.00 ± 0.08 at 70 keV. Results suggest that projection-based keV imaging with the SDCT system and proper processing could enable useful myocardial CTP, much improved over conventional CT.