Normand Robert

Binary vascular reconstruction from a limited number of cone beam projections

This paper describes a method to perform reconstruction of vascular cross-sectional images from a limited number of x-ray angiographic cone-beam projections. It is assumed that the projection data can be simplified by identifying blood vessels in each angiogram and removing signals due to other structures. Under these conditions, the x-ray attenuation coefficient, mu, can be modeled as a binary variable having a value mu 0 within the vessel and 0 outside. The reconstruction is performed by minimizing a cost function using the method of simulated annealing. In this paper, we demonstrate that the introduction of a priori information allows one to reconstruct a sphere and a simulated branched vessel from three views with, respectively, 97% and 93% of voxels having correct values. The addition of a continuity constraint for the reconstruction of the branched vessel resulted in further reduction in the percentage of misplaced voxels. Calculations require from one to six hours of CPU time on a Sun SparcStation 2 computer for the cases investigated here. The effect of noise, cooling schedule, and number of views on the reconstruction are examined using simulated vessel projections. Modifications to our approach to accelerate the reconstruction are also discussed.

The geometric calibration of cone–beam systems with arbitrary geometry

A method is proposed to determine the cone-beam x-ray acquisition geometry of an imaging system using a phantom consisting of discrete x-ray opaque markers defining two parallel rings sharing a common axis. The phantom generates an image of two ellipses which are fitted to an ellipse model. A phantom-centric coordinate system is used to simplify the equations describing the ellipse coefficients such that a solution describing the acquisition geometry can be obtained via numerical optimization of only three of the nine unknown variables. We perform simulations to show how errors in the fit of the ellipse coefficients affect estimates of the acquisition geometries. These simulations show that for ellipse projections sampled with 1200 markers, 25 microm errors in marker positions and a source-detector distance (SDD) of 1.6 m, we can measure angles describing detector rotation with a mean error of <0.002 degrees and a standard deviation (SD) of <0.03 degrees. The SDD has a mean error of 0.004 mm and SD = 0.24 mm. The largest error is associated with the determination of the point on the detector closest to the x-ray source (mean error = 0.05 mm, SD = 0.85 mm). A prototype phantom was built and results from x-ray experiments are presented.

Ultrasound‐guided identification of cardiac imaging windows

PURPOSEnCurrently, the use of cine magnetic resonance imaging (MRI) to identify cardiac quiescent periods relative to the electrocardiogram (ECG) signal is insufficient for producing submillimeter-resolution coronary MR angiography (MRA) images. In this work, the authors perform a time series comparison between tissue Doppler echocardiograms of the interventricular septum (IVS) and concurrent biplane x-ray angiograms. Our results indicate very close agreement between the diastasis gating windows identified by both the IVS and x-ray techniques.nnnMETHODSnSeven cath lab patients undergoing diagnostic angiograms were simultaneously scanned during a breath hold by ultrasound and biplane x-ray for six to eight heartbeats. The heart rate of each patient was stable. Dye was injected into either the left or right-coronary vasculature. The IVS was imaged using color tissue Doppler in an apical four-chamber view. Diastasis was estimated on the IVS velocity curve. On the biplane angiograms, proximal, mid, and distal regions were identified on the coronary artery (CA). Frame by frame correlation was used to derive displacement, and then velocity, for each region. The quiescent periods for a CA and its subsegments were estimated based on velocity. Using Pearsons correlation coefficient and Bland-Altman analysis, the authors compared the start and end times of the diastasis windows as estimated from the IVS and CA velocities. The authors also estimated the vessel blur across the diastasis windows of multiple sequential heartbeats of each patient.nnnRESULTSnIn total, 17 heartbeats were analyzed. The range of heart rate observed across patients was 47-79 beats per minute (bpm) with a mean of 57 bpm. Significant correlations (R > 0.99; p < 0.01) were observed between the IVS and x-ray techniques for the identification of the start and end times of diastasis windows. The mean difference in the starting times between IVS and CA quiescent windows was -12.0 ms. The mean difference in end times between IVS and CA quiescent windows was -3.5 ms. In contrast, the correlation between RR interval and both the start and duration of the x-ray gating windows were relatively weaker: R = 0.63 (p = 0.13) and R = 0.86 (p = 0.01). For IVS gating windows, the average estimated vessel blurs during single and multiple heartbeats were 0.5 and 0.66 mm, respectively. For x-ray gating windows, the corresponding values were 0.26 and 0.44 mm, respectively.nnnCONCLUSIONSnIn this study, the authors showed that IVS velocity can be used to identify periods of diastasis for coronary arteries. Despite variability in mid-diastolic rest positions over multiple steady rate heartbeats, vessel blurring of 0.5-1 mm was found to be achievable using the IVS gating technique. The authors envision this leading to a new cardiac gating system that, compared with conventional ECG gating, provides better resolution and shorter scan times for coronary MRA.

A filtering method for signal equalization in region-of-interest fluoroscopy.

A method to significantly reduce the exposure area product in fluoroscopy using a pre-patient region-of-interest (ROI) attenuator is presented. The attenuator has a thin central region and a gradually increasing thickness away from the center. It is shown that the unwanted brightening artifact caused by the attenuator can be eliminated by attenuating the low spatial frequencies in the detected image using digital image processing techniques. An investigation of the best image processing method to correct for the presence of the attenuator is undertaken. The correction procedure selected is suitable for use with real-time image processors and the ROI attenuator can be permitted to move during image acquisition. Images of an anthropomorphic chest phantom acquired in the presence of the ROI attenuator using an x-ray image intensifier/video chain are corrected to illustrate the clinical feasibility of our approach.

A novel lead attenuator to reduce operator exposure to scattered radiation in transradial coronary procedures.

Introduction The harmful effects of radiation were soon recognised after the discovery of x-rays by Wilhelm Roentgen in 18951. Increased incidences of skin cancer and leukemia confirmed the carcinogenic potential of x-rays in the early twentieth century2. Interventional cardiologists experience frequent radiation exposure through fluoroscopy. Interventional cardiology procedures performed via the radial approach are associated with longer fluoroscopy times and greater cumulative scatter radiation to the operator and staff 3. This approach is gaining popularity and exceeds 90% of cases in dedicated centres4. The cumulative risk associated with a lifetime of exposure could become significant5,6. The lower torso (from the umbilicus and down) acts as a source of scatter radiation to the operator and is not routinely shielded. We developed a uniquely designed, non-disposable lead attenuating material that could shield this region and significantly reduce the radiation scatter exposure to an operator in the catheterisation laboratory.

Characterisation of a novel porcine coronary artery CTO model.

AIMSnTo create a large animal coronary chronic total occlusion (CTO) model. Presence of microvessels within the CTO lumen facilitates guidewire crossing. The patterns and time profiles of matrix changes and microvessel formation during coronary CTO maturation are unknown.nnnMETHODS AND RESULTSnCTO were created in 15 swine by percutaneous deployment of a collagen plug. Matrix changes were assessed by histology. Intraluminal neovascularisation was assessed by histology and several imaging modalities, including conventional and 3D spin angiography, micro-computed tomography (micro-CT) imaging, and contrast-enhanced magnetic resonance imaging (MRI), at six and 12 weeks following CTO creation. Matrix changes included an intense inflammatory reaction at six weeks which had partially abated by 12 weeks. A proteoglycan-rich matrix at six weeks was partially replaced with collagen by 12 weeks. Similar changes were noted in the proximal cap which was acellular. Three patterns of microvessel formation were identified and defined based on the presence and extent of a lead neovessel. No major differences in pattern or extent of neovascularisation were noted between six and 12 weeks.nnnCONCLUSIONSnHeterogeneity in neovascularisation patterns occurs during coronary CTO development in a porcine model. Non-invasive imaging to determine the predominant type of neovascularisation prior to and during CTO revascularisation may improve guidewire crossing success rates. This model may be useful for further exploration of CTO pathophysiology, and may aid in further refinements of in vivo imaging of CTO and development of novel therapeutic approaches to revascularisation of CTO, such as manipulations of the proximal cap, matrix composition, neovessel induction, and device testing.

Improvement of LV functional performance in the chronic total coronary occlusion during the late stage is associated with the extensive collateral development

Methods In nine pigs a CTO was created by percutaneously inserting a fibrin plug (AngiosealΤΜ) into the mid-todistal left anterior descending artery (LAD). Animals were studied six (n=3) or twelve weeks (n=6) later prior to sacrifice. An x-ray angiogram confirmed the LAD CTO at those time points. Cardiac MR (CMR) studies were then conducted on a 1.5T (n=6) or on a 3.0T MRI system (n=3), which included SSFP short axis sections for wall motion and post-gadolinium LGE-MRIs for viability. After sacrifice, both right and left coronary systems were injected with Microfil. X-ray or MSCT angiography of the fixed heart was obtained. A longitudinal cardiac section including the CTO lesion, proximal/distal LAD and the borders of infarction was removed and prepared in gel, then imaged in a microCT system at 27 micron resolution. LV functional parameters including wall thickness at end-systole (WTES) and end-diastole (WTED) were measured in border zone, infarct and remote region. Systolic wall thickening (SWT) was calculated using as (WTES-WTED) x100/ WTED. CMR and micro-CT data were processed using commercial software. The extent of collaterals on micro-CT images was rated qualitatively using a score from 0 to 3, indicating that no, minimal, moderate, or extensive collaterals were observed. A Student’s t-test was used for the statistical significance of differences between measurements at 6 and 12 weeks.

MRI-Compatible C-Arm Imaging for Cardiac Intervention

Minimally invasive interventions limit visual access to the anatomy under treatment requiring the use of imaging technologies for guidance. Organs and vessels located deep within the body can be visualized with imaging modalities having good tissue penetration such as X-ray or magnetic resonance imaging (MRI). X-ray guidance via fluoroscopy provides real-time images of large anatomical territories with a spatial resolution of ~0.2 mm usually with the aid of contrast agents. X-ray guidance also provides excellent percutaneous device visualization. MRI provides superior soft tissue differentiation and three-dimensional (3D) localization. However, intravascular MRI guidance is still in its clinical infancy, and concerns remain over exclusive reliance on this modality (Bock M and Wacker FK, J Magn Reson Imaging 27:326–38, 2008; Hushek SG et al., J Magn Reson Imaging 27:253–66; Ratnayaka K et al., J Cardiovasc Magn Reson 10:62, 2008). MRI and X-ray catheterization imaging exhibit complementary strengths that may potentially improve percutaneous therapies. Efforts to combine these two modalities into a fully hybrid X-ray-MR (XMR) system were first proposed by Fahrig et al. (J Magn Reson Imaging 13:294–300, 2001; Acta Neurochir 145:995–7, 2003). Current approaches include (1) the use of conventional X-ray catheterization and MRI systems in adjacent rooms with the addition of a dual-modality compatible patient table and transport system, (2) “combined” or “hybrid” systems, whereby both modalities can image overlapping volumes of interest with no or minimal patient relocation. Examples of interventions that may benefit from XMR are presented, as well as the imaging requirements associated with these procedures. We describe a novel hybrid cone-beam XMR system built by introducing a rotating anode X-ray catheterization system within 140 cm of a closed-bore 1.5 T MRI. The system is used to acquire images of an MRI-compatible catheter moving inside an aorta phantom.