DEVICE AND METHOD FOR DRIVING A KNOWN PULSED FLOW RATE WITHIN AN MRI, TO CARRY OUT ITS PERFORMANCE EVALUATION IN THE FIELD OF HEMODYNAMIC MEASUREMENTS

Abstract

The invention relates to a method for characterizing the degree of confidence in MRI flow measurements by imaging a phantom, characterized in that: a) a reference pair formed by a digital model and a physical model is implemented, b) a modeling of the MRI flow data of the digital phantom subjected to a reference fluid supply sequence is carried out to record a first reference digital data file, and record the reference fluid supply parameters.This is considered as a digital standard, c) an acquisition of the MRI flow data is carried out by the MRI equipment controlled by the sequence to be characterized, of the physical phantom supplied with said reference fluid supply sequence, to record a second file of experimentally measured digital data, the supply being controlled by said reference fluid supply parameters, d) After having carried out at least the steps of recalibration, correction and formatting of the data, steps necessary for spatio-temporal compatibility, a comparison of said digital data files is carried out to calculate at least one error indicator [speed, wall shear stress, flow rate, vorticity, entropy, etc.] representative of the differences between said digital files.

Description

A flow meter (7) is placed in series to control the actual instantaneous flow supplying the phantom (1). The phantom has openings to which pressure sensors are connected. Examples are the curved part (11) with sensors (23 to 25), but there are also the lateral branches (12, 13) and the cavity (15), which also has a local pressure measuring point (26). The shape is chosen to generate a complex and realistic flow, such as that observed in the large circulation (aorta-heart) of the cardiovascular system. The inner diameter of the phantom (1) is 26 mm and was designed with a 50 mm radius of curvature to mimic the blood flows of the aortic arch. A collateral branch (14) was established by analogy with the collateral arteries. The size of this collateral branch was designed to reproduce the splitting observed in vivo between the aorta and the arteries of the supra-aortic trunk, celiac, renal and iliac arteries. Finally, the protuberance (15) attached to the intersection between the collateral branch (14) and the main branch (13) simulates the blood flows in aortic aneurysms, but also allows to appreciate the vortices present in the cardiac cavities. The shape of the phantom is first created digitally using computer-aided design software. The CAD file then forms the digital basis for preparing both a digital hemodynamic phantom and the physical phantom using a 3D printing technique based on stereolithography. Stereolithography allows the production of complex physical models with excellent surface conditions and a geometric tolerance of the order of 25|im at the date of publication of this document. This fidelity between the geometry of the CAD model and the real physical model makes it possible to guarantee the digital / physical couple in terms of geometric equivalence. Other models obtained by segmentation on medical images are dependent on the spatial resolution of the images and the voxelized nature of the data. This is the reason why it is essential to start from the CAD model towards the physical model in order to preserve geometric equivalence. The physical phantom is placed in the MRI equipment, in place of the patient, after being placed in a bag filled with a gel. It is powered by a circuit comprising a pump (6) and a flow meter (7) as well as a buffer tank (-8). The pump (6) is computer-controlled to provide waveforms that are reproducible over time and conform to the digital model defined by in the reference flow. The fluid has hydrodynamic characteristics similar to those of blood in the general circulation: viscosity = 4 cPoi and density = 1020 kg / m3. The pump delivers a pulsed flow which is controlled at all times by the flow meter (7). The pressure sensors allow validation of the hydrostatic pressure maps reconstructed from the speed maps. To reduce the swirling motion of the incoming fluid, a "honeycomb" (10) makes the flow laminar at the entrance to the region of interest. Functional diagram Figure 4 illustrates the functional diagram of the invention. A numerical model (50) is prepared, consisting of a geometry (1) and flow conditions. A physical phantom (51) is manufactured by 3D printing from the CAD geometry of the digital model. From the digital model, a CFD map is generated by modeling the hemodynamics in accordance with the digital model (52). This map serves as a reference (standard). The physical model is placed in the MRI, the system is fed with the flow conditions set by the digital model and the 4D flow acquisition sequence(s) (53) which control the acquisition are launched. Fed with the parameters conforming to the digital model, the phantom and its power supply system reproduce the reference hemodynamics identically and the MRI generates an MRI map (54) which is the experimental measurement of the phenomenon. The data from the MRI sequence to be validated are also recorded. The data from the MRI and CFD maps are formatted for comparison. By correlating the two data volumes, the deviations (55) between the standard and the measurement are calculated. These treatments make it possible to establish the deviations (55) on useful markers of the specialty. This information then allows the practitioner to assess the degree of confidence he can place in acquiring an image on a patient with the same MRI sequence. Experimental results As an example, the phantom was placed in an MRI machine controlled with a 4D Flow sequence optimized with the following settings: VENC encoding speed = 0.5 m / s in all three speed encoding directions; Echo time TE = 3.43-3.5 ms; Time resolution = 49-52 ms; Tilt angle = 15°. Velocity measurements are corrected for partial volume effects caused by the phantom walls and random noise on the phase images due to the air surrounding the phantom. The blood-mimicking fluid was modeled as an incompressible Newtonian fluid with kinematic viscosity v = 4.02 X 10-6 m2 / s. (4cPoi) The Yales2bio solver was used to solve the fluid mechanics equations. A flow inlet boundary condition is imposed at the inlet, it represents the reference flow rate. A convective type outlet boundary condition has been imposed such that it guarantees mass conservation for the entire flow domain. All walls were assumed to be rigid, in accordance with the material chosen for the fabrication of the phantom, and a no-slip condition was prescribed (velocity=0 at the wall). To ensure that the correct (2D) flow profile is imposed at the entrance to the digital domain during CFD calculation, the relative velocity distribution was adjusted based on the image of a 2D phase contrast acquisition, measured at the entrance of the phantom and perpendicular to the flow. This acquisition is independent of the 4D flow acquisition and allows to correct for shifts in the waveform created by The impedance contained by the series of pipes between the pump and the phantom. The flow meter guarantees the absolute value of the flow rate and the 2D series allows it to be given the correct distribution at the input of the digital domain. This control loop ensures perfect equivalence between the input profiles used in the digital domain and those present in the physical domain. Time average of the numerical solution to construct the reference field During a 4D Flow MRI exploration, the MRI signal is constructed over several hundred cardiac cycles. For example, an acquisition centered on the heart can last 15 minutes and an acquisition at the brain level can last 45 minutes of acquisition. During this time, the MRI signal makes it possible to reconstruct a single cycle representative of the hemodynamics characteristic of the region studied. For its part, fluid mechanics modeling will propose instantaneous solutions that are not exactly identical because they will take into account the fluctuations and instabilities present in complex flows (eddies). In order to reconstruct a numerical solution equivalent to that measured by MRI, all simulations were carried out for 40 cycles, on average over the last 30 cycles after removing the first 10 cycles to cancel the effect of the arbitrarily selected initial condition. The velocity maps modeled by CFD (left) and measured by MRI (right) are presented in Figure 4. Sampling CFD velocity fields Two classical methods can be adopted to compare CFD and MRI velocity fields with each other: voxelized MRI velocity fields can be interpolated onto the high-resolution CFD tetrahedral mesh (called HR-CFD), or HR-CFD velocities, located on the nodes of the tetrahedral mesh, can be taken into account in the voxel of the MRI images, to obtain a voxelized field with the undersampled CFD velocity values, called low-resolution CFD (LR-CFD). This latter method is suitable for studying errors in flow measurement. It is weighted with respect to the volumes of the tetrahedra included in the reconstructed voxel. In the given example, the velocity fields were qualitatively compared with MRI measurements of 2 mm isotropic voxel size. The velocity maps reveal excellent agreement and show very similar flows even in case of complex patterns, as is the case in aneurysm. A statistical analysis of the velocity differences across the entire segmented domain between LR-CFD and 4D Flow MRI (2 mm voxel size) was performed, for each velocity component. For this analysis, the CFD solution was downsampled to the MRI domain, phase-averaged, and the velocities of the voxels overlapping the wall were set to zero. All correlations were found to be statistically significant (p < 10-7). Excellent correlation was found for the two main velocity components (head-foot and right-left). The third component was less well represented by the lack of contrast resolution fixed by the imaging (the maximum velocity in this direction was too low compared to the encoding rate chosen for this measurement). The example presented concept and validation of the speed components are optimized and close to 96% 5 speed maps on this + MRI sequence). here was used as proof of method, the correlations of the excellent on this sequence of confidence for the whole of the imaging system (MRI device

Claims

Claims 1 - Process for characterizing the degree of confidence in MRI flow measurements by imaging a phantom characterized in that: 5 a) we implement a reference pair formed by a digital model and a physical model, b) modeling of the MRI flow data of the digital phantom subjected to a reference fluid supply sequence is carried out to record a first file of 10 reference digital data, and record the reference fluid supply parameters c) an acquisition of the MRI flow data is carried out by the MRI equipment controlled by the sequence to be characterized, of the physical phantom supplied with said reference fluid supply sequence, to record a second file of experimentally measured digital data, the supply being controlled by said fluid supply parameters of £ ete £ ence, d) After having carried out at least the steps of 20 recalibration, correction and formatting of the data, steps necessary for spatio-temporal compatibility, a comparison of said digital data files is carried out to calculate at least one error indicator representative of the differences between said digital files. 2 5 2 - Process for characterizing the degree of confidence of MRI flow data according to claim 1 characterized in that said physical model comprises at least one sensor of a fluid supply parameter. 3 - Process for characterizing the degree of confidence of MRI flow data according to claim 2 characterized in that said physical model comprises at least one pressure sensor. Characterization process read degree confidence of MRI flow data according to claim 2 characterized in that said physical model comprises at least one flow sensor. 5 - Method for characterizing the degree of confidence of the MRI flow data according to claim 1 characterized in that said model is placed in a watery medium.

Images