Study of blood flow parameters in a phantom by magnetic resonance imaging MRI

The study of pulsatile flow through a stenosis is motivated by the need to obtain a better understanding of the impact of flow phenomena on atherosclerosis and stroke. MRI techniques have been employed to characterize flow emerging from a stenosis and non-stenotic tube. Detection and quantificatio...

Πλήρης περιγραφή

Λεπτομέρειες βιβλιογραφικής εγγραφής
Κύριος συγγραφέας: Καζέρου, Ασπασία
Άλλοι συγγραφείς: Ευσταθόπουλος, Ευστάθιος
Μορφή: Thesis
Γλώσσα:English
Έκδοση: 2013
Θέματα:
Διαθέσιμο Online:http://hdl.handle.net/10889/5911
Περιγραφή
Περίληψη:The study of pulsatile flow through a stenosis is motivated by the need to obtain a better understanding of the impact of flow phenomena on atherosclerosis and stroke. MRI techniques have been employed to characterize flow emerging from a stenosis and non-stenotic tube. Detection and quantification of stenosis, serve as the basis for surgical intervention. In the future, the study of arterial blood flow will lead to the prediction of individual hemodynamic flows in any patient, the development of diagnostic tools to quantify disease, and the design of devices that mimic or alter blood flow. Blood flow and pressure are unsteady. The cyclic nature of the heart pump creates pulsatile conditions in all arteries. The heart ejects and fills with blood in alternating cycles called systole and diastole. Blood is pumped out of the heart during systole. The heart rests during diastole, and no blood is ejected. Pressure and flow have characteristic pulsatile shapes that vary in different parts of the arterial system. The experiments demonstrate that stenotic pulsatile flow exhibit flow disturbance phenomena which deviate the flow from the laminar behavior. In vitro measurements can simulate blood flow to a satisfactory degree, under various assumptions for flow. In this study, estimation of various hemodynamic parameters, are achieved by means of a flow phantom. The phantom can simulate pulsatile blood flow in arterial system, in our case blood flow in carotid artery. The phantom consists of an one-headed positive displacement diaphragm pump, driven by an electrocardiogram (ECG) generator, with the tube, creating a closed circuit. Within the circuit, water (as blood mimicking fluid) is driven, simulating blood flow. We studied the flow using velocity-encoded MR phase contrast sequences. Phase contrast angiography relies on dephasing the moving spins submitted to a bipolar gradient. For a bipolar gradient of a given intensity and time, the moving spins will dephase in proportion to their velocity. Similar to spatial encoding in the phase direction, the possible phase values range from – π to + π. Beyond this range of values, aliasing occurs, causing poor velocity encoding. The encoding gradient characteristics are thus defined in order to encode flows within a certain velocity range from -Venc to +Venc to be determined by the user. Any velocity outside this range will be poorly encoded (similar to what happens in pulsed and color Doppler with PRF). The present work refers to blood flow estimation by means of Magnetic Resonance Imaging. The MR imaging system used, is a 1.5 Tesla scanner (Intera 1.5T, Philips Medical Systems, Best, the Netherlands) of Attikon Hospital (Second Department of Radiology). CT imaging system is a Philips Brilliance 64, used to assess the percentage of the stenosis. The experimental set up consists of a flow phantom, simulating blood flow through blood vessels under chosen conditions. Gradient echo (phase contrast) sequences used, precisely: SQ flow and QFP sequences. MR phasecontrast technique quantifies and displays flow velocities in real times. The sequence uses a two-dimensional selective radiofrequency pulse followed by flow-sensitizing gradients with an echo planar readout. It provides the simultaneous display in real time of both an anatomic image for positioning and the through plane flow-velocity data. By controlling scan position and orientation interactively, one can optimize flow signal. The retrospective search of measurements is carried out with the database of a software used, called EVORAD. The software of the workstation automatically provided the following parameters: ROI area (cm2), vessel lumen diameter (cm), blood volume flow (ml/s), mean and maximum blood flow velocities (cm/s). The LOIs in respective, used for velocity profiles determination, were acquired by ImageJ software, by similar procedure at vertical and horizontal direction on the lumens’ plane perpendicular to the flow. Two different geometries were used: a PVC tube mimicking a healthy carotid artery of 6mm internal diameter and a stenotic glass tube to simulate arterial pathology, of 8mm internal diameter. Considering the non-stenotic PVC tube, VFR values are estimated volumetrically (for various bpm and pump output values) and via MRI (for straight and inclined position). VFR values are then compared. MR maximum velocity values are estimated too, and velocity profiles are plotted. The procedure is similar in the case of the stenotic glass tube, for various (bpm and PO; pump output) and at intervals of 1cm across the stenosis reaching 4cm upstream and downstream. In the sequence of estimation, percentage of stenosis follows; estimated from both MRI and CT scans. Finally, variation of pressure and SNR in order to assess the signal loss due to stenosis are estimated. Accounting for the non stenotic tube: The first significant issue to mention, is the greatest cv (correlation of variation) values at lower VFR values (measured at 10%pump output and pressure of 2,5b), among all VFR values for both 60 and 75 bpm, and the greatest std values noticed at the greatest VFRs (60%, 4.6b). VFR values are indeed greater at 75 bpm compared to those at 60 bpm, as expected. Values show no stable relevance between VFR and pump output. There are differences in VFR values from the inclined position, statistically significant, in cases of 5% for both 60,75bpm. Statistical differences (at 5% statistical significance), are noticed between volumetric measurements versus MRI extracted values as compared above, between SQflow and QFP sequences (60/20). VFR values comparison between volumetric and MRI measurements, show statistical differences. Concerning Vmax values from ROIs and LOIs V,H: there are statistical differences in 5, 10%PO, for both 60, 75bpm, indicating higher values in straight position. Concerning Vmax values extracted from MRI ROIs and LOIs V&H: there are statistically significant differences in cases of 5,10% PO, at both 60,75bpm, leading to greater values at straight position. In the case of stenotic tube: Comparison, of VFR values at 75, 120 bpm, result in higher flow at the exit of the stenosis (49.16%, 80.14%). In the vicinity of stenosis (± 1cm), VFR is almost stable in the case of 120 bpm (0,74%), whereas the highest variation is noted at 75 bpm (133,9%). The highest VFR value intrastenotic is noted at 120 bpm (2,06ml/s). As flow increases, VFR variation is noted more distal to the stenosis. Percentage comparison indicate that greater variations for 60,75, 120bpm are noted in the vicinity of stenosis (±1cm), whereas for 100bpm at ±3cm. Considering Vmax extracted values at 4cm post stenosis in all cases of pulsatility are higher than the respective values 4cm pre stenosis. At the neck of the stenosis extracted values are indeed high as expected, since laminar flow persists across the stenosis. The highest Vmax value among all intra-stenotic values, appears at 60 bpm. As pump flow rate increases, maximum value occurs most post along stenosis. Post stenosis variations are expected to be higher at higher pulsatility. Vertical LOIs result in higher R squared values. In lower flow (corresponding to lower pulsatility 60,75bpm as mentioned above), parabolic profiles as noted pre and post stenosis (2cm,1cm pre and 4cm post). For higher flow, (100,120bpm), parabolic profiles are depicted post stenosis (2-4cm) and in the neck of stenosis for 100bpm. Severity of stenosis is calculated as the percentage rate of Vmax upstream or downstream the stenosis to the intrastenotic Vmax, minus the unity. The pump output is set up to 10%, flow rate ranges from 60 to 120 bpm. Calculations account for Vmax values from both ROIs and LOIs (V,H). Measurements from CT scan are also acquired (gold standard) for comparison. Due to turbulence, Xpre values are considered as more reliable. Better agreement for stenosis estimations to ROIs are acquired: in low flow from LOIsV, whereas at higher flow by LOIsH. Overall, values extracted by MR at 60 bpm imply a stenosis of 46% (LOIsH), 98% (LOIsV) and 93% (ROIs), whereas CT scans estimations lead to 90.2% using diameter stenosis and 99% using area stenosis. The LOIsH expectedly underestimate the percentage of stenosis. CT value of 99% is the exact value, that result by the relationship described in Ota et al.(2005) study: A=D*[2-(D/100)], where D=0,902 is the “diameter stenosis”. ΔP values at 60 and 100bpm, exceed the respective at 75, 120 bpm. In Vmax values, higher intrastenotic values were noted at those pulsatilities, indicating higher pressure energy loss converted to kinetic energy. Calculations of ΔP, a value of 4 is used for K factor and Vmax values are calculated in m/s in the neck of the stenosis. Calculations from linear and elliptical ROIs were made. By the same reasoning as before, we assume that the value of 15.92mmHg found at 60bpm from ROIs is the most reliable. A second calculation of ΔP by means of K=4.9 lead to higher values of 5.1%. Signal to noise ratio as indicative of the loss of signal as fluid flows along the stenosis. Rectangular ROIs are designed upstream and downstream the stenosis, thus SNR values: upstream the stenosis, are higher in contrast to all respective values downstream. Calculations lead to values of: 54.15% (60bpm), 71.08% (75bpm), 68.7% (100bpm) and 72.63% for 120 bpm. The highest loss is depicted at 120 bpm, and in descending order at 75, 100 and 60 bpm. There are certain factors that are limiting when it comes to comparing the executed study to clinical flow measurements, many of which are connected to properties of the pump and phantom used. At very low pump output as used, there was instability at several times. On the other hand at high PO the pressure reached maximum value (manometer) and was thus avoided. The PO values of 5, 10, 20% are quite lower than that usually found in patients. Thus, a direct comparison to in vivo values would be invalid. The tube in the phantom differs from that of a blood vessel as it is rigid, tube wall consists of PVC or glass, and BMF has different relaxation properties than those found in vivo. Furthermore, the size of the phantom used is much smaller than that of an actual patient, which can lead to a significant divergence in susceptibility variations in scanned material. Consequently, optimal future projects should include scanning faster flow, higher PO, higher magnitude of 3T, different sequences and modalities (various stenoses, oblique positions, blood mimicking fluids, different vessel walls; to more closely mimic in vivo conditions and to reduce the influence of partial volume effects) and a comparison among different techniques as ultrasound, computed tomography CT. Turbulence in flow is crucial for comprehension and interpretation of the flow across a stenosis. Hence, complete understanding of the interrelationship between pressure, flow, and symptoms for cardiovascular stenoses is a critical problem. New devices to repair stenotic arteries are continuously being developed. Thus fluid mechanics will continue to play an important role in the future diagnosis, understanding, and treatment of cardiovascular diseases.