Validation of first pass magnetic resonance myocardial perfusion imaging using fractional flow reserve

Watkins, Stuart (2009) Validation of first pass magnetic resonance myocardial perfusion imaging using fractional flow reserve. MD thesis, University of Glasgow.

Abstract

Background - Magnetic Resonance Myocardial Perfusion Imaging (MRMPI) has been used for the detection of reversible myocardial ischaemia in humans since the early 1990’s. This non-invasive method of diagnosing reversible myocardial ischaemia has a number of advantages over the other more commonly used non-invasive tests such as ETT, stress echocardiography and radionuclide single photon emission computerised tomography (SPECT). There is no need to perform physical exercise, no image orientation constraints, excellent spatial and temporal resolution, no photon scatter or attenuation artefacts and no exposure to ionising radiation. The use of MRMPI for the detection of reversible myocardial ischaemia has been extensively investigated in the past using other non-invasive tests as the gold standard namely PET and SPECT. Invasive comparisons have been made with visual coronary angiography and quantitative coronary angiography (QCA). This previous work has been summarised in a meta-analysis which estimated the sensitivity and specificity to be 84% and 85% respectively. The majority of previous studies have used QCA or visual estimation of stenosis severity to determine the diameter of stenosis (DS). This however has been shown to correlate poorly with the functional significance of disease within a coronary artery. Prior to the commencement of this study no comparison had been made with the invasive gold standard of FFR. This is measured using a coronary pressure wire at the time of coronary angiography and is regarded by many cardiologists to be the current invasive gold standard for determining if coronary artery disease (CAD) is physiologically significant. We therefore undertook the present study to determine the true accuracy of MRMPI for the diagnosis of physiologically significant CAD. We also assessed the ability of MRMPI to detect isolated microcirculatory disease as determined by thermodilution derived CFR. Our other aims included an analysis of troponin release following PCI and its relation to QCA, pressure wire data and the occurrence of new late gadolinium enhancement (LGE). New LGE post CABG was also quantified and compared with that encountered post-PCI.

Methods - One hundred and three patients with chest pain were referred for coronary angiography and underwent MRMPI in the week prior to the angiogram. This was performed on a Siemens Sonata 1.5Tesla scanner (Erlangen, Germany). Scanning commenced with localisers and cine long and short axis scans (TrueFISP sequence) to provide left ventricular mass, volume and ejection fraction data. This was followed by perfusion imaging of 3 short axis slices using a turboFLASH sequence (TI 90ms, TE 0.99ms, TR 173ms, Flip Angle 8 degrees, Matrix 80 x 128). Thereafter long and short axis slices were acquired for the detection of LGE (turboFLASH). Maximal hyperaemia was achieved using intravenous adenosine (140µg/kg/min). The first pass bolus contained 0.1mmol/kg of gadolinium (Omniscan, Amersham Health, Oslo, Norway) power injected at 5ml/sec (Medrad, Pittsburgh, PA) followed by a 20ml saline bolus. Twenty minutes after the initial bolus of gadolinium a further bolus was administered to obtain rest perfusion images. During coronary angiography the FFR was recorded in all patent major epicardial coronary arteries using a coronary pressure wire (RADI Medical Systems Ltd, Uppsala, Sweden) with hyperaemia induced using intravenous adenosine as above. An FFR value of <0.75 was taken as the cut off for the diagnosis of significant CAD. CFR measurements were obtained at rest and during maximal hyperaemia by means of thermodilution using 3ml boluses of saline. Following coronary angiography those patients who underwent PCI returned for a repeat MRMPI scan at 24 hours and 4 weeks and CABG patients returned for a 4 week scan. PCI patients had a troponin I measurement performed at approximately 24 hours, just prior to their repeat MRMPI. Qualitative MRMPI analysis, left ventricular mass, volume and ejection fraction analysis and QCA were all performed by two blinded independent experienced observers.

Results - Of the 103 enrolled patients, two were excluded from the final analysis. Seventy-six (74%) were male with a mean age of 60 years (SD = 9). 25 (24.8%) of 101 scans were normal, 40 (39.6%) had single-vessel disease, 26 (25.7%) had two-vessel disease and 10 (9.9%) had triple-vessel disease. 121 perfusion defects were reported in 300 coronary territories (3 patients had complete data for only 2 coronary territories) of which 110 had an FFR<0.75. 168 of 179 normally perfused territories were confirmed normal with an FFR greater than or equal to 0.75. The sensitivity, specificity, PPV and NPV of MRMPI for the detection of significant CAD were 91, 94, 91 and 94%. Cohen’s kappa was 0.97 per patient and 0.76 per coronary territory indicating excellent and substantial agreement between observers. Twenty-eight coronary arteries were identified as having an FFR>0.8 and a CFR<2.0 indicative of isolated microcirculatory disease with no physiologically significant epicardial disease. No coronary territories were found to have a perfusion defect on MRMPI suggesting that by visual analysis MRMPI is unable to detect isolated microvascular disease. The median post PCI troponin level was 0.57µg/L (SD=2, Range undetected - 13.1). The only parameters found to correlate with troponin I levels post-PCI were increasing lesion length (r=0.6, p<0.0001) and increasing total stent length (r=0.37, p=0.02). We compared the increase in mass of LGE between the post-PCI scans and the pre-PCI scan and compared this with the troponin measurement. No significant correlation was found to exist between these parameters at 24 hours (r=0.25, p=0.07) or at 4 weeks (r=-0.19, p=0.2). The change in mass of LGE was calculated for the PCI and CABG patients. The mean difference in the PCI group was -0.12g (Median=0, SD=0.8, Interquartile Range 0 – 0) and for the CABG group was 1.08g (Median=0.11, SD=2.3, Interquartile Range -0.11 – 1.38). There is a trend towards the development of more LGE following CABG than PCI however the difference between groups did not reach statistical significance (p=0.07).

Conclusion - MRMPI can accurately detect significant CHD with excellent results using FFR as the gold standard. Interobserver agreement is also very good even when examining individual coronary artery territories. Qualitative analysis of MRMPI is unable to detect isolated microcirculatory disease as defined by an FFR>0.8 and a CFR<2.0. Small troponin releases are common post-PCI and are related to the length of the lesion being treated and the length of stent deployed to treat the lesion. These small troponin releases do not accurately correlate with the occurrence of new LGE. CABG did result in a trend towards more new LGE compared to PCI.

• Item Type: Thesis (MD)
• Qualification Level: Doctoral
• Keywords: myocardial ischaemia, myocardial perfusion, coronary artery disease, magnetic resonance myocardial perfusion imaging, late gadolinium enhancement, troponin, adenosine, caffeine, fractional flow reserve, coronary flow reserve, index of microcirculatory resistance, percutaneous coronary intervention, coronary artery bypass grafting, gadolinium
• Subjects: R Medicine > RZ Other systems of medicine; R Medicine > RC Internal medicine
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cardiovascular & Metabolic Health
• Supervisor's Name: Oldroyd, Dr. Keith and Dargie, Prof. Henry
• Date of Award: 2009
• Depositing User: Dr Stuart Watkins
• Unique ID: glathesis:2009-799
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 23 Jun 2009
• Last Modified: 10 Dec 2012 13:26
• URI: https://theses.gla.ac.uk/id/eprint/799