Diagnostic Methods Magnetic Resonance Imaging

To assess the efficacy of magnetic resonance (MR) imaging in evaluating graft patency after coronary bypass surgery, 20 patients who had prior surgery (average 5.5 years, range 1.5 to 14) and recent cardiac catheterization because of chest pain were studied. No patient had surgical intervention or change in symptoms in the time interval between catheterization and MR imaging. These 20 patients had a total of 47 grafts, defined as proximal anastomoses: 20 to the left anterior descending or diagonal artery (LAD), 13 to the left circumflex artery marginal branches (LCX), and 14 to the right coronary artery or posterior descending artery (RCA). The patients underwent cardiac and respiratory gated MR scans in a 0.5 tesla magnet with an echo time of 22 msec and two repetitions in a 128 x 256 matrix. In-plane resolution was 2.7 mm. Every patient had a scan in the transaxial plane and some underwent scanning in the sagittal and coronal planes as well. A graft was considered patent by MR when a signal-free lumen was visualized in an anatomic position consistent with that of a bypass graft, had a lumen larger than the native vessels, was seen on more than one slice, and was seen at a level higher than that of the native vessels. If a known graft was not seen it was considered occluded. The scans were interpreted by consensus of two physicians aware of the operative but not the cardiac catheterization data. Twenty-six of 29 patent grafts and 13 of 18 occluded grafts were correctly classified (sensitivity 90%, specificity 72%). Eighteen of 20 (90%) LAD grafts, 11 of 14 (79%) RCA grafts, and 11 of 13 (85%) LCX grafts were correctly classified. When the results from three patients with technically poor studies because of poor cardiac gating were excluded, the overall sensitivity and specificity were 92% and 85%, respectively. This study demonstrates the high sensitivity and moderate specificity of MR for evaluating the patency of coronary artery bypass grafts, particularly LAD grafts. Circulation 76, No. 4, 786-791, 1987. INCREASING NUMBERS of patients with previous coronary artery bypass surgery develop chest pain due to ischemic or nonischemic causes. In those patients with ischemia this can be due to closure or atherosclerotic narrowing of one or more bypass grafts, or progression of atherosclerosis in native vessels. Noninvasive evaluation of graft patency is thus of obvious potential importance in the management of patients with chest pain after bypass surgery. It may also help to evaluate the need for repeat cardiac catheterization. In 1983 Herfkens et al.' first reported imaging a coronary artery bypass graft on a transaxial magnetic resFrom the Philadelphia Heart Institute, Division of Cardiology, Department of Medicine, and the Department of Radiology, Presbyterian-University of Pennsylvania Medical Center and the University of Pennsylvania School of Medicine, Philadelphia. Address for correspondence: Dr. Alan D. Askenase, Philadelphia Heart Institute, Presbyterian-University of Pennsylvania Medical Center, 51 North 39th St., Philadelphia, PA 19104. Received April 16, 1987; revision accepted July 9, 1987. *Present address: Jersey Shore Medical Center, Neptune, NJ 07753. **Present address: Department of Radiology, University of Colorado School of Medicine, 4200 E. 9th Ave. C-277, Denver, CO 80262. 786 onance (MR) scan of the chest. Since that time other authors have described the visualization of bypass grafts with MR imaging.2 4 These preliminary studies demonstrated that MR imaging of patients with a history of coronary bypass surgery was safe and artifacts from implanted metals did not interfere with the interpretation of results. Bypass grafts have also been imaged by cine computed tomography,5 6 but this technique requires the injection of radiographic contrast dye and the use of ionizing radiation, which is not needed for graft visualization withMR imaging. WithMR imaging vascular structures are well seen with the use of the spin-echo technique. With this type of pulse sequence, rapid flow in the large arteries of the chest appears signal free because excited spins move out of the imaging plane, and slow flowing blood or occluded vessels will not.7' 8 Currently available MR imaging systems have an inplane resolution on the order of 2 mm. We reasoned that MR imaging should be able to routinely detect bypass grafts in cardiac gated scans because of its sensitivity CIRCULATION by gest on A ril 5, 2017 http://ciajournals.org/ D ow nladed from DIAGNOSTIC METHODS-MAGNETIC RESONANCE IMAGING to blood flow in vessels and excellent spatial resolution. The purpose of this study was to validate this hypothesis by studying patients with bypass grafts who had undergone cardiac catheterization to assess the accuracy of MR imaging for the determination of the patency of bypass grafts. Methods and subjects Study population. All patients with a history of coronary bypass surgery who underwent cardiac catheterization after surgery from January 1, 1986, to March 30, 1987, were eligible for entry. A total of 111 consecutive patients underwent cardiac catheterization after coronary artery bypass grafting in this time period. Thirty were excluded because their conditions were unstable and they had undergone repeat coronary artery bypass grafting or had cardiac events such as myocardial infarction, percutaneous transluminal angioplasty (PTCA) of a bypass graft, or death. A further 61 patients either lived too far to return for an MR study, had contraindications for an MR scan (such as a pacemaker or metallic foreign bodies), or did not wish to participate in the study. Twenty patients formed the study population. There were 18 men and two women. The mean age was 59 years (range 43 to 74), and the patients were studied an average of 5.5 years after coronary bypass surgery (range 1.5 to 14). Cardiac catheterizations were performed at the discretion of a patient's private physician, generally to evaluate chest pain after coronary bypass surgery thought to be ischemic. Three patients had PTCA of native vessels not supplied by a patent bypass graft in the period between cardiac catheterization and MR scanning. The average time between catheterization andMR scan was 3.4 months (range 0 to 1 1) and the median was 1 month. Only five patients were scanned more than 2 months after catheterization. None of the patients had any change in clinical symptoms in the time between catheterization and the MR scan. Two patients underwent MR scanning immediately before cardiac catheterization. MR imaging. A superconducting MR imaging system (Picker International Inc.) with a field strength of 5.0 kG (0.5 tesla) was used for all studies. Studies were performed by a spin-echo technique with both respiratory and cardiac gating. Cardiac gating was triggered from the R wave without delay. Respiratory gating was implemented with a bellows wrapped around the upper abdomen. Data collection occurred only at end-expiration. An echo time of 22 msec and two repetitions were used. Pulse repetition times ranged between 500 and 1000 msec, depending on the heart rate (RR interval). The computer matrix consisted of 128 (number of views in the phase encoding direction) x 256 matrix points. Data were interpolated into a 256 x 256 matrix. The field of view was 35 cm, giving an in-plane resolution of 2.7 mm. MR imaging was performed with a multislice technique composed of contiguous 10 mm thick slices, with a minimum of eight slices in the transaxial plane in all patients. All patients underwent imaging in the transaxial plane and most had imaging in the sagittal or coronal planes as well if time permitted. Studies were generally completed in 1 hr. Angiographic studies. Cardiac catheterization was performed in all patients by the Judkins or Sones technique with standard catheters. All bypass grafts or stumps were selectively visualized in at least two projections. The number of bypass grafts, their origin, and the site of insertion were known for all patients from the surgical operative report at the time of cardiac catheterization and MR imaging. Data analysis and statistics. Angiographic interpretation was done by the physician performing the catheterization (J. A.). Bypass grafts were visualized by selective angiography at carVol. 76, No. 4, October 1987 diac catheterization and classified as patent, totally occluded, or patent with a greater than 70% luminal obstruction. If a graft could not be identified with certainty, an aortic root injection was done to ensure that a patent bypass graft was not overlooked. Evaluation was done from the origin of the bypass grafts in the ascending aorta to their first insertion in the native coronary artery. MR studies were interpreted by consensus of a cardiologist (A. A.) and a radiologist (D. T.) unaware of the results ofcardiac catheterization but aware of the number and positions of bypass grafts from prior surgical reports. If a consensus could not be reached by the two physicians, the graft was considered not seen and, therefore, occluded. A consensus was not reached about only one graft. The transaxial studies were first examined from the arch of the aorta down to the base of the heart to identify signal-free lumens in the mediastinal fat arising from the ascending aorta and descending toward the heart. A graft was considered patent by MR if three conditions were met. A signal-free lumen of greater diameter than native vessels was considered a patent graft if it was seen on at least two contiguous images, if it was seen in a slice above the origin of native coronary arteries, and if it was in an anatomic position consistent with a bypass graft known to have been placed in that patient. A graft seen only at its exit from the aorta and not further down in the chest was considered to be a stump and therefore occluded. A graft known to have been placed at the time of surgery but not visualized in any projection was presumed to be occluded. The typical anatomic positions of coronary artery bypass grafts in the transaxial projection are shown in figure 1. Images taken in other planes were used to determine if a graft not seen in the transaxial plane could be identified, in which case it would be considered patent. Only the proximal portions of grafts were included in the statistical analysis since sequential grafts were rarely seen. Data are presented as sensitivity and specificity of MR imaging for classifying bypass grafts as patent or occluded as compared with angiographic classification.