Diagnostic Methods Congenital Heart Disease

Eleven patients with a total of 17 palliative systemic-pulmonary artery shunts underwent evaluation by electrocardiogram-gated magnetic resonance imaging (GMRI). GMRI successfully imaged 1 1 of 17 shunts (65%), including five of nine Blalock-Taussig shunts, four of six Glenn shunts, and both aortopulmonary shunts. All shunts except for the Waterston were imaged on coronal sections during end-systole. The single Waterston shunt was seen on sagittal and transverse scans. Shunt localization and identification were facilitated by obtaining multiple, contiguous sections through the body. Glenn shunts could be imaged entirely in one section, although multiple sections were required to locate the correct plane. Blalock-Taussig shunts generally required multiple sections to image different segments of the shunt. Both aortopulmonary shunts were seen as direct side-to-side connections of the aorta and pulmonary artery. GMRI permitted assessment of the size, course, patency, and distribution of systemic-pulmonary artery shunts as well as the size and morphology of the proximal pulmonary arteries. We conclude that GMRI is a useful, noninvasive method for imaging the anatomy of systemic-pulmonary artery shunts. Circulation 70, No. 4, 650456, 1984. CONGENITAL heart defects characterized by cyanosis and decreased pulmonary blood flow are often palliated by creation of systemic-pulmonary artery shunts. All of the operations that have been devised increase pulmonary blood flow and differ primarily in terms of which specific vessels are anastomosed. The evaluation of patients who have previously undergone palliative shunt procedures frequently includes imaging of the shunt and pulmonary arterial anatomy. This generally requires invasive angiographic studies, either during cardiac catheterization or, more recently, with digital subtraction techniques. Noninvasive imaging of systemic-pulmonary artery shunts has been unsatisfactory. Ultrasonography is limited by the poor echographic accessibility of these structures, which are extracardiac and enveloped by lung tissue. We have previously reported our experience with electrocardiogram-gated magnetic resonance imaging (GMRI) in patients with congenital heart defects.1 2 Included in our early work were two patients in whom palliative systemic-pulmonary artery From the Division of Pediatric Cardiology and the Department of Radiology, Rainbow Babies and Childrens Hospital and University Hospitals of Cleveland, Case Western Reserve University, Cleveland. Address for correspondence: Mark D. Jacobstein, M.D., Pediatric Cardiology. Rainbow Babies and Childrens Hospital, 2101 Adelbert Rd., Cleveland, OH 44106. Received May 11, 1984; revision accepted July 12, 1984. 650 shunts were successfully visualized.' We have since had the opportunity to study a much larger number of patients with these shunts. This article presents our experience with GMRI in depicting a variety of systemic-pulmonary artery shunts in young patients with cyanotic congenital heart disease. Patients and methods Eleven patients with cyanotic congenital heart disease in whom 17 palliative systemic-pulmonary artery shunts were present underwent evaluation by GMRI. Five patients had pulmonic atresia: two with intact ventricular septum, two with large ventricular septal defects, and one in association with a univentricular heart. Three patients had tricuspid atresia and the remaining three had atrioventricular septal defects, with severe pulmonic stenosis in two and a tight pulmonary arterial band in the other. Six patients had two shunts. There were nine BlalockTaussig shunts, six Glenn shunts, one Waterston shunt, and one Potts shunt. Patients ranged in age from 1 to 21 years (mean 13). Only one patient was under 2 years of age. This patient was sedated with 2 mg/kg meperidine. 1 mg/kg promethazine, and 1 mg/kg chlorpromazine. Informed consent was obtained in all cases from the patient or the patient's guardian. Images were obtained with a superconducting magnet operating at 0.3 T (Technicare; Solon, OH). At this field strength the resonant frequency for protons is 12.85 MHz. Infants were placed within a smaller 29 cm head-coil radio antenna inserted into the 100 cm magnet bore to improve signal-to-noise ratio. A 90 degree saturation-recovery pulse was followed in 15 msec by a 180 degree pulse, resulting in a peak ''readecho 30 msec after the 90 degree pulse. Image acquisition was gated to systole by initiating the 90 degree saturation-recovery pulse at a predeCIRCULATION by gest on A ril 9, 2017 http://ciajournals.org/ D ow nladed from DIAGNOSTIC METHODS-CONGENITAL HEART DISEASE termined time after the R wave of the patient's electrocardiogram to coincide with end-systole. The electrocardiogram was transmitted to the controlling computer telemetrically as previously described.2 Multiple, nonsimultaneous, parallel sections could be obtained in three orthogonal planes: transverse, coronal, and sagittal. All patients were evaluated in transverse and coronal planes, with five to eight scans obtained in each plane. Because of time constraints, only four patients were imaged in the sagittal plane. Each scan had a thickness of 1.4 cm and, since multiple contiguous sections were obtained, an entire volume was imaged and displayed tomographically. Depending on the patient's heart rate, individual scans required from 3 to 5 min, with data collected over 256 consecutive heart beats. Since simultaneous multislice capabilities were not available at the time of this study, a complete examination required about 60 min. GMRI studies were evaluated by two of the authors (M. D. J. and B. D. F.). Since the authors were aware of the patients' clinical diagnoses, strict criteria were established for identifying successful studies. For Blalock-Taussig and Glenn shunts, a positive study required longitudinal imaging of the shunt and had to include imaging of the pulmonic anastamosis. For aortopulmonary shunts, a positive study had to identify the confluence of aorta and pulmonary artery on at least two separate, preferably orthogonal planes. GMRI findings were compared with cineangiograms or digital subtraction angiograms to verify morphology and patency of the shunt. Two-dimensional and Doppler echocardiographic studies were also available for review.