Imaging was performed on a 3T GE MR 750 scanner (GE Healthcare, Waukesha, WI). A 32 channel phased array coil was placed on the abdomen and pelvis. To avoid prolonging the exam time for coil repositioning, the body coil was used for the extremities. Patients were scanned in the supine position. A single dose (0.12 mL/kg body weight, i.e. 0.03 mmol/kg) of gadofosveset trisodium (Ablavar, Lantheus Medical Imaging, Billerica, MA) was diluted in saline to a volume of at least ten milliliters and administered intravenously at a rate of 0.5 mL per sec using a power injector (MEDRAD, Inc., Warrendale, PA) via peripheral venous access (22 gauge angiocatheter in an antecubital vein), followed by a 20 mL normal saline solution flush.
After a two minute delay, post-contrast axial images were obtained in multiple stations, from superior to inferior, with a dual-echo RF-spoiled gradient-recalled echo T1-weighted sequence with Dixon-based fat-water separation (LAVA-Flex). Parameters were: a ~4.5 msec repetition time, a ~2 msec echo time, a 12° flip angle, a field of view between 28–38 cm, and an 166 kHz bandwidth. For the abdomen and/or pelvis, images were obtained with 320 × 256 matrix, 3 mm slice thickness, 1.5 mm slice spacing, 72 – 244 per station, and acceleration factor of 2–3; the sequence was either performed as a single breath-hold of approximately 25 seconds, or modified to enable navigation for respiratory motion correction. In the lower extremities, images were obtained with 320 × 384 matrix; 1.6–3 mm slice thickness; 0.8–1.5 mm slice spacing; 152–504 slices per station, two signal averages, and 1–5 minutes per station. The total acquisition time for the entire exam averaged 11 ± 3.5 minutes (range 7 – 15 min). By comparison, based on a repetition time of 20 ms and a phase matrix of 200, 75 cm of coverage obtained at 3 mm slice thickness would conservatively require approximately 40 minutes for a time-of-flight MR venogram.
After a rapid MR venography technique was established, exams were performed for clinical purposes. However, given the new imaging approach and our intention to evaluate it, the study was approved by our Institutional Review Board, and informed consent/assent was obtained prior to inclusion in the study. Nine consecutive patients referred for clinically indicated studies were recruited for our study from October, 2009, to April, 2011. Patients ranged from 1 – 16 years of age (average 12.4 ± 5.2 years). Depending on the clinical indication, patients were imaged in different stations from the diaphragm to the knees. Representative images from venograms are displayed ( – ). Examples of the various pathologies identified in these patients using this method are also included ( – ).
Figure 1 Representative MR venographic images of the suprarenal inferior vena cava of a 14 year old female patient. White arrows on the a) axial source image and b) coronal oblique maximum intensity projection image indicate the position of the suprarenal inferior (more ...)
Representative MR venographic image of the popliteal veins of the same patient as in . White arrows on the a) axial source image and b) coronal oblique maximum intensity projection image indicate the position of the popliteal veins.
Figure 9 Example of a venous malformation. a) Axial source image and b) coronal maximal intensity projection image centered below the knee joint of a 15 year old male demonstrate findings of a venous malformation (white arrows) located in the subcutaneous fat (more ...)
Figure 12 Example of mass effect from a tumor narrowing a pelvic vein. a) Axial source image and b) coronal curved planar reformatted maximum intensity projection image of the pelvis demonstrate a large, enhancing mass in the right pelvis, compatible with a desmoid (more ...)
The veins were assessed segmentally to demonstrate that this imaging technique results in adequate and homogeneous conspicuity of the veins relative to the surrounding anatomy. Noise in the setting of parallel imaging is spatially varying, so a simple region of interest in the air surrounding the patient is not a valid measurement of noise [8
]; therefore, in the abdomen and pelvis where parallel imaging was used, a contrast to background ratio (CBR) was calculated. A region of interest was placed in the venous segment, and the signal intensity of the vessel segment was divided by the similarly derived signal intensity of adjacent muscle that demonstrated homogeneous signal intensity. In the lower extremities, where parallel imaging was not utilized, the contrast to noise ratio (CNR) was also determined by measuring the signal intensity in the vein, subtracting the signal intensity in the adjacent muscle, and dividing the result by the standard deviation of the signal in the air [8
]. Calculations for the bilateral common and external iliac veins were averaged together, as were results for the bilateral common femoral and proximal, mid and distal superficial femoral veins, and also the bilateral popliteal veins, respectively. The CBR for the inferior venae cavae was 3.6 ±
0.6; for the iliac veins, 3.2 ±
1.0; for the femoral veins, 2.9 ±
0.6; and for the popliteal veins, 3.0 ±
0.7 (numbers represent the mean ±
SD from between 3 – 9 different patients). The CNR was 27.4 ±
6.8 for the femoral veins and 26.5 ±
7.2 for the popliteal veins (mean ±
SD, 3 – 9 different patients). These numbers suggest the method will consistently yield good contrast between veins and adjacent tissues.