Utility of magnetic resonance imaging in establishing a venous pressure gradient in a patient with possible nutcracker syndrome

A 24-year-old Caucasian woman presented with chronic pelvic congestion syndrome, interstitial cystitis, and endometriosis. Her pain was nearly constant, located in the left lower quadrant, and exacerbated by prolonged sitting or standing and after exercise. The pain radiated down the left inner thigh. She also reported dyspareunia. About 1 year ago, she also developed left upper quadrant pain that required multiple hospitalizations for pain control. Imaging at that time revealed markedly dilated left ovarian and uterine veins, and coil embolization of the left ovarian vein was performed at another hospital. However, there was no significant symptom relief, and she presented to our institution. An MRI from another hospital demonstrated tight compression of the LRV with dilation of mesenteric collaterals after left ovarian vein ligation. Results of urinalysis were negative, however. Hemodynamic evaluation of the LRV was deemed necessary to help determine if management should be focused upon alleviation of possible persistent nutcracker syndrome as the etiology. Additional velocity-sensitive MRI was then suggested. A noncontrast MR angiography evaluation of the abdomen was performed on a 1.5-Tesla MRI machine ( Fig 1 ). Axial and coronal balanced steady-state images of the abdomen and pelvis were acquired for anatomic evaluation of the arterial and venous structures and for planning of the flow-quantification sequences. These images depict flowing blood as “bright” and are often called “bright-blood images.” Velocity-encoding scout images were acquired in the peripheral LRV, ∼1 cm from the renal hilum and at the SMA crossing. These images determine if the parameters controlling velocity measurement are sufficient to measure the maximum velocity present in that vessel cross section. A velocity that is too high results in “aliasing,” which is a misrepresentation of higher velocities as incorrectly low velocity or noise. Because no aliasing was seen with a velocity-encoding maximum of 20 cm/s and 80 cm/s in the respective locations, these were used in through-plane flow sequences that measure the bulk flow and maximum velocity through a particular cross section. Through-plane quantification of flow and velocities was performed by dedicated three-dimensional laboratory technologists using Circle software (Wayne, NJ; Fig 2 ). Fig 1 A, Axial two-dimensional bright-blood images demonstrate compression of the left renal vein (LRV) by the superior mesenteric artery (SMA). Cross-section phase-contrast magnetic resonance images (PC-MRI) of the (B)  peripheral and (C) central LRV images ( red arrows ) demonstrate a minimum diameter decrease from 1.5 cm to 0.5 cm in the central LRV compared with the peripheral LRV. The green arrow indicates the aorta, and the blue arrow indicates the SMA. Fig 2 Velocity curves of the (A) peripheral and (B) compressed left renal vein (LRV). The peak velocity in the peripheral vein is 15 cm/s ( bottom curve in A ), in agreement with previously published values. Peak velocity at the stenosis was measured at 78 cm/s. There is expected variation in velocity with the cardiac cycle, due to the overriding adjacent aorta and superior mesenteric artery (SMA). The x -axis is time (seconds) and y -axis is velocity (cm/s). A, Axial two-dimensional bright-blood images demonstrate compression of the left renal vein (LRV) by the superior mesenteric artery (SMA). Cross-section phase-contrast magnetic resonance images (PC-MRI) of the (B)  peripheral and (C) central LRV images ( red arrows ) demonstrate a minimum diameter decrease from 1.5 cm to 0.5 cm in the central LRV compared with the peripheral LRV. The green arrow indicates the aorta, and the blue arrow indicates the SMA. Velocity curves of the (A) peripheral and (B) compressed left renal vein (LRV). The peak velocity in the peripheral vein is 15 cm/s ( bottom curve in A ), in agreement with previously published values. Peak velocity at the stenosis was measured at 78 cm/s. There is expected variation in velocity with the cardiac cycle, due to the overriding adjacent aorta and superior mesenteric artery (SMA). The x -axis is time (seconds) and y -axis is velocity (cm/s). The modified Bernoulli equation was used to calculate the pressure gradient across the compressed LRV segment as 4 v 2 , where v is the velocity across the stenosis. Therefore, the estimated pressure gradient caused by the SMA narrowing in this patients is: 4 × (0.78 cm 2 ) = 2.44 mm Hg. A confirmatory LRV venography was obtained via right common femoral vein access. A 5F glide catheter was advanced into the renal pelvis, and pullback pressures were measured. The pressure gradient across the compressed LRV segment was between 2 mm Hg and 3 mm Hg ( Fig 3 ). Because the pressure elevation was only moderately elevated and the urinalysis result was negative, no intervention was pursued at that time. The patient tolerated the pressure without any complications. Fig 3 Pullback pressure tracing across the left renal vein (LRV) compression. The pressure gradient was between 2 and 3 mm Hg. The x -axis is time (seconds) and y -axis is pressure in mm Hg. IVC, Inferior vena cava. Pullback pressure tracing across the left renal vein (LRV) compression. The pressure gradient was between 2 and 3 mm Hg. The x -axis is time (seconds) and y -axis is pressure in mm Hg. IVC, Inferior vena cava.

This case report demonstrates a novel application of MRI to quantify LRV hypertension via measurement of LRV flow acceleration and associated pressure gradients. The absence of high-velocity pulsatile flow makes PC-MRI less prone to artifactual results in the venous circulation. 16 The anatomic and physiologic information were both acquired in a 20-minute examination, without ionizing radiation (CT), multiple examinations, intravenous contrast material, or an invasive procedure. MRI can provide anatomic detail and physiologic pressure data to limit exposure to diagnostic venography as well as guide therapeutic decisions. The relevance is illustrated in cases such as this, in which the decision to not pursue an intervention would ideally be guided by a noninvasive MRI. An additional advantage of MRI is the potential for quantification of flow through additional contiguous vessels. Such multisite flow quantifications are commonly used in the chest to quantify shunts and coarctations. Similar application to venous structures contiguous with the LRV would characterize the distribution of antegrade vs collateral flow. Although application of MRI to measure blood flow in noncardiac vascular beds is established, the clinical literature is relatively sparse. However, the technique described here is equally applicable to evaluation of pathophysiology in other medium-large arterial and venous systems, such as femoral and carotid vessels. The challenge is in the logistics of a workflow that identifies sites of stenosis while the examination is in progress. Traditional two-dimensional PC-MRI cannot be applied generally throughout a three-dimensional imaging volume. Rather, peak velocity data must be acquired across the cross section of a vessel lumen, specified at a particular segment of the vessel. Therefore, the region of interest is identified while the imaging examination is in progress. This requires the attention of a radiologist or an experienced technologist during the examination. Nevertheless, this is quite feasible in most moderately large hospital settings, with real-time review of preliminary bright blood or postcontrast images provided by a radiologist with standard MRI training. There is no added cost for including velocity quantification. Furthermore, the continued evolution of new four-dimensional flow techniques will likely render such considerations obsolete because volumetric velocity data can be acquired over a specified time interval. This allows retrospective analysis of velocity components anywhere in the volume. 16

Further studies of both two-dimensional and four-dimensional flow applied to nutcracker syndrome and other stenotic conditions will help dissemination of this technique into routine clinical use for noninvasive hemodynamic assessment. In time, it is feasible that this will replace invasive venography but more validation and optimization will be needed.