Imagers achieve peak results while cutting gadolinium dose

At the 2007 RSNA meeting, Dr. Reza Habibi presented three lower extremity MR angiography images and asked his audience to guess which one had been performed with a low dose of gadolinium.

“It’s hard to tell which belongs to the low-dose group,” said Habibi, a radiology research fellow at the University of California, Los Angeles.

The images came from a new study of 3T MRA images taken on a 32-channel scanner at three different doses of gadopentetate dimeglumine (Magnevist): high (0.3 mmol/kg), intermediate (0.19 mmol/kg), and low (0.12 mmol/kg).

Most practices are still using 1.5T systems for MRA studies, but some are in the process of migrating toward 3T. Higher field magnets offer an increased signal-to-noise ratio and higher sensitivity, making it feasible to do more with less contrast.

This ability is a distinct advantage, given the link between gadolinium-enhanced MR and the deadly skin condition nephrogenic systemic fibrosis. The association in patients with severely impaired renal function has been particularly strong at very high doses of gadolinium.

Before NSF came to light, two to four times the standard dose was commonly used in MRA. But since the safety scare, researchers have been focusing on ways to lower gadolinium dose without sacrificing image quality.

“For a long time, we were using more gadolinium than we needed to. There was no real pressure on us to adopt dose minimization strategies. We are now discovering that it is not necessary to use very large doses of gadolinium, not even for demanding MR angiography procedures,” said Dr. J. Paul Finn, chief of diagnostic cardiovascular imaging and director of MR research at UCLA.

DOSE ON SHOW

In the study presented by Habibi, researchers scanned 45 patients at the three dose levels using a dual-injection protocol. The first injection was used to image the calf station, and the second one followed after five to 10 minutes in the abdomen and thigh stations. An advantage of dual injection is that it avoids risk of venous contamination in the calves, because the imaging is always done with the first pass of contrast through the calves, Habibi explained in an interview with Diagnostic Imaging.

Two radiologists blinded to the dose categories rated overall image quality and visibility of 400 to 500 arterial segments, severity of stenoses, and venous contamination and its effect on diagnosis. From 91% to 96% of arterial segments were of diagnostic quality across the three groups. Furthermore, there was no significant difference in vessel definition between the high- and intermediate dose groups, while quality in the low-dose group was actually statistically significantly better than that in the other two groups.

“It may be that with lower dose in the calves, there is less background enhancement ‘hangover’ for the second injection than with the higher dose calf protocol,” Habibi said.

In another study presented at the RSNA session, University of Munich researchers compared images involving a 0.2-mmol/kg dose of gadobutrol (Gadovist) on a 1.5T MR scanner with those of half this dose (0.1 mmol/kg) at 3T. The images were for the assessment of renal arteries.

Image quality was comparable for proximal renal arteries and slightly higher at 3T for distal arteries, according to assessments by two independent radiologists. When it came to performance for all vessel segments, there was no statistically significant difference in image quality. Stronger vessel enhancement and better delineation of vessels were achieved at 3T compared with 1.5T, said Dr Ulrike Attenberger, who reported results at the RSNA meeting.

LOW DOSE, HIGH QUALITY

Experience at UCLA suggests contrast-enhanced MRA studies can also be performed on 1.5T systems at substantially reduced gadolinium doses compared with amounts used in the past, with excellent image quality, Finn said. Typically, MRA studies at UCLA are performed at half the pre-NSF dose at 1.5T and one-third the old dose at 3T.

The literature suggests that NSF typically occurs in patients with severe kidney failure who were exposed repeatedly to gadolinium at extremely high doses. But even at higher doses, the majority of patients with very severe kidney failure still don’t develop NSF, Finn said.

UCLA patients are screened for renal function, and, if the contrast-enhanced exam is deemed medically necessary in those with weak kidneys, it is performed at the lowest dose possible and with informed consent. Patients with normal renal function also undergo imaging at reduced gadolinium doses. In some cases, studies are done with one-quarter the dose used in the past.

“The images are terrific at lower doses. There is [often] no advantage at a higher dose,” Finn said.

The field is in a state of flux at the moment, but Finn is optimistic about the future, as lower dose imaging becomes the norm.

“NSF may disappear,” he said.

NONCONTRAST APPROACHES

Researchers have also recently reported success with noncontrast thoracic imaging, using the 3D steady-state free precession MRA technique. SSFP is a gradient-echo method that offers high temporal resolution and high SNR.

Dr. Mayil Krishnam, a cardiovascular and thoracic radiologist at UCLA, presented data on imaging of chest veins and arteries at the North American Society for Cardiac Imaging meeting in Washington, DC, in October 2007. In one study, 30 consecutive patients with known or suspected thoracic aortic disease underwent free-breathing navigator-gated ECG-triggered 3D SSFP noncontrast MRA and high-resolution 3D contrast-enhanced MRA of the thorax.

The investigators read the enhanced and unenhanced MRA data to check for pathology and to evaluate image quality from four aortic segments (root, arch, ascending aorta, and descending aorta) obtained at 1.5T. They found that noncontrast MRA with SSFP provided sufficient vascular delineation and SNRs to support the diagnosis of thoracic aortic disease.

Two readers diagnosed aneurysms of the ascending aorta in 12 patients using noncontrast SSFP MRA. They also confirmed root dilatation and coarctation in seven and three patients, respectively. The unenhanced MRA data helped the diagnosis of one patient with mural thrombosis of the aortic arch and three patients with aortic dissection as well. Total scanning time with this sequence ranged from five to 10 minutes.

The investigators found that the noncontrast MRA approach provided higher SNR than contrast-enhanced MRA in all aortic segments. As a result, there was no significant difference between readers in their ability to visualize ascending/descending aorta, the aortic arch, or the motion artifacts affecting these vessels. The noncontrast MRA sequence afforded a higher visibility of the aortic root compared with the other aortic segments.

The UCLA team assessed the feasibility of the noncontrast MRA technique in another paper focused on the central chest veins of the same set of 30 patients with known or suspected cardiovascular disease of the thorax. Researchers evaluated noncontrast SSFP scans of eight venous segments and artifacts and compared their results with conventional contrast-enhanced MR.

Two readers successfully evaluated 240 venous segments from both noncontrast and contrast-enhanced data sets. They correctly diagnosed dilatation/thrombosis and coronary sinus drainages of the superior vena cava in one and five patients, respectively. They also confirmed the patency of pulmonary artery and inferior vena cava shunts in five patients.

Again, the investigators found no statistically significant difference between readers in their ability to visualize all venous segments using contrast or noncontrast MRA.

“Noncontrast SSFP MRA provides sufficient vascular visibility to support confident diagnosis. This emerging technique appears to be valuable in the morphological evaluation of central thoracic veins in various diseases, including veno-occlusive pathologies,” Krishnam said.