AJR: MRI may be safe for assessing great vessel stents
Some MRI sequences may provide a less invasive, safer and sufficiently accurate alternative to angiogram and CT for assessing great vessel stents, according to a study in the October issue of American Journal of Roentgenology (AJR). The study also cautioned against the more standard black-blood MRI technique, which “performed poorly” in assessing great vessel stenosis in stents.
Conventional angiography is the most accurate form of imaging for endovascular stent follow-up, but is not well-adopted because of its invasive nature. CT is currently the standard of care for routine stent imaging, but lead author Andrew M. Taylor, MD, of Great Ormond Street Hospital for Children in London, noted, “there are concerns regarding the long-term health effects of ionizing radiation exposure,” generated by CT scans. MR may offer a safer alternative. “MRI does not use radiation; however, because of artifacts it has traditionally been thought of as an unreliable method of assessing stents,” explained Taylor.
Seeking to establish a robust imaging protocol for great vessel stents, Taylor and colleagues compared angiography, CT and 10 MR sequences on nitinol, platinum-iridium and stainless steel stents.
Two reviewers assessed six image sequences for each stent to measure lumen and obstruction for non-stenosis, external stenosis and internal stenosis in in vitro pig aortas. The stents were implanted into excised, descending thoracic pig aortas and connected to heart bypass pumps to obtain visualizations with flow that could be realistically transferred to clinical imaging.
Percutaneous angiogram, the standard method for follow-up imaging of great vessel stents, demonstrated almost perfect results for the four criteria for assessing the quality of imaging technique (positive predictive value of stenosis, negative predictive value of stenosis, specificity and sensitivity). It achieved 100 percent sensitivity for all three stent types. Specificity was 100, 92 and 100 percent for nitinol, platinum-iridium and stainless steel stents, respectively.
CT performed only slightly less favorably, while MRI sequences yielded mixed results.
High flip angle through-plane gradient recalled echo (GRE) MRI performed best, scoring results roughly comparable to CT and angiogram. Like angiography, GRE MRI achieved 100 percent sensitivity for all three stent types. Specificity was 96, 100 and 88 percent for nitinol, platinum-iridium and stainless steel stents, respectively.
The authors hypothesized that GRE MRI performed so well because the high flip angle-through plane was able to overcome reduced signal intensity or signal loss of radiofrequency shielding and magnetic field homogeneity.
In-plane and through-plane steady state free precession (SSFP) and high flip angle MR angiography (MRA) were slightly less accurate than GRE MRI.
Taylor noted that black-blood MRI was highly susceptible to radiofrequency shielding, stating: “Black-blood imaging performs poorly, particularly for stents with internal stenoses.” He warned that the black-blood MR’s results “should be viewed with caution,” since this poorly performing MR technique is commonly used to assess stents.
While researchers noted that in vitro trials are associated with some challenges in extrapolation to clinical settings, the study provided a comprehensive comparison of imaging techniques for follow-up assessment of stents.
The authors acknowledged that all grades and angles of MRI suffered from some systematic mismeasurements of stent diameters, particularly with the stainless steel models. In addition, the study did not include visualizations of various levels of internal and external stenoses, which may show up more or less accurately in MRI.
Taylor concluded that the various MRI sequences provided sufficient accuracy in assessing stent stenosis, which could “allow more frequent assessment of stents at lower risk to patients and represent a significant change in clinical practice.”
Conventional angiography is the most accurate form of imaging for endovascular stent follow-up, but is not well-adopted because of its invasive nature. CT is currently the standard of care for routine stent imaging, but lead author Andrew M. Taylor, MD, of Great Ormond Street Hospital for Children in London, noted, “there are concerns regarding the long-term health effects of ionizing radiation exposure,” generated by CT scans. MR may offer a safer alternative. “MRI does not use radiation; however, because of artifacts it has traditionally been thought of as an unreliable method of assessing stents,” explained Taylor.
Seeking to establish a robust imaging protocol for great vessel stents, Taylor and colleagues compared angiography, CT and 10 MR sequences on nitinol, platinum-iridium and stainless steel stents.
Two reviewers assessed six image sequences for each stent to measure lumen and obstruction for non-stenosis, external stenosis and internal stenosis in in vitro pig aortas. The stents were implanted into excised, descending thoracic pig aortas and connected to heart bypass pumps to obtain visualizations with flow that could be realistically transferred to clinical imaging.
Percutaneous angiogram, the standard method for follow-up imaging of great vessel stents, demonstrated almost perfect results for the four criteria for assessing the quality of imaging technique (positive predictive value of stenosis, negative predictive value of stenosis, specificity and sensitivity). It achieved 100 percent sensitivity for all three stent types. Specificity was 100, 92 and 100 percent for nitinol, platinum-iridium and stainless steel stents, respectively.
CT performed only slightly less favorably, while MRI sequences yielded mixed results.
High flip angle through-plane gradient recalled echo (GRE) MRI performed best, scoring results roughly comparable to CT and angiogram. Like angiography, GRE MRI achieved 100 percent sensitivity for all three stent types. Specificity was 96, 100 and 88 percent for nitinol, platinum-iridium and stainless steel stents, respectively.
The authors hypothesized that GRE MRI performed so well because the high flip angle-through plane was able to overcome reduced signal intensity or signal loss of radiofrequency shielding and magnetic field homogeneity.
In-plane and through-plane steady state free precession (SSFP) and high flip angle MR angiography (MRA) were slightly less accurate than GRE MRI.
Taylor noted that black-blood MRI was highly susceptible to radiofrequency shielding, stating: “Black-blood imaging performs poorly, particularly for stents with internal stenoses.” He warned that the black-blood MR’s results “should be viewed with caution,” since this poorly performing MR technique is commonly used to assess stents.
While researchers noted that in vitro trials are associated with some challenges in extrapolation to clinical settings, the study provided a comprehensive comparison of imaging techniques for follow-up assessment of stents.
The authors acknowledged that all grades and angles of MRI suffered from some systematic mismeasurements of stent diameters, particularly with the stainless steel models. In addition, the study did not include visualizations of various levels of internal and external stenoses, which may show up more or less accurately in MRI.
Taylor concluded that the various MRI sequences provided sufficient accuracy in assessing stent stenosis, which could “allow more frequent assessment of stents at lower risk to patients and represent a significant change in clinical practice.”