Response to Letter by Py and André
We thank Drs Py and André for their interest in our meta-analysis.1 One of our primary objectives was to determine whether noninvasive magnetic resonance angiography (MRA) could potentially replace the invasive and costly gold-standard, digital subtraction angiography, for the detection of stenoses of the internal carotid arteries. Our results showed high sensitivities and specificities for both time-of-flight and contrast-enhanced MRA in the detection of high-grade stenoses, usually defined as 70% to 99% narrowing. The data presented by Dr Py and colleagues, although more limited, are in accordance with our results.
MRA has been extensively studied for the detection of carotid artery disease. As we described, our initial literature searches of 3 databases revealed over 11 000 articles and abstracts. Following these initial searches, 72 time-of-flight MRA articles—including that of Dr Py and colleagues2—and 24 contrast-enhanced MRA articles were thoroughly screened and reviewed. The characteristics, methodology and quality of each study were examined and a number of inclusion and exclusion criteria were applied. Eventually 37 time-of-flight and 21 contrast-enhanced MRA articles (5 overlapping) were included. After careful consideration, our 2 reviewers jointly decided on exclusion of the Py et al data in the meta-analysis. Details regarding the image analyses were not provided. The article by Dr Py and colleagues also had a very small sample size (10 patients, 2 of whom died during the study): among the included articles investigating the accuracy of time-of-flight MRA, the mean sample size was 45 patients. There was also a large variability in the period of time between imaging procedures, particularly for such a small population. The range of 9 to 104 days was much longer than the ranges used in the other studies (typically between 1 and 14 days) and generally accepted for comparison and evaluation of diagnostic techniques.
We agree with Dr Py and colleagues that the combined use of duplex ultrasound with MRA may reduce or eliminate deficits present in any individual test. In our article, we suggested that the combined use of any 2 noninvasive tests, including MRA, ultrasound, or computed tomography angiography may improve the detection of moderate-grade stenoses and aid in decision-making regarding surgical or medical treatment. Prospective, large-scale studies investigating the use of duplex ultrasound with MRA, in comparison to digital subtraction angiography, are needed to quantify this combined effect.