Use of PETRA-MRA to assess intracranial arterial stenosis: Comparison with TOF-MRA, CTA, and DSA

BACKGROUND AND PURPOSE: Non-invasive and accurate assessment of intracranial arterial stenosis (ICAS) is important for the evaluation of intracranial atherosclerotic disease. This study aimed to evaluate the performance of 3D pointwise encoding time reduction magnetic resonance angiography (PETRA-MRA) and compare its performance with that of 3D time-of-flight (TOF) MRA and computed tomography angiography (CTA), using digital subtraction angiography (DSA) as the reference standard in measuring the degree of stenosis and lesion length. MATERIALS AND METHODS: This single-center, prospective study included a total of 52 patients (mean age 57 ± 11 years, 27 men, 25 women) with 90 intracranial arterial stenoses who underwent PETRA-MRA, TOF-MRA, CTA, and DSA within 1 month. The degree of stenosis and lesion length were measured independently by two radiologists on these four datasets. The degree of stenosis was classified according to DSA measurement. Severe stenosis was defined as a single lesion with >70% diameter stenosis. The smaller artery stenosis referred to the stenosis, which occurred at the anterior cerebral artery, middle cerebral artery, and posterior cerebral artery, except for the first segment of them. The continuous variables were compared using paired t-test or Wilcoxon signed rank test. The intraclass correlation coefficients (ICCs) were used to assess the agreement between MRAs/CTA and DSA as well as inter-reader variabilities. The ICC value >0.80 indicated excellent agreement. The agreement of data was assessed further by Bland–Altman analysis and Spearman's correlation coefficients. When the difference between MRAs/CTA and DSA was statistically significant in the degree of stenosis, the measurement of MRAs/CTA was larger than that of DSA, which referred to the overestimation of MRAs/CTA for the degree of stenosis. RESULTS: The four imaging methods exhibited excellent inter-reader agreement [intraclass correlation coefficients (ICCs) > 0.80]. PETRA-MRA was more consistent with DSA than with TOF-MRA and CTA in measuring the degree of stenosis (ICC = 0.94 vs. 0.79 and 0.89) and lesion length (ICC = 0.99 vs. 0.97 and 0.73). PETRA-MRA obtained the highest specificity and positive predictive value (PPV) than TOF-MRA and CTA for detecting stenosis of >50% and stenosis of >75%. TOF-MRA and CTA overestimated considerably the degree of stenosis compared with DSA (63.0% ± 15.8% and 61.0% ± 18.6% vs. 54.0% ± 18.6%, P < 0.01, respectively), whereas PETRA-MRA did not overestimate (P = 0.13). The degree of stenosis acquired on PETRA-MRA was also more consistent with that on DSA than with that on TOF-MRA and CTA in severe stenosis (ICC = 0.78 vs. 0.30 and 0.57) and smaller artery stenosis (ICC = 0.95 vs. 0.70 and 0.80). In anterior artery circulation stenosis, PETRA-MRA also achieved a little bigger ICC than TOF-MRA and CTA in measuring the degree of stenosis (0.93 vs. 0.78 and 0.88). In posterior artery circulation stenosis, PETRA-MRA had a bigger ICC than TOF-MRA (0.94 vs. 0.71) and a comparable ICC to CTA (0.94 vs. 0.91) in measuring the degree of stenosis. CONCLUSION: PETRA-MRA is more accurate than TOF-MRA and CTA for the evaluation of intracranial stenosis and lesion length when using DSA as a reference standard. PETRA-MRA is a promising non-invasive tool for ICAS assessment.