The diagnostic value of non-contrast enhanced quiescent interval single shot (QISS) magnetic resonance angiography at 3T for lower extremity peripheral arterial disease, in comparison to CT angiography

Peripheral arterial disease (PAD) affects more than 5 million adults in the United States [1], and the morbidity associated with PAD is increasing duo to the aging population. Accurate diagnosis of PAD is essential for treatment and management [2]. Ankle brachial index (ABI) measurement is a sensitive test and is often the first line test performed to evaluated PAD [3]. Color Doppler ultrasonography (CDU) is widely utilized in patients with chronic symptomatic limb ischemia, as it is a safe, noninvasive, and inexpensive examination [4]. However, CDU does not fully depict the entirety of the lower extremity vasculature and is operator dependent. Digital Subtraction Angiography (DSA) is considered the clinical gold standard for the diagnosis of PAD due to its high spatial and temporal resolution [5]. In addition, treatment of significant stenoses with balloon angioplasty and stenting can be performed concurrently. However, DSA is utilized relatively infrequently in practice as it is an invasive and expensive test and relies upon the use of ionizing radiation. Computed tomography angiography (CTA) and magnetic resonance angiography (MRA) are non-invasive alternatives to DSA for the depiction of PAD [6, 7]. The diagnostic performance of CTA and contrast-enhanced magnetic resonance angiography (CE MRA) have been reported as comparable to that of DSA in many studies [8, 9]. However, the high incidence of renal insufficiency in patients with PAD raises the concern for nephrogenic systemic fibrosis (NSF) with respect to contrast enhanced MRA [10, 11]. The risk of NSF is eliminated with non-contrast enhanced magnetic resonance angiography (NCE-MRA) techniques which have been developed as alternatives to CE-MRA [12‐14]. The Quiescent Interval Single-shot (QISS) technique was developed as a safe and simple “push-button” NCE MRA technique. It uses saturation pulses to suppress background and venous signal, and is ECG-gated to synchronize data acquisition with maximal arterial inflow. The use of a single-shot 2D balanced Steady State Free Precession (b-SSFP) sequence results in clear depiction of arteries. The diagnostic performance of QISS MRA has been reported to be comparable to that of CE MRA [15]. To the best of our knowledge, there is no study that has compared QISS at 3 T with CTA. Imaging at 3 T is an attractive option due to the increased SNR and the possibility to use higher parallel imaging acceleration factors. The aim of the current study is to compare non-contrast enhanced QISS MRA at 3 T with CTA for the evaluation of lower extremity arterial disease.

All 32 patients completed the CTA and MRA examinations successfully, without any adverse events. With 19 lower extremity arterial segments assessed per patient, a total of 608 segments were evaluated. Segments with nondiagnostic image quality (score 1) were not observed with QISS MRA or CTA in the present study. 3.78 % (23/608) MRA segments were rated as fair in quality (score 2) by Reader1, and 3.62 % (22/608) by Reader 2. 77.96 % (474/608) MRA segments were rated as excellent in quality (see Fig. 1) by Reader 1, and 78.29 % (476/608) by Reader 2. 0.66 % (4/608) CTA segments were rated as fair in quality (score 2) by Reader 1, and 0.33 % (2/608) by Reader 2. 92.27 % (561/608) CTA segments were rated as excellent in quality (score 4) by Reader1, and 94.08 % (572/608) by Reader 2. Total image quality of CTA was significantly higher than that of QISS MRA (P < 0.001 for both readers, see Table 1). Notably, the image quality of the two methods was not statistically significantly different at some segments (P > 0.05 for both readers, see Table 1). Of the 608 segments evaluated, 77 were heavily calcified segments. For these segments, image quality of QISS were significantly higher than that of CTA (P < 0.001 for both readers, see Table 1).

15 of fair (score 2) segments at QISS were caused by limb motion. In the pelvic region of 3 patients, vessel blurring and misregistration due to respiratory artifacts was observed. For one patient with orthopedic implant at the right femur, image quality degradation of CTA was not caused, while the QISS segment neighbor to fixator (right femoral artery) was rated as fair (score 2). Image degradation caused by the implant was not observed at other segments of this patient. For 3 patients with cardiac arrhythmia, QISS examination time was lengthened, because data acquisition was not performed every heart beat. Significant artifacts or image degradation was not observed.

Intermodality agreement in stenosis ratings was calculated on a per segment basis. Stenosis ratings were equivalent with the two methods (Fig. 2) in 94.24 % segments (573/608) for Reader 1, and 94.74 % (576/608) for Reader 2. Intermodality agreement was excellent for rating stenoses (kappa > 0.8, for both readers, see Table 2). Interobserver agreement was excellent for both CTA (0.936 ± 0.012) and MRA (0.935 ± 0.011).

In the present study DSA was performed on the segments where severe vascular disease was identified by CTA/MRA. In total, 178 segments were evaluated with DSA. Significant stenoses (≥50 %) were found in 85 segments (Fig. 3). Using DSA as the reference standard, sensitivity of QISS MRA was 94.25 and 93.26 %, and specificity was 96.70 and 97.75 %. Sensitivity, specificity and accuracy of QISS MRA for the detection of significant stenosis (≥50 %) were not statistically significantly different from those of CTA (see Table 3). Of 178 segments evaluated with DSA, 41 segments were heavily calcified. For these segments, sensitivity of QISS in detecting significant stenoses (≥50 %) was significantly higher than that of CTA (see Table 3).

Given the high incidence of renal insufficiency in PAD patients, the risk of contrast-induced nephropathy with CTA and MRA are important concerns [17, 18]. In such patients, non-contrast enhanced MRA is a promising alternative. Various NCE MRA technologies have been studied; however, none of them have been widely adopted in clinical practice duo to various difficulties. For example, 2D TOF has fallen into disuse because of lengthy imaging time and poor image quality. Electrocardiographic gated 3D fast spin-echo sequences have been shown to enable accurate imaging of the calf and pedal arteries, but these techniques requires optimal selection of systolic trigger delays [19, 20]. Flow-sensitive dephasing (FSD) prepared balanced steady state free precession sequence requires systolic and diastolic blood signal acquisition, and permits visualization of arteries with subtraction [21]. However, the FSD technique is not user-friendly as it requires the MR technologist to adjust multiple parameters from patient to patient [22]. QISS MRA is a bright blood sequential 2D NCE MRA technique originally developed for the evaluation of PAD [15]. Compared with other NCE MRA technologies, QISS has an easy “push-button” workflow, eliminating the need for extensive patient-to-patient parameter modification [23]. QISS has been compared with ABI in screening PAD [24], and has demonstrated a comparable diagnostic performance to CE-MRA for PAD evaluation [15, 25]. QISS has also been compared favorably to other NCE-MRA technique [26]. However, to the authors’ knowledge, QISS performance at 3 T has not been compared with CTA, the predominant test for PAD in many institutions because of its speed and accuracy. In the current study, QISS at 3 T was compared to CTA finding excellent intermodality agreement in the assessment of stenoses. For heavily calcified segments, the sensitivity of QISS in detecting significant stenoses (≥50 %) was significantly higher than that of CTA.

In the present study, the overall image quality of QISS MRA was lower than that of CTA, the difference being statistically significant. One possible explanation is that the spatial resolution of CTA (1 mm × 1 mm × 0.625 mm) is greater than that of QISS MRA (1 mm × 1 mm × 3 mm). Another possible explanation is that the acquisition time of CTA is much shorter than MRA, rendering it less prone to motion artifacts. However, it is worth noting that the image quality of the two methods was not significantly different at some segments.

Sufficiently suppressed background was observed in the source images of QISS MRA, resulting from the in-plane saturation pulse. The use of a gated acquisition with a b-SSFP sequence ensured high blood signal and adequate vessel-to-background contrast. Compared to other NCE-MRA techniques, QISS has superior flow contrast, because the thin-slice 2D QISS acquisition enables much greater replenishment of saturated arterial spins compared to other thick-slab 3D acquisitions. QISS is also relatively insensitive to patient motion due to the 2D acquisition in the transverse plane. NCE QISS MRA offers users the option of repeating acquisitions of arterial segments in cases of poor image quality due to motion or other technical limitations. For the current study no acquisitions were repeated due to imaging time considerations; however, this could be done in clinical practice to potentially further improve image quality.

The intrinsically high blood signal intensity with the b-SSFP sequence contributed to the image quality of QISS [27]. QISS MRA was also performed at 3 T where the intrinsic SNR is increased over lower field strengths [28, 29]. The assessment of stenoses with MRA and CTA was equivalent in most segments, and intermodality agreement was excellent. Thus, QISS MRA is a promising alternative to CTA, in particular because ionizing radiation and iodinated contrast are avoided.

2D QISS MRA with a slice thickness of 3 mm was utilized in the present study. A thinner slice thickness could be chosen; however, the number of slices required to cover the field of view would be increased with an associated increase in scan acquisition time. For example, the acquisition time doubles with slice thickness reduced from 3 mm to 1.5 mm. A lengthy acquisition poses difficulty for PAD patients with back or leg pain and the inability to keep still in supine position for extended periods. Fortunately, a 3 mm thickness was adequate for assessment of PAD in the present study, and the corresponding image acquisition time was acceptable (6.5 min for heart rate of 80 per minute).

QISS was performed in the transverse plane, similar to CTA. Because the scan direction was perpendicular to the lower extremity artery, they were sensitive to stenoses in both anterior-posterior and left-right directions, which may explain the excellent agreement in stenoses rating. In the future clinical practice, through-plane resolution could be further increased in a repeat-scan for a more detailed analysis of suspected segments.

The reported sensitivity of CTA for the detection of greater than 50 % stenosis is on the order of 89–100 % [30]. Stenosis degree may be overestimated when severely calcified antherosclerotic plaques were present [31]. Calcified plaque is not problematic with QISS MRA, likely accounting for the higher sensitivity of the technique. The diagnostic accuracy of QISS shown in the current study indicates that it may be a promising alternative to DSA.

The present study has several limitations. First, the sample size is small. More patients with lower extremity PAD should be included in the future studies. Second, DSA was available for only 19 patients. This reflects the fact that CTA currently performs sufficiently well in PAD patients to be considered the first-line examinations for most patients. Due to its invasiveness and cost, DSA was only performed on the patients with need of definitive diagnosis or endovascular therapy. Third, DSA was only performed on the diseased segments of the lower extremity arteries. Ionizing radiation and contrast administration would increase if all parts of lower extremity arterial vasculature were imaged. The number of segments evaluated by DSA was thus relatively small, weakening the statistical power. Fourth, the pedal arteries were not evaluated in the present study. Arteries of the foot are smaller in size and tortuous, thus making the image quality provided by a 2D acquisition inadequate for evaluation. However, accurate diagnosis for pedal PAD is as important as in the remainder of the lower extremities, especially in patients with diabetes [32]. Further advancement of NCE techniques will be required for sufficient evaluation of the pedal vasculature. Finally, a head-to-head comparison was not performed between QISS and other NCE MRA techniques as the focus was on comparison with CTA.

In conclusion, the extent and severity of lower extremity PAD is accurately evaluated by NCE QISS MRA with a typical acquisition time of 7–9 min. QISS is a reliable alternative to CTA for this assessment, and may be suitable as a first-line screening examination for lower extremity PAD patients with contraindications to intravenous contrast administration.