Single-Shot Free-Breathing 2D bSSFP Stack for Non-contrast Aortic MRA: Paired Comparison with iNAV 3D bSSFP

Background: Non-contrast MRA avoids gadolinium use, offering safe vascular imaging for patients with contrast contraindications. While free-breathing 3D balanced SSFP (bSSFP) MRA with image-based navigation (iNAV) enables clinically feasible non-contrast aortic imaging^1,2, residual respiratory motion, prolonged acquisition times, saturation-related contrast loss, and flow-related dephasing artifacts can limit image quality and throughput. We evaluated Overlap-MSTAR—a rapid technique reconstructing 3D aortic MRA from overlapping single-shot 2D ECG-triggered bSSFP slices using reference-free 2D-to-3D registration³. By preserving blood signal in rapid 2D slice-selective acquisitions, Overlap-MSTAR achieves ~1-minute scan times while mitigating saturation, motion, and flow artifacts. This pilot study determines whether Overlap-MSTAR provides accurate aortic diameter measurements equivalent to iNAV 3D bSSFP (±3 mm margin⁴) and evaluates its capability to rescue suboptimal iNAV cases.

Methods: Fifty consecutive patients (mean age 51.4 ± 17.3 years; 46% males) underwent both techniques in a single session on a 1.5T scanner (Siemens Healthineers, MAGNETOM Sola). Imaging parameters are summarized in Table I. A blinded cardiologist scored image quality (IQ, 1–5: blood signal, edge sharpness, artifacts) and measured inner-to-inner wall diameters at the aortic root, ascending aorta, and proximal arch. Diagnostic adequacy was IQ ≥3. Statistical analysis included Two One-Sided Tests (TOST) for equivalence (±3 mm margin), Bland–Altman agreement, and concordance correlation coefficient (CCC). McNemar's test evaluated rescue capability where iNAV failed (IQ < 3) but Overlap-MSTAR succeeded (IQ≥3).

Results: All patients completed both acquisitions successfully. Overall image quality was higher with iNAV (3.69 ± 0.93) than Overlap-MSTAR (2.88 ± 0.39), see Fig. 1A. Non-diagnostic cases (IQ < 3): 6 iNAV-only, 7 Overlap-MSTAR–only, and 1 both (36/50 diagnostic on both). McNemar's analysis demonstrated that Overlap-MSTAR rescued 6/7 iNAV failures (85.7%; 95% CI: 0.49–0.97), reducing failure rates from 14% to 2% (p=0.031). Aortic diameter measurements showed a systematic +1.0 mm bias, but TOST confirmed clinical equivalence within the ±3 mm margin at all locations, see Table II. Bland–Altman plots (Fig. 1B) demonstrated a consistent small bias with tight agreement and no proportional bias. CCC indicated substantial to almost perfect measurement agreement between two methods.

Conclusion: Overlap-MSTAR effectively complements iNAV, achieving clinically equivalent aortic sizing within ±3 mm margins and 85.7% rescue rate for failed cases while reducing scan time >70% to ~1 minute. The systematic +1.0 mm underestimation is correctable with calibration. Technical limitations include reduced edge sharpness from thicker slices. Future optimization of slice thickness and reconstruction algorithms should enhance spatial resolution while preserving time advantages, enabling clinical implementation as backup technique.

Comparison of Imaging Parameters Between 3D iNAV and 2D Overlap-MSTAR SSFP Sequences

	iNAV	Overlap-MSTAR
MR acquisition type	3D	2D
Image Matrix	272 × 272	256 × 192
Pixel Size (mm)	1.47 × 1.47	1.56 × 1.56
Flip Angle	90	60
TE (ms)	1.47	1.18
Bandwidth (Hz/Pixel)	875	849
Acquisition Window(ms)	242	214
# of Slices	113 ± 15	63 ± 13
Acquisition Time(sec)	250 ± 76	66 ± 22
Slice Orientation	Coronal	Sagittal
FOV (mm)	400 × 400	400 × 300
Slice Thickness/Gap (mm)	1.2 / 1.2	6.0 / 2.0
T2 Preparation	Yes	No
Fat Saturation	Yes	No

Aortic Diameter Measurements (mm, mean ± SD)

Location	iNAV	Overlap-MSTAR	Difference	TOST p-value	CCC [confidence Interval]
Aortic Root	36.06 ± 4.77	35.05 ± 5.18	+1.01	< 0.001	0.95 [0.91 0.97]
Ascending Aorta	36.25 ± 7.43	35.25 ± 7.25	+1.00	< 0.001	0.98 [0.96 0.99]
Proximal Arch	32.23 ± 6.61	31.14 ± 6.69	+1.09	< 0.001	0.97 [0.95 0.99]

Figure 1 A). Aortic root views from 3D iNAV (left) and Overlap-MSTAR (right). iNAV shows residual respiratory motion artifacts, whereas Overlap-MSTAR is artifact-free at the expense of mild edge blurring. B). The Bland-Altman plot of aortic root, ascending aorta, and proximal arch measurements. The mean difference of three measurements are: 1.01 mm, 1.00mm, and 1.09mm. The upper/lower limit of agreements of three measurements are: (3.52 mm, -1.49 mm), (3.21 mm, -1.21 mm), and (3.28 mm, -1.09 mm).