Free‐breathing 3D pulmonary ventilation mapping at 0.55 T using stack‐of‐spiral out‐in bSSFP

Abstract

Purpose

To develop a free‐breathing pulmonary imaging technique that provides three‐dimensional (3D) structural images and regional ventilation maps, and to evaluate its repeatability and accuracy compared with two‐dimensional phase‐resolved functional lung (PREFUL) and global tidal volume.

Methods

A free‐breathing 3D stack‐of‐spiral out‐in (SOS out‐in) balanced steady‐state free precession (bSSFP) sequence with self‐navigators was designed to achieve 2‐mm isotropic resolution in 5 min. Respiratory‐resolved images were reconstructed using spatial L1 wavelet and temporal finite‐difference constraints. The 3D ventilation maps were generated based on Jacobian determinant of the estimated nonrigid deformations. Six healthy volunteers were scanned in supine and prone positions. Ventilation maps were compared with two‐dimensional PREFUL from two matched slices. Test–retest repeatability was assessed using Bland–Altman analysis. Correlations among the proposed method, PREFUL, and global tidal volume were evaluated.

Results

In healthy volunteers, the SOS out‐in lung images provided sufficient vessel‐parenchyma contrast and boundary sharpness to support accurate ventilation estimation. Regional ventilation measurements from 3D SOS out‐in demonstrated good repeatability (relative differences < 10%). Ventilation maps from 3D SOS out‐in strongly correlated with PREFUL on a slice‐matched basis as well as with global tidal volume (R²  > 0.7, p < 0.001).

Conclusion

The proposed method provides high‐quality respiratory‐resolved structural images and 3D ventilation mapping in a single 5‐min scan at 0.55 T. Ventilation measurements are sensitive, consistent, and in good agreement with PREFUL and spirometry.

1. INTRODUCTION

Lung MRI provides both structural and functional information for the screening, diagnosis, and longitudinal assessment of lung diseases without ionizing radiation. ¹ , ² , ³ , ⁴ , ⁵ Proton‐based lung MRI suffers from a low signal‐to‐noise ratio (SNR) at conventional field strengths (≥ 1.5 T). ⁶ However, contemporary 0.55T MR systems offer significant advantages for pulmonary MRI. ⁷ , ⁸ , ⁹ At 0.55 T, lung parenchyma signals are higher compared with conventional field strengths due to reduced susceptibility differences between air (alveoli) and tissue (parenchyma and blood). ¹⁰ The lung parenchyma T₂* is approximately 10 ms at 0.55 T, compared with ≤ 2 ms at 1.5 T, ⁶ , ¹¹ resulting in substantially higher SNR. Additionally, contemporary gradient hardware enables the development of fast imaging sequences, such as spiral ¹² , ¹³ , ¹⁴ and echo‐planar imaging, ¹⁵ which can improve acquisition efficiency.

Structural lung imaging at 0.55 T has been recently developed to achieve three‐dimensional (3D) isotropic high resolution under breath‐hold and free‐breathing conditions. The balanced steady‐state free precession half‐radial dual‐echo (bSTAR) ¹⁶ , ¹⁷ lung imaging leverages the high SNR of balanced steady‐state free‐precession (bSSFP) at 0.55 T with very short repetition time (TR) (≤ 2.14 ms). Under free breathing, bSTAR has demonstrated superior parenchyma details (0.9‐mm isotropic resolution) but with relatively long scan times (13 min). Acquisition time can be shortened by adopting k‐space trajectories with higher SNR efficiency. Recently, a stack‐of‐spiral out‐in (SOS out‐in) bSSFP variant has been explored to provide superior vessel‐to‐parenchyma contrasts for structural evaluation under breath‐hold conditions. ¹⁸ The spiral‐based trajectory offers more efficient sampling and flexible TR selection compared with radial trajectories. ¹⁹ , ²⁰ , ²¹ However, respiratory‐resolved lung structural imaging with bSSFP at 0.55 T has not yet been explored.

One technical challenge is the design of a respiratory navigator bSSFP. Javed et al. ¹⁴ used a superior–inferior navigator to estimate respiratory motion every 200 ms at 0.55 T using ultrashort–echo time (UTE) stack‐of‐spirals imaging. However, inserting a Cartesian‐based superior–inferior navigator is likely to disrupt the steady state of bSSFP, leading to unwanted image artifacts. ²² Free induction decay (FID) navigators, which sample the k‐space center every TR, capture respiratory motion without interrupting the bSSFP steady state and have been combined with various imaging trajectories, including 3D cones ²³ , ²⁴ and wobbling Archimedean spiral pole. ¹⁷ The high sampling rate of the FID signal is especially advantageous for capturing fast breathing motions, such as those common in pediatric subjects (e.g., infants with a resting breathing frequency of about 1 Hz ²⁵ ). Therefore, an FID‐based respiratory navigator is well‐suited for respiratory‐resolved lung structural imaging with bSSFP.

Pulmonary functional analysis is a valuable endpoint for respiratory‐resolved lung MRI. Lung ventilation can be measured based on either parenchyma deformation ²⁶ or tissue signal intensity changes. ²⁷ , ²⁸ Free‐breathing ¹H functional lung MRI has been studied for years. Recently, Capaldi et al. ²⁹ , ³⁰ validated ventilation maps derived from proton MRI against hyperpolarized gas MRI in healthy volunteers and patients with asthma. Two‐dimensional (2D) phase‐resolved functional lung (PREFUL) measures regional lung ventilation from lung parenchyma signals that are modulated by tissue density changes throughout the respiratory cycle. At 0.55 T, PREFUL has demonstrated its utility in detecting pulmonary dysfunction in patients with persistent symptoms after COVID‐19, as well as in both adults ²⁰ and pediatric populations. ³¹ However, 3D ventilation measurements using PREFUL at 0.55 T have not yet been studied. Ventilation can also be measured based on parenchyma deformation, which primarily relies on detailed vessel structures and diaphragm movements during breathing. Tan et al. proposed a respiratory‐compensated, low‐rank regularized reconstruction using 3D UTE MRI, which provides both structural and functional lung imaging at 3T. ²⁶ However, the Jacobian‐based ventilation, which relies on structural images with low vessel‐parenchyma contrast and limited vessel details, has not been rigorously evaluated.

In this study, we propose a free‐breathing high‐resolution 3D bSSFP‐based lung imaging method at 0.55 T that simultaneously provides respiratory‐resolved structural images and regional ventilation information. The proposed method uses a 3D SOS out‐in bSSFP sequence with an embedded FID navigator, constrained reconstruction, and regional ventilation mapping using Jacobian‐based methods. ²⁶ We hypothesize that the proposed method will be able to reveal detailed vessel features and provide robust, reproducible 3D ventilation maps.

The proposed technique involves several innovations: (i) Spiral‐based trajectories enable higher sampling efficiency resulting in shorter scan time, compared with 3D radial center‐out trajectories; (2) the structural lung images provide high vessel‐to‐parenchyma contrast with short TR, leveraging the high signal intensities of bSSFP at 0.55 T, which enables accurate registration for ventilation estimates ⁸ ; (iii) to our knowledge, this is the first demonstration of 3D ventilation mapping using spiral‐based bSSFP at 0.55 T, capturing regional volume changes throughout the entire respiratory cycle.

2. METHODS

2.1. Experimental methods

Imaging experiments were performed on a whole‐body 0.55T system (prototype MAGNETOM Aera; Siemens Healthineers, Forchheim, Germany) equipped with high‐performance shielded gradients (45 mT/m amplitude, 200 T/m/s slew rate). A six‐channel body array (anterior) and six elements from an 18‐channel spine array coil (posterior) were used for phantom and in vivo experiments. The study was approved by the Institutional Review Board of the University of Southern California. All subjects provided written informed consent. Six healthy volunteers (3 female/3 male, 28–35 years old) were scanned in both supine and prone positions, following a protocol described in Table 1 (see Section 2.6 for detailed description).

TABLE 1.

In vivo experimental design. The scan consisted of three sections (A, B, and C). Sections A versus B were scanned in the supine position, with a 10‐min break between the two sections to prevent gradient heating.

Start time	Postures	Section	Imaging sequences	Acquisition duration
0:00		15 min rest
15:00		A	3D SOS out‐in a	4 min 49 s
			2D PREFUL—two slices a Slice 1 repeated (a) and (b) Slice 2 repeated (a) and (b)	4 min 4 s
24:00		10‐min break
34:00		B	3D SOS out‐in a	4 min 49 s
			2D PREFUL—two slices a Slice 1 repeated (c) and (d) Slice 2 repeated (c) and (d)	4 min 4 s
43:00	Change position (3 min) and stabilization (15 min)
61:00		C	3D SOS out‐in a	4 min 49 s
			2D PREFUL—two slices a Slice 1 repeated (e) and (f) Slice 2 repeated (e) and (f)	4 min 4 s

Note: Each section included both a 3D SOS out‐in scan and a 2D PREFUL scan, with two coronal slices, each repeat twice. Section C was scanned in the prone position. After positioning, there was a 15‐min break to allow volunteers to stabilize. The center frequency was measured and adjusted before and after the SOS out‐in acquisition to correct for any frequency shifts caused by gradient heating. Total experiment time: 1 h 10 min.

^a The order of 3D SOS out‐in and 2D PREFUL is randomly switched within each section.

Abbreviations: 2D, two‐dimensional; 3D, three‐dimensional; PREFUL, phase‐resolved functional lungddd; SOS, stack‐of‐spirals.

All subjects were scanned using a free‐breathing 2‐mm isotropic‐resolution SOS out‐in imaging protocol with the following imaging parameters: flip angle = 25$°$, field of view (FOV) = 432 $\times$ 432 $\times$ 240 mm, matrix size = 216 $\times$ 216 $\times$ 120, TE₁/TE₂/TR = 0.64/2.60/3.25 ms, readout duration = 1.86 ms, peak gradient amplitude = 23 mT/m, maximum gradient slew rate = 175 T/m/s, total acquisition time = 4 min 49 s, and coronal acquisition. Center frequency was monitored and readjusted before and after each SOS out‐in sequence. PREFUL imaging parameters and scan planes selections are summarized in Appendix S1.

2.2. Pulse sequence

Figure 1A illustrates the proposed free‐breathing SOS out‐in pipeline. Nonselective excitation was generated with a 200‐μs hard pulse. For each k_z partition, a rapid spiral out‐in readout created two echoes in the k_x‐k_y plane. ¹⁸ It took approximately 14 400 interleaves (46.8 s) to fully sample the 3D image. Cartesian sampling was used along the anterior–posterior k_z direction, following a “ping‐pong” order to avoid rapid switching of k_z phase encoding in balanced SSFP, which induces eddy currents and results in image artifacts. ³² , ³³ The k_x‐k_y spiral takeoff angle started at 0$°$ and was increased by a tiny golden angle ($\mathit{GA}$ = 23.6281$°$) for every 120 TRs (120 k_z encoding steps) to minimize the eddy currents. ³⁴ A Kaiser‐Bessel windowed ramp preparation ³⁵ using seven TRs was added to minimize oscillatory transients in bSSFP imaging. The pulse sequence was implemented in the Pulseq framework. ³⁶

FIGURE 1.

Free‐breathing stack‐of‐spirals out‐in pipeline. (A) Three‐dimensional (3D) sampling scheme. Spiral out‐in interleaves are acquired in the k_x‐k_y plane, with k_z sampling following a “ping‐pong” ordering. The spiral rotation angles increase by a tiny golden angle after each complete set of k_z steps. (B) Reconstruction and ventilation analysis. Respiratory data are sorted according to the respiratory direction–dependent manner, with raw k‐space data binned accordingly. A constrained reconstruction is applied to generate respiratory‐resolved images. Regional 3D ventilation maps are generated following motion field extraction and image registration.

2.3. Respiratory navigator and data binning

An FID navigator was designed and embedded into a 3D SOS out‐in sequence for respiratory signal extraction. Specifically, a 20‐sample apparent diffusion coefficient (ADC) event of duration 0.05 ms was inserted every TR between a nonselective radiofrequency pulse and a k_z partition‐encoding gradient. The FID navigator module has a duration of 0.07 ms including two ADC deadtimes, 0.01 ms each, before and after the ADC event. The respiratory signals were then passed through angular filtering ³⁷ in the [0–0.04]rad⁻¹ angular frequency domain, followed by bandpass filtering with a passband [0.05–0.6]Hz. Coil clustering was performed to weight the signals based on highly correlated coils, resulting in a one‐dimensional respiratory signal with a length equal to the total number of TRs. Appendix S2 provides a detailed description. We verified the FID‐based navigator against pilot‐tone at 0.55 T. ³⁸

Figure 1B demonstrates the data‐binning strategy and remaining pipeline. The extracted navigator was used to bin the corresponding k‐space data in a respiratory direction–dependent manner. Respiratory directions were defined by inhalation (from end‐of‐exhalation to end‐of‐inhalation) and exhalation (from end‐of‐inhalation to end‐of‐exhalation) periods, based on the sign of the derivatives of the respiratory signals. Different respiratory stages were grouped to have an equal displacement range of the respiratory signal. The sorted signals from the two extreme displacements were combined, representing end‐of‐exhalation and end‐of‐inhalation states. k‐Space data were binned into different number of respiratory stages based on the motion estimates.

2.4. Image reconstruction

All reconstructions were performed offline using MATLAB R2021a (MathWorks, Natick, MA, USA). Raw data sets were transferred and processed using ISMRMRD format. ³⁹ The spiral out‐in trajectory was measured using the method proposed by Zhao et al. ⁴⁰ with a spherical ball phantom of 14 cm in diameter. The k‐space data were prewhitened using the correlation matrix calculated from noise measurements. ⁴¹ Coil sensitivity maps were estimated with ESPIRiT ⁴² and were assumed to be time‐invariant. An estimation‐subtraction method ⁴³ was used to mitigate the spiral aliasing artifacts originating from arm regions outside the FOV.

Respiratory‐resolved images $X$ were reconstructed with an iterative constrained reconstruction using spatial L1‐based wavelets ($\mathrm{\Phi }$) and temporal finite difference along the respiratory dimension. The cost function is given as follows:

$$\underset{X}{\text{argmin}}\sum _{m=1}^M{‖A{X}_m-{\mathrm{b}}_m‖}_2^2+{\lambda }_s{\Vert \mathrm{\Phi }X\Vert }_1+{\lambda }_m{‖{\mathit{FD}}_m(X)‖}_1$$ (1)

where $X_m$ is vectorized respiratory‐resolved 3D images parametrized by the respiratory state index $m$ ($m$ = 1,2,…, $M$); $A$ is the encoding function (including coil sensitivity and nonuniform fast Fourier transform; ${\mathrm{b}}_m$ represents multicoil k‐space data for each sorted respiratory dimension; $\mathrm{\Phi }$ represents a spatial 3D L1‐wavelet transform; ${\mathit{FD}}_m$ represents a first‐order finite difference along the respiratory motion dimension; and ${\lambda }_s$ and ${\lambda }_m$ are the spatial and motion regularization parameters, respectively.

This optimization problem was solved using the alternating direction method of multipliers algorithm, ⁴⁴ as implemented in the Berkeley Advanced Reconstruction Toolbox. ⁴⁵ The spiral out and in acquisitions were reconstructed separately and were then combined using the square root of the sum of squares.

2.5. Ventilation analysis

All quantitative measurements were performed in MATLAB. Lung parenchyma was segmented using the Total Segmentator. ⁴⁶ The end‐of‐exhalation and end‐of‐inhalation states of each subject were segmented. A 3D tidal volume change (in percentage) was then calculated by dividing the difference in volumes between the two extreme stages by the end‐of‐exhalation stage. PREFUL images were processed using the MR Lung 2.2.0 research package (Siemens Healthineers), ²⁷ resulting in a fractional ventilation map (in percentage).

The end‐of‐exhalation frame, typically considered as the stable state, was selected as the reference frame for registration. All other respiratory state images were registered to this reference via Demons nonrigid registration ⁴⁷ (4 pyramid levels and 100 iterations). Jacobian‐based regional ventilation maps ($R_{\mathit{\text{vent}}}$) were then calculated from the estimated respiratory fields as follows:

$$R_{\mathit{\text{vent}}}(\%)=\frac{V_i-V_e}{V_e}\times 100\%=\frac{V_i}{V_e}-1=\mid \det (\nabla M+\mathit{Id})\mid -1$$ (2)

where $V_i$ and $V_e$ are the lung volumes at ith respiratory state and the end‐of‐exhalation state, respectively; $\nabla M$ is a 3 $\times$ 3 matrix for each voxel, representing the gradient ($\nabla$) of the respiratory motion fields ($M$); $\mathit{Id}$ is a 3 $\times$ 3 identity matrix; $\det$ demotes the determinant of the matrix; and $R_{\mathit{\text{vent}}}>0$ indicates volume expansion, whereas $R_{\mathit{\text{vent}}}<0$ indicates volume contraction. Because the reference frame is the end‐of‐exhalation, where the volume is often at a minimum, $R_{\mathit{\text{vent}}}$ is typically positive, representing voxel‐wise expansion throughout the respiratory cycle.

2.6. In vivo experimental study design

Table 1 illustrates the experimental design. The total scan included three sections (A, B, and C), lasting 1 h 20 min. Sections A versus B were used to test intrascan test–retest repeatability for 3D SOS out‐in and PREFUL. Sections A or B versus C were used to evaluate posture‐related gravity dependence of ventilation. Sections A and B were scanned in the supine position, with the volunteer remaining on the patient table, with a 10‐min break between the two sections. Section C was scanned in the prone position with the arms aside the torso. Each scan section lasted approximately 9 min and included both a 3D SOS out‐in and 2D PREFUL with each coronal slice measured twice. The order of the 3D SOS out‐in and 2D PREFUL acquisitions was randomly alternated within each section. A 15‐min break was taken before each repositioning, which includes the time for changing of the position (approximately 3 min) and allowing the volunteer to stabilize. Appendix S1 provides detailed descriptions and considerations related to the experiments design.

2.7. Statistical analysis

Bland–Altman analysis of the mean of the ventilation maps was performed between the anterior and posterior slices of PREFUL within 2 min and within 15 min, as well as between repeated SOS out‐in scans within 15 min. Correlation plots of regional ventilation were generated for comparisons between 2D PREFUL and 3D SOS out‐in, and between 3D SOS out‐in and 3D volume changes (%). Relative differences were calculated by dividing the difference between two measurements by their mean, and these were used to assess repeatability, with a threshold of less than 10% defining good repeatability. A significant level of p < 0.001 was considered for correlation analyses.

3. RESULTS

Figure 2 contains representative structural images and ventilation maps from the entire respiratory cycle. Structural images exhibit clear diaphragm boundaries and well‐defined vessel structures. The end‐of‐exhalation frame shows superior image quality with sharpest diaphragm boundaries in coronal views compared with other respiratory frames. Ventilation maps reveal overall homogeneous ventilation patterns during the respiratory cycle across all three views, with the highest values occurring during the end‐of‐inhalation frame. In sagittal views, higher ventilation values are observed in the posterior regions.

FIGURE 2.

Structural images and ventilation maps from 1 representative healthy volunteer (31 years old; male). Coronal (first row), axial (second row), and sagittal (third row) views reconstructed from the combined echoes in different respiratory states across the respiratory cycle (A) and ventilation maps (B). Diaphragm movement can be seen by comparing with the fixed red dashed lines. All respiratory‐motion states show clear vessel delineations and sharp diaphragm boundaries. (B) Quantitative ventilation maps show an overall increase in ventilation values from the end‐of‐exhalation frame, with maximum values observed during the end‐of‐inhalation frame, indicating a realistic ventilation trend during tidal breathing. Video S1 demonstrates a movie display of three‐dimensional structural images and ventilation maps throughout the slices, including images from separate echoes.

Figure 3 represents the results of the repeatability studies for regional ventilation in 3D SOS out‐in and 2D PREFUL. Figure 3A shows the Bland–Altman analysis of the mean regional ventilation values is shown for PREFUL scans within 2‐min (i), 15‐min (ii), and two 15‐min separated SOS out‐in scans (iii), along with mean and median ventilation values for 3D SOS out‐in (iv). The mean bias between PREFUL scans within 2 min was 0.3% (−1.69, 2.30) in the anterior slice and 0.74% (−1.54, 3.02) in the posterior slice. The bias between PREFUL scans within 15 min was 0.08% (−1.48, 1.63) for the anterior slice and 0.21% (−2.92, 2.5) for the posterior slice. For SOS out‐in, the bias within 15‐min repeated measurements was 0.19% (−0.88, 1.26) for the anterior slice and 0.78% (1.63, 3.18) for the posterior slice. The bias for 3D ventilation values of SOS out‐in are 0.35% (−0.33, 1.03) and 0.25% (−0.85, 1.35) of mean and median values. Overall, posterior slices showed slightly higher bias compared with the anterior slices for both PREFUL and SOS out‐in. The relative differences in all repeated experiments (i, ii, and iii) were less than 10%, indicating good repeatability. Figure 3B shows the representative SOS out‐in ventilation maps for single‐slice measurements in coronal, sagittal, and axial views for 2 healthy male volunteers (30 and 31 years old). The ventilation patterns exhibit strong agreements between the two scans. Orange boxes highlight similar ventilation patterns in the coronal and axial views in Volunteer 1 and 2, respectively. Figure S1 shows the SOS out‐in ventilation maps from 15‐min repeated scans of the other 4 healthy volunteers. Figure S2 displays PREFUL regional ventilation maps in 2‐min and 15‐min repeated scans of all 6 healthy volunteers.

FIGURE 3.

Repeatability studies of ventilation. (A) Bland–Altman analysis of the mean regional ventilation for phase‐resolved functional lung (PREFUL) in 2 min (i) and 15 min (ii), stack‐of‐spirals (SOS) out‐in with matched two‐dimensional slices in 15 min (iii), and mean and median three‐dimensional ventilation of SOS out‐in in 15 min (iv). Anterior and Posterior slices are represented by different colors. Solid lines represent the mean bias, and dashed lines indicate the 95% confidence interval. (B) Representative SOS out‐in ventilation maps from 15‐min repeated scans in three views for 2 healthy male volunteers (30 and 31 years old). The mean ventilation values for the two‐dimensional slices are reported. Red dashed boxes highlight the regional ventilation areas that are matched between Scan A and Scan B.

Figure 4 compares the ventilation results between SOS out‐in and PREFUL. Ventilation quantification using SOS out‐in was significantly correlated with global tidal volume change (R² = 0.8738, p < 0.0001) and PREFUL measurements (R² = 0.8967, p < 0.0001) at matched slice locations. Regional ventilation patterns generally match well between PREFUL and SOS out‐in. In particular, posterior slices show consistently high ventilation values. Orange boxes highlight regions of higher regional ventilation in the same coronal areas in Volunteer 5, as seen in both PREFUL and SOS out‐in. However, unmatched regions, such as the corners of the diaphragm in Volunteer 1 and 2, may result from registration errors. Figure S3 illustrates the Bland–Altman analysis of mean of regional ventilation for SOS out‐in and PREFUL, SOS out‐in, and 3D tidal volume changes.

FIGURE 4.

Ventilation (Vent) comparison with phase‐resolved functional lung (PREFUL). (A) Correlation plots showing the correlation between the mean ventilation values for three‐dimensional (3D) stack‐of‐spirals (SOS) out‐in and 3D volume changes (left) and the correlation between the mean ventilation values for 3D SOS out‐in and two‐dimensional (2D) PREFUL (right). (B) Representative slice‐matched regional ventilation maps (anterior and posterior slices) for PREFUL (first row) and SOS out‐in (second row) from all 6 volunteers. The means of the ventilation values are reported. Regional ventilation patterns are well‐matched in the indicated areas (red dashed boxes). A notable increase in ventilation values is observed at the corner of the diaphragm in the SOS out‐in results (red arrow), probably due to the misregistration.

The selection of flip angle of the sequence, regularization parameters during reconstruction, and number of respiratory states have been optimized (see Figures S4–S6). Further details have been summarized in Appendix S3.

4. DISCUSSION

In this study, we demonstrate a 5‐min free‐breathing lung imaging method that provides 2‐mm isotropic structural images and 3D regional ventilation maps. The structural lung images exhibit good vessel‐parenchyma contrasts, enabled by short TR (3.25 ms) and bSSFP readout ⁴⁸ at 0.55 T, which supports ventilation estimation. The ventilation results show good intrascan repeatability (relative differences < 10%) within a 15‐min time gap and are strongly correlated with 2D PREFUL and global tidal volume (R² > 0.7, p < 0.001). To our knowledge, this is the first demonstration of 3D regional ventilation mapping using spiral‐based bSSFP, showing both feasibility and repeatability. This method warrants further investigation in a larger cohort, including in patients with pulmonary dysfunction and a greater range of regional ventilation defects, such as chronic obstructive pulmonary disease and asthma.

The 3D spiral out‐in bSSFP sequence can be combined with variant magnetization preparation pulses, expanding the possibilities for lung imaging at 0.55 T. The high vessel‐to‐parenchyma contrast provided by this sequence allows for detailed exploration of pulmonary perfusion, such as through arterial spin labeling with spiral sampling. ⁴⁹ By combining SOS out‐in ventilation mapping with perfusion assessments, we can achieve a comprehensive 3D Ventilation/Perfusion analysis of the lung at 0.55 T, offering enhanced diagnosis capabilities and deeper insights into pulmonary physiology. T₂ and/or diffusion preparation could also be combined with the proposed method to create tissue contrasts that are useful for lung cancer screening and nodule characterization.

Our results reveal a gravitational gradient in regional ventilation, with increased ventilation in the dependent lung regions (see Supporting Information S2 and Figure S7). This gravity dependence in ventilation has been observed using PREFUL as well as other imaging modalities such as single‐photon emission computed tomography and positron emission tomography. ⁵⁰ , ⁵¹ , ⁵² All prior work has reported physiologic vertical gradients with increased ventilation with decreasing lung height. Our results show a larger gradient in the prone position (11.56% difference in ventilation across the lung) compared with the supine position (5.76% difference), which is consistent with the studies using PREFUL. ⁵⁰ Similar results have been reported in positron emission tomography studies, which found an averaged of 8% ventilation heterogeneity. ⁵²

The proposed method estimated regional ventilation of 7.59%–17.08%, which was consistent with 7.76%–16.07% measured in the same subjects using PREFUL. It is worth noting that these values are comparable with 11% reported in the literature for healthy subjects using the gold standard method (spirometry). ⁵³ This suggests that comparison with spirometry, and a nuanced examination of potential discrepancies between modalities, would be useful to include in future studies.

This work has limitations. First, parameter optimization was performed using long breath holds, because we had access to an exceptional volunteer who could perform them with ease. When reproducing this work, we suggest shorter breath‐hold durations of about 17 s with slightly coarser spatial resolution, as was used by Tian et al. ¹⁸ Next, this study only involves healthy volunteers; future work includes validating the ventilation measurements and analysis on patients with ventilation defects. The ventilation results rely on a Jacobian‐based method, which uses the calculated respiratory fields during nonrigid registration. However, intensity‐based ventilation measurements (i.e., specific ventilation ²⁶ ) were not considered in this study. Signal differences between end‐of‐inhalation and end‐of‐exhalation were not adequately sensitive. Further investigation into the reconstruction is needed to account for signal differences between respiratory states. Additionally, because this study provides global repeatability and accuracy comparisons between SOS out‐in and PREFUL in a slice‐matched manner, there are limitations in the registration process. Specifically, PREFUL is limited in 2D registration and does not account for 3D through‐plane motion, which may lead to subtle regional differences in ventilation between 2D PREFUL and 3D SOS out‐in, especially in regions near distal vessels and the heart. Future studies should include a complete comparison in 3D ventilation, potentially using hyperpolarized gas ⁵⁴ MRI, to further validate the 3D ventilation maps.

5. CONCLUSION

We present a free‐breathing 3D pulmonary structural and ventilation method at 0.55 T that combines a 3D stack‐of‐spiral out‐in bSSFP acquisition with an embedded respiratory navigator, a constrained reconstruction, and a nonrigid registration. This method provides respiratory‐resolved lung images with 2‐mm isotropic resolution and generates 3D regional ventilation maps in just 5 min. The proposed ventilation analysis demonstrates good repeatability and strong correlations in both 2D PREFUL in matched slices and 3D tidal volumes during free breathing. Overall, 3D‐SOS out‐in ventilation offers a feasible, repeatable, and sensitive method of measuring regional ventilation in healthy volunteers.

CONFLICTS OF INTEREST

Ziwei Zhao is currently an employee of Vista AI, Inc. and completed all work associated with this manuscript while at the University of Southern California. Sophia X. Cui is an employee of Siemens Medical Solutions, USA.

Supporting information

Data S1. Supporting information.

Video S1. Individual echo and echo‐combined images: three‐dimensional (3D) stack‐of‐spiral (SOS) out‐in structural lung images and ventilation maps from 1 healthy volunteer. All coronal slices from anterior to posterior are shown. The movie includes coronal (first row), axial (second row), and sagittal (third row) views of images reconstructed from the first echo, second echo, and combined echoes in exhalation state (A) and inhalation states (B). (C) Representative ventilation maps in the corresponding coronal, axial, and sagittal views. Diaphragm positions are marked with red dashed lines.

Video S2. Gravity dependence of the stack‐of‐spiral (SOS) out‐in ventilation maps. Axial views of three‐dimensional (3D) ventilation maps in prone (left) and supine (right) positions for 1 healthy volunteer. In the prone position, ventilation values are higher in anterior area (red arrows), whereas in the supine position, ventilation values are higher in the posterior areas (red dashed boxes).

Zhao Z., Lee N. G., Tasdelen B., et al., “Free‐breathing 3D pulmonary ventilation mapping at 0.55 T using stack‐of‐spiral out‐in bSSFP ,” Magnetic Resonance in Medicine 95, no. 1 (2026): 430–441, 10.1002/mrm.70069.

DATA AVAILABILITY STATEMENT

Sample reconstructed data are available via Zenodo: https://zenodo.org/records/14788613?preview=1&token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6IjdkNTg1MmMzLTYxNmItNDA4MC1hZjY3LTNlOTFkZTBhYTU4ZiIsImRhdGEiOnt9LCJyYW5kb20iOiI4YzMxYTEzY2EwYzEwYzg0ZDUyMDJmMzNiMTRmN2YwNyJ9.3YihelhqXpzJuYq5QS7ROdEniiL93j6VKVC1sG_2khopJx2CVRc‐pxIiP1RMb25y1cVJOWdOcJXM5h9oJsIgMA.