Simultaneous T1, T2, and T2* Mapping of Carotid Plaque: The SIMPLE* Technique
Abstract
Background
Quantitative T1, T2, and T2* measurements of carotid atherosclerotic plaque are important in evaluating plaque vulnerability and monitoring its progression.
Purpose
To develop a sequence to simultaneously quantify T1, T2, and T2* of carotid plaque.
Materials and Methods
The simultaneous T1, T2, and T2* mapping of carotid plaque (SIMPLE*) sequence is composed of three modules with different T2 preparation pulses, inversion-recovery pulses, and acquisition schemas. Single-echo data were used for T1 and T2 quantification, while the multiecho (ME) data were used for T2* quantification. The quantitative accuracy of SIMPLE* was tested in a phantom study by comparing its measurements with those of reference standard sequences. In vivo feasibility of the technique was prospectively evaluated between November 2020 and February 2022 in healthy volunteers and participants with carotid atherosclerotic plaque. The Pearson or Spearman correlation test, Student t test, and Wilcoxon rank-sum test were used.
Results
T1, T2, and T2* estimated with SIMPLE* strongly correlated with inversion-recovery spin-echo (SE) (correlation coefficient [r] = 0.99), ME-SE (r = 0.99), and ME gradient-echo (r = 0.99) sequences in the phantom study. In five healthy volunteers (mean age, 25 years ± 3 [SD]; three women), measurements were similar between SIMPLE* and modified Look-Locker inversion recovery, or MOLLI (1151 msec ± 71 vs 1098 msec ± 64; P = .14), ME turbo SE (31 msec ± 1 vs 31 msec ± 1; P = .32), and ME turbo field echo (24 msec ± 2 vs 25 msec ± 2; P = .19). In 18 participants with carotid plaque (mean age, 65 years ± 9; 16 men), quantitative T1, T2, and T2* of plaque components were consistent with their signal characteristics on multicontrast images.
Conclusion
A quantitative technique for simultaneous T1, T2, and T2* mapping of carotid plaque with 100-mm3 coverage and 0.8-mm3 resolution was developed using the proposed SIMPLE* sequence and demonstrated high accuracy and in vivo feasibility.
© RSNA, 2023
Summary
A quantitative MRI technique for simultaneous T1, T2, and T2* mapping of carotid plaque (SIMPLE*) demonstrated comparable accuracy with reference standard sequences and in vivo feasibility.
Key Results
■ A technique for simultaneous T1, T2, and T2* mapping of carotid plaque (SIMPLE*) using interleaved single- and multiecho (ME) three-dimensional golden angle radial acquisition was developed.
■ In a phantom study, the SIMPLE* technique correlated with inversion-recovery spin-echo (SE) (r = 0.99), ME-SE (r = 0.99), and ME gradient-echo (r = 0.99) sequences in T1, T2, and T2* measurements, respectively.
■ In an in vivo study of 18 participants with carotid plaque, SIMPLE* quantitative images showed similar characteristics to conventional imaging.
Introduction
Rupture of carotid atherosclerotic plaque is a major cause of ischemic stroke. While US, CT angiography, and digital subtraction angiography are recommended for clinical assessment of atherosclerosis (1,2), MRI takes advantage of good soft-tissue contrast to distinguish vulnerable plaque characteristics, such as intraplaque hemorrhage (IPH), lipid-rich necrotic core (LRNC), fibrous cap, and calcification (3). Patients with the presence of vulnerable plaque characteristics with or without symptoms would benefit from MRI assessment. Traditionally, two-dimensional multicontrast MRI techniques have been widely used (4). However, such techniques result in long scan times, limited section resolution and coverage, and misregistration between multicontrast images. Furthermore, diagnosis using these qualitative techniques relies on signal intensities of the surrounding tissues, which may vary among scans due to coil sensitivity differences, field inhomogeneity, and other imaging system imperfections (5).
In the past 2 decades, techniques for quantitative assessment of carotid atherosclerotic plaques have been proposed (6–9). By directly evaluating intrinsic quantitative values (T1, T2, and T2*), these mapping methods can be used to distinguish plaque components independently of the signal intensity of surrounding tissue, with improved reproducibility. However, traditional quantitative techniques are time consuming, as they require multiple scans for just one parameter, which may lead to image misregistration and postprocessing complications (7,9). To address these challenges, newly developed techniques acquire single or multiple maps in one scan, such as golden angle radial k-space sampling-simultaneous noncontrast angiography and intraplaque hemorrhage (or GOAL-SNAP) (T1 mapping) (5), simultaneous T1 and T2 mapping of carotid plaque (or SIMPLE) (10), and multitasking (T1 and T2 mapping) (11). A study from the past 2 decades shows that iron deposition, which can be quantified by T2* mapping, may play a key role in plaque vulnerability and progression (8). Therefore, comprehensive and efficient evaluation of carotid atherosclerotic plaque requires simultaneous T1, T2, and T2* mapping.
We developed a technique for simultaneous T1, T2, and T2* mapping of carotid plaque (SIMPLE*) with large coverage and isotropic spatial resolution by using interleaved single-echo and multiecho (ME) three-dimensional golden angle radial acquisition. We compared its performance with that of conventional MRI techniques in both a phantom study and in vivo study.
Materials and Methods
Sequence Design
MRI sequence design was implemented on the Philips Advanced Research and Development Integrated Sequence Programming Environment (Yajie Wang, with 6 years of experience in MRI). The SIMPLE* sequence was composed of three modules with different T2 inversion-recovery preparation pulses and acquisition schemas (Fig 1). The T2 inversion-recovery preparation pulse consisted of the adiabatic T2 preparation (T2prep) pulse and inversion-recovery pulse and was used to generate T2 and T1 contrast. For the first module with T2prep = 0 and inversion recovery, single-echo data were first acquired, followed by ME data. For the second and the third modules with T2prep ≠ 0 and inversion recovery, only single-echo data were acquired. Fat suppression was achieved by using a water excitation pulse. The acquisition trajectories were designed to satisfy the three-dimensional golden angle strategy (12) (Appendix S1). Typical scan parameters of the SIMPLE* sequence are shown in Table 1.
Figure 1: Diagram of the proposed simultaneous T1, T2, and T2* mapping of carotid plaque (SIMPLE*) sequence. The SIMPLE* sequence is composed of three modules with different T2 preparation (T2prep) pulses, inversion-recovery (IR) pulses, and acquisition schemas. For the first module with T2 preparation = 0 and inversion recovery, single-echo data are first acquired, and then multiecho data are obtained. For the second and the third modules with T2 preparation ≠ 0 and inversion recovery, only single-echo data are acquired. The acquisition trajectories (S1–S6) (Appendix S1) are designed to satisfy the three-dimensional (3D) golden angle strategy. Tex = time between the last gradient-echo acquisition and the next T2 preparation pulse, Tgap = time between the inversion-recovery pulse and the first excitation pulse.
 |
Image Reconstruction and T1, T2, T2* Fitting
Sliding window reconstruction with a temporal width of 25 spokes was applied for the single-echo data to generate T1- and T2-weighted compound images. ME data were used to reconstruct the T2*-weighted images. Low-rank modeling and sparsity constraint reconstruction was then performed to further improve image quality (5,13).
The vessel wall single-echo image was reconstructed using the data obtained around the nulling point of the blood signal. T1 and T2 were estimated using the T1- and T2-weighted compound images by combining two extra spoiled gradient-recalled echo images with different flip angles with use of the Bloch equation (10). For the T2*-weighted images reconstructed from the SIMPLE* sequence, a monoexponential decay curve was fitted to estimate T2*.
Phantom Study
All scans were acquired with a 3.0-T MRI scanner (Ingenia CX; Philips Healthcare). The phantom study was performed with nine tubes filled with different concentrations of agarose and gadolinium-diethylenetriamine pentaacetic acid to create different T1, T2, and T2* values. Apart from the proposed sequence, traditional two-dimensional inversion-recovery spin echo (SE) with different inversion times; ME-SE; and ME gradient echo were used as the reference standard sequences for T1, T2, and T2* quantification of the phantom, respectively (Appendix S1).
In Vivo Study
A prospective, in vivo study was approved by the institutional review board, and all participants provided written informed consent. Between November 2020 and February 2022, healthy volunteers and participants with carotid atherosclerosis clinically diagnosed by means of US (1) or MRI (14) were recruited. Exclusion criteria included any contraindication to MRI (eg, claustrophobia, metal implants). Participants whose MRI scans had severe motion artifacts were also excluded. All scans were acquired using a dedicated eight-channel carotid coil on a 3.0-T MRI scanner (Ingenia CX; Philips Healthcare).
In healthy volunteers, the proposed sequence was compared with two-dimensional modified Look-Locker inversion recovery, or MOLLI; ME turbo SE; and ME turbo field echo in T1, T2, and T2* measurements. For each volunteer, elliptical regions of interest were drawn in the sternocleidomastoid muscle on both sides, and the T1, T2, and T2* within each muscle region of interest were averaged.
For participants with carotid atherosclerosis, scans with the SIMPLE* sequence and conventional multicontrast MRI scans including time-of-flight, black-blood T1- and T2-weighted turbo SE (14), simultaneous noncontrast angiography and intraplaque hemorrhage (or SNAP) (15), and ME gradient echo (8) were acquired. Conventional multicontrast plaque MRI scans were reviewed by an experienced radiologist (X.L., with 15 years of experience in MRI) to localize the plaque and identify different plaque components, including IPH, LRNC, fibrous tissue, and calcification (4,14). The reconstructed SIMPLE* images were reformatted to 0.8-mm section thickness and 0.4 × 0.4 mm2 in-plane resolution for quantitative analysis. Plaque components identified on conventional multicontrast MRI scans were used as references to guide the delineation of components on SIMPLE* vessel wall images. Calcification had a low signal-to-noise ratio on contrast images; thus, its T1, T2, and T2* values could not be reliably estimated.
Nine of the recruited participants with carotid plaque underwent a repeat examination on a separate day to test the scan-rescan reproducibility of the proposed SIMPLE* sequence. Blinded to the quantitative results, an author (Yajie Wang) matched the first and the second scans based on the reconstructed T1- and T2-weighted compound images, with use of the carotid bifurcation as reference.
Statistical Analysis
In the phantom study, the mean T1, T2, and T2* values of each tube calculated by the SIMPLE* sequence were compared with those calculated by the reference standard sequences with use of the Pearson or Spearman correlation test, as appropriate. In healthy volunteers, the Student t test or Wilcoxon rank-sum test was used to compare the quantitative muscle parameters. For participants with carotid plaque, T1, T2, and T2* among the IPH, LRNC, and fibrous tissue estimated with use of SIMPLE* were compared using one-way analysis of variance or the Kruskal-Wallis test with post hoc tests. In the reproducibility study, the T1, T2, and T2* measurements from the two scans on matched axial sections were compared using the Student t test or Wilcoxon rank-sum test and intraclass correlation coefficient.
P < .05 was considered indicative of a statistically significant difference. All statistical analyses were performed in SPSS (version 23.0; IBM).
Results
Phantom Study
The proposed SIMPLE* sequence showed similar T1, T2, and T2* distributions of the phantom compared with reference sequences (Fig 2). Due to the T2*-shortening effect at the junction of the tube and the surrounding air, only the uniform central regions were averaged for the T2* comparison. The mean T1, T2, and T2* values within each tube estimated with the proposed SIMPLE* sequence showed good agreements with inversion-recovery SE (correlation coefficient [r] = 0.99, P < .001), ME-SE (r = 0.99, P < .001), and ME gradient echo (r = 0.99, P < .001). The mean relative errors in T1, T2, and T2* estimations were 0.5% ± 2.8 (SD), 3.4% ± 4.3, and 4.0% ± 4.7, respectively, compared with the reference standard sequences.
Figure 2: Quantitative mapping results of the phantom study. (A) T1 maps estimated with an inversion-recovery spin-echo (IR-SE) sequence and the proposed simultaneous T1, T2, and T2* mapping of carotid plaque (SIMPLE*) sequence with quantitative comparison. (B) T2 maps estimated with a multiecho spin-echo (ME-SE) sequence and the proposed SIMPLE* sequence with quantitative comparison. (C) T2* maps estimated with a multiecho gradient-echo (ME-GRE) sequence and the proposed SIMPLE* sequence with quantitative comparison.
In Vivo Study
Finally, MRI scans in the five volunteers (mean age, 25 years ± 3 [SD]; three women) and 18 participants with carotid plaque (mean age, 65 years ± 9; 16 men) (Table 2) were used for analysis.
 |
In healthy volunteers, we found no evidence of differences in the T1, T2, and T2* estimates of the muscle between SIMPLE* and modified Look-Locker inversion-recovery (n = 10; 1151 msec ± 71 vs 1098 msec ± 64; P = .14), ME turbo SE (n = 10; 31 msec ± 1 vs 31 msec ± 1; P = .32), and ME turbo field-echo (n = 10; 24 msec ± 2 vs 25 msec ± 2; P = .19) sequences, respectively. Example quantitative mapping results for one healthy volunteer (26-year-old woman) are shown in Figure S1. Similar quantitative parameter distributions were found between the proposed sequence and the compared sequences. The proposed sequence also showed clear delineation of the vessel wall on the reconstructed T1- and T2-weighted compound images in healthy volunteers (Fig S1D).
In the 18 participants with carotid plaque, 109 sections from 19 plaques were analyzed. IPH was found in seven plaques, LRNC was identified in seven plaques, and the remaining five plaques were fibrotic. Figure 3 shows comparisons of T1, T2, and T2* values of different plaque components estimated with use of SIMPLE*. IPH (641 msec ± 179) had lower T1 than LRNC (1181 msec ± 91, P < .001) and fibrous tissue (1316 msec ± 196, P < .001). LRNC (36 msec ± 4) showed the lowest T2 value among the three plaque components (IPH, 38 msec ± 4 [P = .03]; fibrous tissue, 40 msec ± 6 [P < .001]). IPH (18 msec ± 5) had lower T2* values than LRNC (24 msec ± 4, P < .001) and fibrous tissue (24 msec ± 8, P = .001). Examples of conventional multicontrast MRI scans and SIMPLE* quantitative images in a 74-year-old man and a 57-year-old man are shown in Figures 4 and 5, respectively.
![Bar graphs show comparisons of (A) T1, (B) T2, and (C) T2* values of different plaque components (intraplaque hemorrhage [IPH], lipid-rich necrotic core [LRNC], and fibrous tissue) estimated with the proposed simultaneous T1, T2, and T2* mapping of carotid plaque, or SIMPLE*, sequence. The error bars represent the SDs.](/cms/10.1148/radiol.222061/asset/images/medium/radiol.222061.fig3.gif "Bar graphs show comparisons of (A) T1, (B) T2, and (C) T2* values of different plaque components (intraplaque hemorrhage [IPH], lipid-rich necrotic core [LRNC], and fibrous tissue) estimated with the proposed simultaneous T1, T2, and T2* mapping of carotid plaque, or SIMPLE*, sequence. The error bars represent the SDs.")
Figure 3: Bar graphs show comparisons of (A) T1, (B) T2, and (C) T2* values of different plaque components (intraplaque hemorrhage [IPH], lipid-rich necrotic core [LRNC], and fibrous tissue) estimated with the proposed simultaneous T1, T2, and T2* mapping of carotid plaque, or SIMPLE*, sequence. The error bars represent the SDs.
Figure 4: Conventional axial multicontrast MRI scans acquired with the simultaneous noncontrast angiography and intraplaque hemorrhage (SNAP) sequence, T1- and T2-weighted turbo spin-echo (TSE) sequences, and multiecho gradient-echo (ME-GRE) sequence, along with a vessel wall image and T1, T2, and T2* MRI quantitative mapping results from the proposed simultaneous T1, T2, and T2* mapping of carotid plaque , or SIMPLE*, sequence of one carotid plaque with intraplaque hemorrhage (arrow) in a 74-year-old man.
Figure 5: Conventional axial multicontrast MRI scans acquired with the simultaneous noncontrast angiography and intraplaque hemorrhage (SNAP) sequence, T1- and T2-weighted turbo spin-echo (TSE) sequences, and multiecho gradient-echo (ME-GRE) sequence, along with a vessel wall image and T1, T2, and T2* quantitative mapping results from the proposed simultaneous T1, T2, and T2* mapping of carotid plaque, or SIMPLE*, sequence of one carotid plaque with a lipid-rich necrotic core (arrow) in a 57-year-old man.
For the nine participants with repeat scans, 57 pairs of matched axial sections were identified and analyzed. We found no evidence of differences in the T1 (P = .91), T2 (P = .48), and T2* (P = .58) values of each paired section between the first and the second scans. The intraclass correlation coefficients for T1, T2, and T2* measurements of the two scans were 0.99, 0.87, and 0.96, respectively.
Discussion
Quantifying T1, T2, and T2* values of carotid atherosclerotic plaque is essential for comprehensive evaluation. However, previous techniques estimated T1, T2, and T2* separately. Particularly, T1 measurements relied on multiple sequences (16,17). Thus, misregistration between multiple sequences was a major challenge. Consequently, multiparameter techniques, which have the advantage of producing coregistered parametric maps, are gaining interest (5,11,18–20). In our in vivo study, important high-risk plaque components, including intraplaque hemorrhage and lipid-rich necrotic core, were detected and showed consistent features between the quantitative maps and the multicontrast images.
Our quantitative comparison among different plaque components showed comparable distributions with those in previous studies (21,22). In the study by Qiao et al (21), IPH also showed a lower T1 value than other plaque components. Consistent with our study, Biasiolli et al (22) found that LRNC had a lower T2 value than fibrous tissue and IPH; however, the estimated T2 value of IPH differed from our findings (107 msec vs 38 msec). This may be caused by the limited number of plaques with IPH in their study (n = 1) (22). The T2* value of carotid plaque reported in our study was comparable with that in a previous study (8). Our results indicate the potential of the proposed SIMPLE* sequence in plaque component diagnosis and identification.
The SIMPLE* technique has the potential to achieve highly undersampled image reconstruction by using the three-dimensional golden angle strategy (12). In the future, deep learning–based methods (23) can be applied to extremely improve the acquisition efficiency. In addition, as inspired by another study (24), self-motion compensation can be applied to detect and correct abrupt motion during image acquisition in the future. Furthermore, additional in vivo studies are needed to further investigate the clinical feasibility of the proposed sequence.
Our study had some limitations. First, the scan time of the proposed sequence was 10 minutes in our preliminary experience, which may still be too long for extensive clinical application. Second, the number of the recruited participants with carotid plaque was limited, resulting in a small sample size.
In conclusion, a quantitative technique using interleaved single- and multiecho three-dimensional golden angle radial acquisition for simultaneous T1, T2, and T2* MRI mapping of carotid plaque was developed, which demonstrated high accuracy and in vivo feasibility.
Author Contributions
Author contributions: Guarantors of integrity of entire study, J.W., D.L., H.Z., H.C.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; agrees to ensure any questions related to the work are appropriately resolved, all authors; literature research, Yajie Wang, X.L., D.L., J.L., H.Z., H. Qiao, H.C.; clinical studies, Yajie Wang, X.L., J.W., X.K., D.L., J.L., H.Z., F.X., L.Z., X.F., X.Z.; experimental studies, Yajie Wang, X.L., H. Qi, D.L., J.L., H.Z., R.G., H. Qiao, Z.C., D.S.; statistical analysis, Yajie Wang, X.L., D.L., J.L., H.Z.; and manuscript editing, Yajie Wang, X.L., Yishi Wang, D.L., J.L., H.Z., H.C.
* Yajie Wang and X.L. contributed equally to this work.
Supported by the Beijing Municipal Natural Science Foundation (grant Z190024), National Natural Science Foundation of China (grants 81930119, 81371540, and 81571667), and Natural Science Foundation of Hubei Province (grant 2021CFB442).
Data sharing statement: Data generated or analyzed during the study are available from the corresponding author by request.
References
- 1. . Recommendations for the assessment of carotid arterial plaque by ultrasound for the characterization of atherosclerosis and evaluation of cardiovascular risk: from the American Society of Echocardiography. J Am Soc Echocardiogr 2020;33(8):917–933.
- 2. . Imaging modalities to diagnose carotid artery stenosis: progress and prospect. Biomed Eng Online 2019;18(1):66.
- 3. . Roadmap consensus on carotid artery plaque imaging and impact on therapy strategies and guidelines: an international, multispecialty, expert review and position statement. AJNR Am J Neuroradiol 2021;42(9):1566–1575.
- 4. . Carotid atherosclerotic plaque: noninvasive MR characterization and identification of vulnerable lesions. Radiology 2001;221(2):285–299.
- 5. . Highly undersampled kooshball reconstruction with low-rank modeling and sparsity constraints for high-resolution T1 mapping. In:
Proceedings of the 26th Annual Meeting of ISMRM , Paris, France, 2018; 0248. https://cds.ismrm.org/protected/18MProceedings/PDFfiles/0248.html. - 6. . Quantification of lipid-rich core in carotid atherosclerosis using magnetic resonance T2 mapping: relation to clinical presentation. JACC Cardiovasc Imaging 2017;10(7):747–756.
- 7. . Three-dimensional quantitative T1 and T2 mapping of the carotid artery: sequence design and in vivo feasibility. Magn Reson Med 2016;75(3):1008–1017.
- 8. . In vivo atherosclerotic plaque characterization using magnetic susceptibility distinguishes symptom-producing plaques. JACC Cardiovasc Imaging 2008;1(1):49–57.
- 9. . Quantitative T1, T2 and T2* mapping of carotid artery normal wall and atherosclerotic plaque. In:
Proceedings of the 26th Annual Meeting of ISMRM , 2011; 3311. https://archive.ismrm.org/2011/3311.html. - 10. . Simultaneous T1 and T2 mapping of the carotid plaque (SIMPLE) with T2 and inversion recovery prepared 3D radial imaging. Magn Reson Med 2018;80(6):2598–2608.
- 11. . Quantitative Multi-contrast Atherosclerosis Characterization (qMATCH): initial study of plaque analysis. In: Proceedings of the SCMR. Orlando, FL, 2020; 749492. https://www.eventscribe.com/2020/SCMR/fsPopup.asp?efp=SkJBUE9PSFI2NzY0&PosterID=253095&rnd=0.3612675&mode=posterinfo.
- 12. . Temporal stability of adaptive 3D radial MRI using multidimensional golden means. Magn Reson Med 2009;61(2):354–363.
- 13. . Low-rank plus sparse matrix decomposition for accelerated dynamic MRI with separation of background and dynamic components. Magn Reson Med 2015;73(3):1125–1136.
- 14. . Classification of human carotid atherosclerotic lesions with in vivo multicontrast magnetic resonance imaging. Circulation 2002;106(11):1368–1373.
- 15. . Simultaneous noncontrast angiography and intraplaque hemorrhage (SNAP) imaging for carotid atherosclerotic disease evaluation. Magn Reson Med 2013;69(2):337–345.
- 16. . Relationship between the carotid plaque T1 relaxation time and the plaque-to-muscle signal intensity ratio on black-blood magnetic resonance imaging scans. J Stroke Cerebrovasc Dis 2016;25(11):2580–2584.
- 17. . Quantitative characterization of carotid plaque components using MR apparent diffusion coefficients and longitudinal relaxation rates at 3T: a comparison with histology. J Magn Reson Imaging 2018;48(6):1657–1667.
- 18. . Novel method for rapid, simultaneous T1, T2*, and proton density quantification. Magn Reson Med 2007;57(3):528–537.
- 19. . Simultaneous three-dimensional myocardial T1 and T2 mapping in one breath hold with 3D-QALAS. J Cardiovasc Magn Reson 2014;16(1):102.
- 20. . Magnetic resonance fingerprinting. Nature 2013;495(7440):187–192.
- 21. . Quantitative evaluation of carotid atherosclerotic vulnerable plaques using in vivo T1 mapping cardiovascular magnetic resonaonce: validation by histology. J Cardiovasc Magn Reson 2020;22(1):38.
- 22. . In-vivo quantitative T2 mapping of carotid arteries in atherosclerotic patients: segmentation and T2 measurement of plaque components. J Cardiovasc Magn Reson 2013;15(1):69.
- 23. . Deep learning-enhanced T1 mapping with spatial-temporal and physical constraint. Magn Reson Med 2021;86(3):1647–1661.
- 24. . Free-running 3D whole heart myocardial T1 mapping with isotropic spatial resolution. Magn Reson Med 2019;82(4):1331–1342.
Article History
Received: Aug 14 2022Revision requested: Oct 5 2022
Revision received: Dec 9 2022
Accepted: Dec 22 2022
Published online: Feb 28 2023
