Improved myocardial T1 mapping technique to eliminate device-induced image artefacts for patients with implanted cardiac devices
Jiaxin Shao1, Shams Rashid1, Kim-Lien Nguyen2,3, and Peng Hu1,4

1UCLADepartment of Radiological Sciences, David Geffen School of Medicine, University of California, Los Angeles, CA, United States, 2Department of Medicine, Division of Cardiology, David Geffen School of Medicine, University of California, Los Angeles, CA, United States, 3Division of Cardiology, Veterans Affairs Greater Los Angeles Healthcare System, Los Angeles, CA, United States, 4Biomedical Physics Inter-Departmental Graduate Program, University of California, Los Angeles, CA, United States

Synopsis

Current cardiac T1 mapping techniques, including the modified Look-Locker inversion-recovery (MOLLI), cannot be used effectively in patients with implanted cardiac devices due to large off-resonance induced by the device. To eliminate the device-induced image artefacts, we developed a technique by modifying the MOLLI sequence to use spoiled gradient echo readout and a wideband inversion pulse, with a new acquisition scheme and T1 estimation algorithm. The feasibility of our new technique was tested in phantom studies and validated in eight healthy volunteers and ten patients with implanted cardiac devices at 1.5 Tesla.

Background

With recent development of a wideband LGE CMR sequence (1), detection of focal fibrosis is feasible in patients with implanted cardiac devices, such as implantable cardioverter defibrillators (ICDs). One main drawback of LGE is its limited ability to detect diffuse fibrosis. Many patients with cardiac devices have non-ischemic cardiomyopathy resulting from conditions associated with diffuse fibrosis. Therefore, myocardial T1 mapping techniques (2–4) may help further diagnose and monitor the disease processes in these patients. Unfortunately, current widely used cardiac T1 mapping techniques, including the modified Look-Locker inversion-recovery (MOLLI) (5) sequence, cannot be used to image patients with ICDs. The large off-resonance induced by the device and associated banding artifacts when using bSSFP readouts limit the reliability and accuracy of existing techniques. Accordingly, we sought to develop and validate an improved myocardial T1 mapping technique that would mitigate the device-induced image artefacts in patients with ICDs.

Methods

A MOLLI-based pulse sequence, named Wideband-FLASH-MOLLI, was developed by incorporating a fast low angle shot (FLASH) readout (6) and a wideband inversion pulse that was previously developed for wideband LGE (1). Figure 1 demonstrates the acquisition scheme of the proposed Wideband-FLASH-MOLLI and the bandwidth of the wideband inversion pulse compared to the traditional inversion pulse used in MOLLI. With Wideband-FLASH-MOLLI, ten FLASH images were acquired over 13 heart beats during a single breath-hold with a resolution of 1.9×2.5×8.0 mm3. The Bloch equation simulation with slice profile correction (BLESSPC) T1 estimation algorithm, which was originally developed for FLASH readout at 3.0T (6), was used in our current technique to reconstruct the Wideband-FLASH-MOLLI T1 maps. The bSSFP-MOLLI with InSiL fitting (7) and FLASH-MOLLI with BLESSPC fitting (6) was implemented for comparisons, both of which uses the conventional inversion pulse and a 3-(3)-3-(3)-5 acquisition scheme.

Wideband-FLASH-MOLLI was evaluated using eight phantoms and validated in eight healthy volunteers and ten patients with ICDs using a 1.5T MR scanner (Avanto, Siemens Healthcare; Erlangen, Germany). The effects of off-resonance frequency variation, heart rate variation, and presence of ICD on T1 estimation accuracy were investigated using phantom studies. To mimic device-induced image artifacts, an ICD generator was attached to the body coil and close to the left shoulder of healthy volunteers. The bSSFP-MOLLI and FLASH-MOLLI images were acquired in all healthy volunteers before and after the ICD attachment and in three patients with ICDs for comparisons.

The coefficient of determination (R2) was used to determine the quality of the fit for each pixel. T1 values with R2 < 0.95 were set to zero in T1 maps.

Results

Wideband-FLASH-MOLLI generated consistent T1 values over a wide range of off-resonance frequencies (Figure 2) and showed no dependence on heart rate variation. The maximum T1 estimation errors using wideband-FLASH-MOLLI with and without an ICD were less than 3% for T1 range of 212ms - 1673ms. For all eight healthy volunteers without ICD, the average myocardial T1 values in septal region measured using Wideband-FLASH-MOLLI was close to that measured using bSSFP-MOLLI with InSiL fitting (1173±24 ms vs. 1175±30 ms, p=0.8). No statistically significant difference in T1 values was found between four different myocardial regions (septal, anterior, lateral and inferior) of left ventricular (LV) using Wideband-FLASH-MOLLI (p>0.2). After ICD attachment in healthy volunteers, the Wideband-FLASH-MOLLI T1 values were not significantly changed in any of the four myocardial regions (p>0.2).

Due to the presence of an ICD, the magnitude images acquired using bSSFP-MOLLI and FLASH-MOLLI showed hyper-intensity and dark band artifacts within the myocardium (Figure 3-4). The dark bands in the bSSFP-MOLLI are expected off-resonance banding related to bSSFP readouts. The dark bands in the FLASH-MOLLI images occurred at the boundary of inverted and non-inverted regions, presumably due to signal cancelations within the boundary pixels. In contrast, device-induced artifacts were eliminated within the myocardial regions of all healthy volunteers and patients with ICDs when using Wideband-FLASH-MOLLI.

Figure 5 shows the pre- and post-contrast Wideband-FLASH-MOLLI T1 maps and the wideband LGE image acquired in a patient with an ICD, who was referred for assessment of cardiac amyloidosis. The average native T1 value within the entire LV myocardium was 1336±86.6 ms, which is ~14% greater than the average myocardial T1 values in healthy volunteers. This finding is concordant with those reported in the literature for cases of amyloidosis.

Conclusion

This study demonstrates the feasibility of using Wideband-FLASH-MOLLI to eliminate device-induced image artifacts in patients with implanted cardiac devices and can be used for accurate myocardial T1 mapping. Wideband-FLASH-MOLLI enables assessment of diffuse fibrosis in patients with implanted cardiac devices.

Acknowledgements

None

References

1. Rashid S, Rapacchi S, Vaseghi M, et al.: Improved late gadolinium enhancement MR imaging for patients with implanted cardiac devices. Radiology 2014; 270:269–74.

2. Burt JR, Zimmerman SL, Kamel IR, Halushka M, Bluemke DA: Myocardial T1 mapping: techniques and potential applications. Radiographics 2014; 34:377–95.

3. Iles LM, Ellims AH, Llewellyn H, et al.: Histological validation of cardiac magnetic resonance analysis of regional and diffuse interstitial myocardial fibrosis. Eur Heart J Cardiovasc Imaging 2015; 16:14–22.

4. Karamitsos TD, Piechnik SK, Banypersad SM, et al.: Noncontrast T1 mapping for the diagnosis of cardiac amyloidosis. JACC Cardiovasc Imaging 2013; 6:488–97.

5. Messroghli DR, Radjenovic A, Kozerke S, Higgins DM, Sivananthan MU, Ridgway JP: Modified Look-Locker inversion recovery (MOLLI) for high-resolution T1 mapping of the heart. Magn Reson Med 2004; 52:141–6.

6. Shao J, Rapacchi S, Nguyen K-L, Hu P: Myocardial T1 mapping at 3.0 tesla using an inversion recovery spoiled gradient echo readout and bloch equation simulation with slice profile correction (BLESSPC) T1 estimation algorithm. J Magn Reson Imaging 2015. doi: 10.1002/jmri.24999.

7. Shao J, Nguyen K-L, Natsuaki Y, Spottiswoode B, Hu P: Instantaneous signal loss simulation (InSiL): an improved algorithm for myocardial T1 mapping using the MOLLI sequence. J Magn Reson Imaging 2015; 41:721–9.

8. Rashid S, Rapacchi S, Shivkumar K, Plotnik A, Finn JP, Hu P: Modified wideband three-dimensional late gadolinium enhancement MRI for patients with implantable cardiac devices. Magn Reson Med 2015. doi: 10.1002/mrm.25601.

Figures

Figure 1: (a) Acquisition scheme of Wideband-FLASH-MOLLI. The M0 weighted image is acquired first without inversion, followed by three inversion groups, acquiring 5, 3 and 1 image respectively. A wideband inversion pulse proposed by Rashid et al. (1) is used. (b) A simulated longitudinal magnetization profile of the conventional inversion pulse and the applied wideband inversion pulse are shown to compare the difference in their bandwidths (8).

Figure 2: Calculated Wideband-FLASH-MOLLI T1 values as a function of center frequency shift. Wideband-FLASH-MOLLI produced consistent T1 estimation for a range of ±1.7 kHz. In all eight phantoms, the measured values were within 4% of the reference T1 values (Ref T1) measured using standard inversion recovery spin-echo sequence.

Figure 3: Comparative examples of magnitude images and corresponding T1 maps acquired using the bSSFP-MOLLI (column 1), FLASH-MOLLI (columns 2) and Wideband-FLASH-MOLLI (columns 3) in a healthy volunteer. Images were acquired in the mid left ventricular short axis plane without and with an ICD taped near the volunteer’s left shoulder. Compared to bSSFP-MOLLI and FLASH-MOLLI, Wideband-FLASH-MOLLI effectively eliminated the device-induced artifacts within the myocardium.

Figure 4: Comparative examples of native magnitude images and T1 maps acquired using bSSFP-MOLLI, FLASH-MOLLI and wideband-FLASH-MOLLI in a patient with an ICD. Wideband-FLASH-MOLLI was effective in mitigating the device-induced artifacts. Although mild residual artifacts remain present in the anterior wall of the Wideband-FLASH-MOLLI magnitude image, the degree of contamination is reduced in the reconstructed T1map.

Figure 5: Pre and post-contrast Wideband-FLASH-MOLLI T1 maps and wideband LGE image acquired in a patient with an ICD and suspected cardiac amyloidosis. Images were acquired in mid left ventricular short axis and in the same slice position. The average T1 value of the entire myocardium was 1336±86.6 ms, which is much longer than normal T1 values measured in healthy volunteers.



Proc. Intl. Soc. Mag. Reson. Med. 24 (2016)
0589