Introduction

At present, most lung IVIM studies have been carried out using echo planar imaging (EPI) technique, which has fast imaging speed and high SNR, but due to the characteristics of lung tissue, image quality is frequently deteriorated by susceptibility artifacts. This could lead to lower accuracy of IVIM assessment¹. Turbo spin-echo (TSE) technique has the advantage of negligible distortion but low scan efficiency². GRASE sequence has shown high application value in cholangiopancreatography³, but its application in lung IVIM has not been reported. This study explored the application value of GRASE technique in lung cancer and compared it with the TSE and EPI technique.

Materials and Methods

Patients
A total of 28 patients with lung cancer (17 males and 21 females; age range, 42–84 years; mean age, 62.6 years) were included with written informed consent obtained from each patient. This study was approved by the local IRB.
MRI Experiments
All the MRI examinations were performed on a 3.0T MR scanner (Ingenia CX; Philips Healthcare, Best, The Netherlands) equipped with a body phase array coil. The examination included routine MR scans and IVIM sequences: (1) Coronal single-shot fast spin echo breath-hold T2-weighted imaging (T2WI); (2) Axial fat-suppressed respiratory-triggered T2WI. In order to save the scan time, both IVIM and T2WI use 12 slices, with coronal T2WI as the positioning image to ensure that the largest layer of the lesion was in the center of the IVIM scan. Table 1 shows the detailed scan parameters of the IVIM sequences.
Data analysis
The quantitative assessments of the image quality of all 28 patients were analyzed by 2 radiologists with 3 and 6 years of experience for MR diagnosis, respectively. Image distortions of the lesions in each sequence were quantitatively compared using the maximum displacement in the phase-encoding direction of the lesion between T2WI/SPIR and IVIM by superimposing IVIM on T2WI(with b = 800 s/mm² ), signal-to-noise ratio(SNR), and contrast-to-noise ratio (CNR) from TSE-IVIM ,EPI-IVIM and GRASE-IVIM (with b = 800 s/mm² ) were measured. 3 circular regions of interest (ROIs, ≥70 pixels per ROI) were delineated by 2 experienced readers independently and placed in lesions, erector spinae muscle, and the lung for the measurement of the SNR and CNR.One ROI in the lung was put right next to the lesion to assess local noise. Mean signal intensity (SI) and the standard deviation (SD) were measured for each ROI and the results from the two readers were averaged. SNR and CNR were calculated using Eq. (1–2)²:

(1) $$$\mathrm{SNR}=\mathrm{SI}_{\text {lesion }} / \mathrm{SD}_{\text {lung }} $$$
(2) $$$\mathrm{CNR}=\left(\mathrm{SI}_\text {lesion }-\mathrm{SI}_{\text {muscle }}\right) / \sqrt{{\mathrm{SD}_{\text {lesion }}}^{2}+{\mathrm{SD}_{\text {muscle}}}^{2}}$$$

Where SI_lesion and SI_muscle are the mean SI in the lesion and muscle; SD_lesion, SD_lung, and SD_muscle are the standard deviations of the SI in the lesion, lung, and muscle, respectively.
ADC maps were generated automatically inline by the scanner. IVIM paraments were processed offline with an open source medical imaging interaction toolkit (MITK Diffusion, v. 2018; DKFZ, Heidelberg, Germany).
Statistical analyses were conducted using SPSS (Version 26; IBM Corporation) and MedCalc software (Version 19; Mariakerke, Belgium). ADC value, IVIM parameters, SNR and CNR from TSE-IVIM, GRASE-IVIM and EPI-IVIM were compared using ANOVA, and the distortion level were compared between each pair of sequences with Wilcoxon test (P value < 0.05 was considered to be statistically significant for all the analyses).

Results

The imaging efficiency of GRASE-IVIM and TSE-IVIM was 100%, and the imaging efficiency of EPI-IVIM was 89% (3 cases of lesions could not be analyzed on EPI-IVIM because of severe geometric deformation). In terms of SNR, EPI-IVIM>TSE-IVIM>GRASE-IVIM, pairwise comparisons showed significant differences (P<0.05)(Figure 2); CNR was significantly lower for EPI-IVIM compared with GRASE-IVIM and TST-IVIM (P<0.001), whereas the CNR of GRASE-IVIM was higher than that of TSE-IVIM, but the difference was not significant (P>0.05)(Figure 2).The image distortions of lesions in GRASE-IVIM and TSE-IVIM were significantly lower than that of EPI-IVIM (P<0.001)(Figure 2) . In terms of IVIM and ADC parameters, the ADC and D value of TSE-IVIM were higher than those of EPI-IVIM and GRASE-IVIM (P<0.001), The f value of TSE-IVIM were higher than those of EPI-IVIM in the lesions(P<0.05)(table 2).

Discussion

The SS-GRASE technique can be viewed as a hybrid of TSE and EPI⁴. Our data showed that GRASE-IVIM could reduce the lesion deformation artifacts (Figure 1) and had a higher CNR, while the SNR was lower than EPI-IVIM and TSE-IVIM. ADC, D and f values of IVIM from the three techniques showed statistically significant differences, thus ADC and IVIM parameters from different techniques should be interpreted with great caution.

Conclusion

Our results showed that GRASE technology is a useful technique for IVIM imaging of pulmonary space-occupying lesions. It has fast imaging speed and small lesion distortions. The IVIM parameters using different techniques should be interpreted with caution.

Acknowledgements

We thank Yishi Wang from Philips Healthcare for this valuable support on the IVIM sequence.