A subjective and objective comparison of tissue contrast and imaging artifacts present in routine spin echoes and in iterative decomposition of asymmetric spin echoes for soft tissue neck MRI

Abstract

Objective: FSE sequences play key roles in neck MRI despite the susceptibility issues in neck region. Iterative decomposition of asymmetric echoes (IDEAL, GE) is a promising method that separates fat and water images resulting in high SNR and improved fat suppression. We tested how neck tissue contrasts, image artifacts and fat separation as opposed to fat suppression in terms of image quality compare between routine and IDEAL FSE.

Methods: IDEAL based and routine T₁ and T₂-weighted FSE sequences were applied for neck MRI at 1.5T and 3T. Overall image quality including fat suppression, tissue contrast, image artifacts and lesion conspicuity were subjectively assessed for 20 patients clinically indicated for neck MRI. Quantitative tissue contrast estimates from parotid area were compared between IDEAL and routine FSE for 7 patients. Four patients with oncocytoma were also reviewed to assess benefits of separately reconstructed fat specific image sets.

Results: Subjective tissue contrast and overall image quality including image sharpness, fat suppression and image artifacts were superior for IDEAL sequences. For oncocytoma fat specific IDEAL images provided additional information. Objective CNR estimates from a central slice were equivalent for IDEAL and routine FSE at both field strengths.

Conclusions: We demonstrated that high SNR inherent in IDEAL FSE consistently translates into high tissue contrast with image quality advantages in neck anatomy where large susceptibility variation and physiological motions reduce image quality for conventional FSE T₁ and T₂. However, the objective contrast estimates for parotid gland at isocenter were statistically equivalent for IDEAL and conventional FSE perhaps because at or near isocenter routine FSE works well. Additionally, fat specific IDEAL image sets add to diagnostic specificity for fat deficient lesions.

Keywords: Fat suppression; IDEAL; Image quality; Soft tissue neck; Susceptibility; Tissue contrast.

Figure 1.

ROI placement is shown on selected central slice to compute fat/parotid CNR. The water only T₂ weighted IDEAL FSE image (panel A) is compared with T₂ weighted FSEIR (panel B).

Figure 2.

CNR from a central slice at 3T, at 1.5T. The mean CNR values for buccal fat/parotid are plotted from IDEAL and non-IDEAL FSE, ROI placement is shown in Figure 1.

Figure 3.

(A) Non-fat suppressed FSE T₁ axial image, and (B) fat+water combined IDEAL based in-phase T₁ image (IDEAL version of non-fat suppressed) image. IDEAL FSE seems to show greater thyroid nodule conspicuity as well as less sensitivity to respiratory/carotid flow induced artifacts compared to routine T₁ (panel A, arrowhead).

Figure 4.

Benefit of preserving fat-selective images instead of discarding fat signal for a biopsy-proven “vanishing parotid” oncocytoma. The lesion is less conspicuous on routine FSE T₁ image (panel A) or in fat-removed, water only IDEAL T₁ image (panel B). However, the water-removed, fat only IDEAL T₁ image (arrow, panel C) clearly reveals a highly fat deficient lesion, in this case, an oncocytoma.

Figure 5.

Benefit of IDEAL based imaging when patient may have hardware. Routine FSE T₁ (Panel A: ETL 3, BW 15.6 kHz) shows artifactual background signal due to metallic braces; IDEAL FSE T₁ (Panel B, arrows, ETL 3, BW 15.6 kHz) shows reduction of artifacts although some reconstruction artifact is still present and Panel C shows mostly artifact-free image for FSE IDEAL T₂ (ETL 20, BW 41.7 kHz).