Dynamic contrast enhanced MRI of the head and neck region using a VIBE sequence with Cartesian undersampling and...

Area of Science:

• Radiology and Medical Imaging
• Magnetic Resonance Imaging
• Image Reconstruction Techniques

Background:

• Compressed sensing (CS) is crucial for reconstructing images from sparse k-space data, especially in dynamic contrast-enhanced MRI (DCE-MRI).
• Head and neck MRI is essential for diagnosing various conditions, but image quality and acquisition speed can be limiting factors.

Purpose of the Study:

• To evaluate the diagnostic performance of a 3D T1-weighted spoiled gradient-echo sequence using variable density Cartesian undersampling and compressed sensing (CS-VIBE) for head and neck MRI.
• To compare the image quality and diagnostic value of CS-VIBE against standard Cartesian VIBE in head and neck DCE-MRI.

Main Methods:

• Seventy-one patients underwent 3 Tesla DCE-MRI using CS-VIBE with a temporal resolution of 3.4 seconds and 1 mm slice thickness.
• Image quality, lesion conspicuity, fat saturation, and artifacts were assessed by three experienced readers using a 5-point Likert scale.
• Comparison was made with standard Cartesian VIBE, and DCE-derived time intensity curve (TIC) types were determined.

Main Results:

• CS-VIBE showed significantly improved vessel contour, lesion conspicuity, fat saturation, and reduced motion artifacts compared to standard VIBE.
• While overall image quality scores were similar, standard VIBE demonstrated better delineation of pharyngeal mucosa.
• Lesion size measurements were consistent across both methods, and TIC assessment exhibited good interobserver agreement.

Conclusions:

• CS-VIBE with variable density Cartesian undersampling is a valuable technique for head and neck DCE-MRI.
• The method achieves diagnostic, high image quality and high temporal resolution, facilitating accurate diagnosis.
• CS-VIBE offers a promising approach for accelerated and improved head and neck imaging.