Dynamic Contrast-Enhanced pituitary MRI with higher temporal resolution using deep learning Reconstruction: A...

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Pituitary MRI is crucial for evaluating sellar lesions.
• Conventional dynamic contrast-enhanced imaging (C-DCE) has limitations in temporal resolution and spatial sampling.
• Deep learning-based reconstruction dynamic contrast-enhanced imaging (DLR-DCE) is a novel technique aiming to improve MRI acquisition.

Purpose of the Study:

• To compare the image quality and diagnostic performance of DLR-DCE with C-DCE for pituitary MRI in evaluating sellar lesions.
• To assess the impact of DLR-DCE on scan time, temporal resolution, and spatial sampling.

Main Methods:

• A retrospective study involving 43 patients who underwent both DLR-DCE and C-DCE pituitary MRI.
• DLR-DCE protocol involved six sequential phases with reduced scan time and gap, and increased slices per phase.
• Two neuroradiologists independently evaluated image quality, lesion diagnosis, and cavernous sinus invasion.

Main Results:

• DLR-DCE showed significantly improved delineation of the normal pituitary gland and oculomotor nerve.
• Lesion conspicuity and signal-to-noise ratio (SNR) were significantly higher with DLR-DCE.
• Diagnostic accuracy for cavernous sinus invasion was numerically higher with DLR-DCE, though not statistically significant.
• Overall diagnostic performance for sellar lesions was comparable between DLR-DCE and C-DCE.

Conclusions:

• DLR-DCE pituitary MRI achieves higher temporal resolution and improved spatial sampling compared to C-DCE.
• DLR-DCE enhances image quality and lesion visualization without compromising diagnostic performance in routine clinical practice.
• This technique holds promise for more efficient and effective evaluation of sellar lesions.