Related Experiment Video
Updated: Jun 8, 2026
Assessing Cortical Cerebral Microinfarcts on High Resolution MR Images
Published on: November 20, 2015
High-Resolution MRI Using Artificial Intelligence-Assisted Acceleration and Radial Dynamic Contrast Enhancement for
Shanshan Liu1, Xuwen Zhang2, Qiang Fang2
1From the Department of Radiology (S.L.), The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, Jinan, China.
Background And Purpose:
Accurate detection of pituitary microadenomas is critical for the diagnosis and treatment of Cushing disease (CD). However, conventional MRI often has limited resolution and thick slices, leading to missed lesions and suboptimal surgical planning. This study investigates the diagnostic utility of artificial intelligence-assisted compressed sensing (ACS) applied to conventional anatomic MRI, combined with dynamic contrast-enhanced (DCE)-MRI using united compressed sensing with radial acquisition (uCSR), aiming to improve spatial resolution and lesion detection without prolonging scan time, while uCSR enhances temporal resolution and motion robustness in dynamic contrast imaging.
Materials And Methods:
This prospective study included 61 patients with surgically confirmed CD who underwent both conventional and ACS-accelerated MRI sequences, including T2WI, contrast-enhanced T1-weighted imaging (T1WI-C), and delayed FLAIR, along with DCE-MRI using uCSR technique. Image quality assessments and lesion detection rates were compared. Pharmacokinetic parameters (volume transfer constant [Ktrans], rate constant [Kep], and extravascular extracellular volume fraction) derived from DCE were evaluated across lesion types.
Results:
A total of 61 patients (median age, 42 years; 56% women) were included, with 71 lesions identified, including 9 patients with multiple lesions and 2 patients with ectopic lesions. ACS-T1WI-C achieved higher image clarity scores compared with conventional T1WI-C (4.7 ± 0.3 versus 4.1 ± 0.6; P < .001) and higher SNR (30.1 ± 3.4 versus 22.3 ± 2.4; P < .001). Similarly, ACS-T2WI showed higher contrast-to-noise ratio (CNR; 12.4 ± 3.1 versus 8.5 ± 2.3; P < .001). Across all sequences, the combination of ACS-T1WI-C and delayed FLAIR detected all 71 lesions, corresponding to a sensitivity of 94.9% and specificity of 93.5%, significantly higher than conventional sequences (P < .001). Interobserver agreement for lesion detection was excellent (κ = 0.91) for ACS sequences. Multiple lesions (14.7%) showed significant pharmacokinetic differences; adrenocorticotropic hormone-secreting adenomas demonstrated significantly lower Ktrans and Kep compared with Rathke cysts and nonfunctional adenomas (P < .01).
Conclusions:
ACS markedly improves image quality and lesion detection in CD, providing high-resolution imaging without extending acquisition time. uCSR-based DCE-MRI further aids lesion-type differentiation, contributing to more accurate preoperative localization and diagnosis.
