Deep learning imrproves MRI signal

VRU 2023 64(5): 873-880

Background: Magnetic resonance imaging (MRI) is a diagnostic imaging modality for brain diseases, but it requires long scan time and optimal image quality. Deep learning-based reconstruction (DLR) is a novel technique that can improve image quality and reduce scan time.

Study: The authors compared the performance of DLR with conventional methods for canine brain MRI using T2-weighted and fluid-attenuated inversion recovery (FLAIR) sequences.

Methods: Twelve clinically healthy beagles underwent brain MRI with different numbers of excitations (NEX) and DLR. The scan times, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and image quality scores were compared among the groups.

Results: DLR reduced scan time by 50% and 75% in T2-weighted and FLAIR sequences, respectively, and improved SNR and CNR compared with conventional methods. DLR also improved the image quality scores in all three indices: overall image quality, contrast, and perceived SNR.

Limitations: The study used a vendor-supplied DLR algorithm with a small sample size. The effects of DLR alone could not be compared in FLAIR sequences because compressed sensing was applied. The study did not evaluate DLR performance in other sequences and scan planes.

Conclusions: DLR can be used to improve the diagnostic efficacy of canine brain MRI while reducing scan time. DLR has the potential to replace conventional brain MRI in veterinary neuroimaging practice.

Visual comparison of qualitative analysis. NEX2DL and NEX1DL images show greater visual quality compared with conventional NEX4 images. FLAIR images show inferior image quality than T2W images. A–C, transverse T2W; D–F, transverse FLAIR; A, D NEX4; B, E NEX2DL; C, F NEX1DL. CSF, cerebrospinal fluid; DLR, deep learning-based reconstruction; FLAIR, fluid-attenuated inversion recovery; NEX, number of excitations; T2W, T2-weighted.

How did we do?

Login or Subscribe to participate in polls.

Disclaimer: The summary generated in this email was created by an AI large language model. Therefore errors may occur. Reading the article is the best way to understand the scholarly work. The figure presented here remains the property of the publisher or author and subject to the applicable copyright agreement. It is reproduced here as an educational work. If you have any questions or concerns about the work presented here, reply to this email.