Germany: Deep Learning-Accelerated Brain MRI Study

Deep learning (DL)-accelerated brain MRI is four times faster, offers better image quality and diagnostic confidence, and hence improves the detection of acute ischemic lesions compared to conventional MRI, a recent study published in Radiology has shown.

The researchers assert that the technique could translate to better patient care in that it could cut healthcare costs and improve imaging workflow.

Implementing this technique may be of great value, considering the increasing demand for medical examinations and increasing financial constraints placed on healthcare systems, the study stated.

DL-accelerated MRI can substantially reduce examination times. However, there is a lack of studies prospectively evaluating the diagnostic performance of DL-accelerated MRI reconstructions in acute suspected stroke. Therefore, Sebastian Altmann, University Medical Center Mainz, Germany, and colleagues investigated the interchangeability of DL-accelerated MRI with conventional MRI in patients with suspected acute ischemic stroke at 1.5 T.

For this purpose, they conducted a prospective study comprising 211 participants with suspected acute stroke who underwent clinically indicated MRI at 1.5 T between 2022 and 2023.

For each participant, conventional MRI (including T1-weighted, T2-weighted, T2*-weighted, T2 fluid-attenuated inversion-recovery, and diffusion-weighted imaging; 14 minutes 18 seconds) and DL-accelerated MRI (same sequences; 3 minutes 4 seconds), were conducted.

The primary endpoint was the interchangeability between conventional and deep learning-accelerated MRI for detecting acute ischemic infarction. Secondary endpoints included interchangeability regarding the impacted vascular territory and clinically relevant secondary findings (e.g., neoplasm, microbleeds).

The overall occurrence of acute ischemic stroke, clinically relevant secondary findings, affected vascular territory, diagnostic confidence, and overall image quality were evaluated by three readers. For acute ischemic lesions, size and signal intensities were assessed.

The interchangeability margin was chosen as 5%. For interrater reliability analysis and interrater agreement analysis, intraclass correlation coefficient and multi-rater Fleiss κ, respectively, were determined.

Findings

• The study sample comprised 211 participants (mean age, 65 years); 123 male and 88 female. Acute ischemic stroke was confirmed in 79 participants.
• Interchangeability was demonstrated for all primary and secondary endpoints. No individual equivalence indexes (IEIs) exceeded the interchangeability margin of 5% (IEI, −0.002).
• Almost perfect interrater agreement was observed.
• DL-accelerated MRI provided higher overall image quality and diagnostic confidence.
• The signal properties of acute ischemic infarctions were similar in both techniques and demonstrated good to excellent interrater reliability (intraclass correlation coefficient, ≥0.8).

"Despite being four times faster, DL-accelerated brain MRI was interchangeable with conventional MRI for detecting acute ischemic lesions," the researchers concluded.

Reference

Altmann S, Grauhan NF, Brockstedt L, Kondova M, Schmidtmann I, Paul R, Clifford B, Feiweier T, Hosseini Z, Uphaus T, Groppa S, Brockmann MA, Othman AE. Ultrafast Brain MRI with Deep Learning Reconstruction for Suspected Acute Ischemic Stroke. Radiology. 2024 Feb;310(2):e231938. doi: 10.1148/radiol.231938. PMID: 38376403.