Deep learning reconstruction for accelerated high-resolution upper abdominal MRI improves lesion detection without time penalty

Jan M Brendel; Johann Jacoby; Reza Dehdab; Judith Herrmann; Stephan Ursprung; Sebastian Werner; Sebastian Gassenmaier; Dominik Nickel; Konstantin Nikolaou; Saif Afat; Haidara Almansour

doi:10.1016/j.diii.2024.09.008

Deep learning reconstruction for accelerated high-resolution upper abdominal MRI improves lesion detection without time penalty

Diagn Interv Imaging. 2024 Nov 19:S2211-5684(24)00205-5. doi: 10.1016/j.diii.2024.09.008. Online ahead of print.

Affiliations

¹ Department of Radiology, Diagnostic and Interventional Radiology, Tuebingen University Hospital, University of Tuebingen, 72076 Tuebingen, Germany.
² Institute of Clinical Epidemiology and Applied Biometry, Tuebingen University Hospital, University of Tuebingen, 72076 Tuebingen, Germany.
³ Department of MR Application Predevelopment, Siemens Healthineers, 91301 Forchheim, Germany.
⁴ Department of Radiology, Diagnostic and Interventional Radiology, Tuebingen University Hospital, University of Tuebingen, 72076 Tuebingen, Germany; Cluster of Excellence iFIT (EXC 2180) "Image-guided and Functionally Instructed Tumor Therapies", University of Tübingen, 72076 Tuebingen, Germany.
⁵ Department of Radiology, Diagnostic and Interventional Radiology, Tuebingen University Hospital, University of Tuebingen, 72076 Tuebingen, Germany. Electronic address: haidar.almansour@gmail.com.

PMID: 39567306
DOI: 10.1016/j.diii.2024.09.008

Abstract

Purpose: The purpose of this study was to compare a conventional T1-weighted volumetric interpolated breath-hold examination (VIBE) sequence with a DL-reconstructed accelerated high-resolution VIBE sequence (HR-VIBE_DL) in terms of image quality, lesion conspicuity, and lesion detection.

Materials and methods: Consecutive patients referred for upper abdominal MRI between December 2023 and March 2024 at a single tertiary center were prospectively enrolled. Participants underwent 1.5 T upper abdominal MRI with acquisition of spectrally fat-saturated unenhanced and gadobutrol-enhanced conventional VIBE (fourfold acceleration, 3.0 mm slice thickness, 72 axial slices) and HR-VIBE_DL (sixfold acceleration, 2.0 mm, 108 slices). Both sequences had an identical acquisition time of 16 s. Image analysis was performed by three readers in a blinded and randomized fashion, with respect to image quality, lesion conspicuity, and lesion detection in liver, pancreas, spleen, lymph nodes and adrenal glands. Image quality parameters were compared using repeated measures analysis of variance. Lesion detection rates were compared using Fisher exact test. Inter-reader agreement was assessed using Fleiss κ test.

Results: Among 744 consecutive patients, 50 participants were evaluated. There were 30 men and 20 women, with a mean age of 60 ± 15 (standard deviation [SD]) years (age range: 18-88 years). HR-VIBE_DL images demonstrated superior signal-to-noise ration and edge sharpness by comparison with conventional VIBE images (P < 0.001 for both), with substantial interreader agreement (κ: 0.70-0.90). Lesion conspicuity was higher with for HR-VIBE_DL images (3.50 ± 0.83 [SD]) by comparison with conventional VIBE images (3.21 ± 0.98 [SD]) (P = 0.005). There were 171 upper abdominal lesions, yielding a total of 513 for all three readers. HR-VIBE_DL images yielded higher lesion detection rate (97.5 %; 500/513) compared to conventional VIBE images (93.2 %; 478/513) (P = 0.002).

Conclusion: HR-VIBE_DL images of the upper abdomen result in superior image quality, better lesion conspicuity, and improved lesion detection without time penalty by comparsion with conventional VIBE images.

Keywords: Deep learning; High-resolution; Lesion detection; Upper abdominal MRI; Volumetric interpolated breath-hold examination (VIBE).