The aim of this study was to evaluate the performance of an automated workflow for whole-body magnetic resonance imaging (WB-MRI), which reduces user interaction compared with the manual WB-MRI workflow.
This prospective study was approved by the local ethics committee. Twenty patients underwent WB-MRI for myopathy evaluation on a 3 T MRI scanner. Ten patients (7 women; age, 52 ± 13 years; body weight, 69.9 ± 13.3 kg; height, 173 ± 9.3 cm; body mass index, 23.2 ± 3.0) were examined with a prototypical automated WB-MRI workflow, which automatically segments the whole body, and 10 patients (6 women; age, 35.9 ± 12.4 years; body weight, 72 ± 21 kg; height, 169.2 ± 10.4 cm; body mass index, 24.9 ± 5.6) with a manual scan. Overall image quality (IQ; 5-point scale: 5, excellent; 1, poor) and coverage of the study volume were assessed by 2 readers for each sequence (coronal T2-weighted turbo inversion recovery magnitude [TIRM] and axial contrast-enhanced T1-weighted [ce-T1w] gradient dual-echo sequence). Interreader agreement was evaluated with intraclass correlation coefficients. Examination time, number of user interactions, and MR technicians' acceptance rating (1, highest; 10, lowest) was compared between both groups.
Total examination time was significantly shorter for automated WB-MRI workflow versus manual WB-MRI workflow (30.0 ± 4.2 vs 41.5 ± 3.4 minutes, P < 0.0001) with significantly shorter planning time (2.5 ± 0.8 vs 14.0 ± 7.0 minutes, P < 0.0001). Planning took 8% of the total examination time with automated versus 34% with manual WB-MRI workflow (P < 0.0001). The number of user interactions with automated WB-MRI workflow was significantly lower compared with manual WB-MRI workflow (10.2 ± 4.4 vs 48.2 ± 17.2, P < 0.0001). Planning efforts were rated significantly lower by the MR technicians for the automated WB-MRI workflow than for the manual WB-MRI workflow (2.20 ± 0.92 vs 4.80 ± 2.39, respectively; P = 0.005). Overall IQ was similar between automated and manual WB-MRI workflow (TIRM: 4.00 ± 0.94 vs 3.45 ± 1.19, P = 0.264; ce-T1w: 4.20 ± 0.88 vs 4.55 ± .55, P = 0.423). Interreader agreement for overall IQ was excellent for TIRM and ce-T1w with an intraclass correlation coefficient of 0.95 (95% confidence interval, 0.86–0.98) and 0.88 (95% confidence interval, 0.70–0.95). Incomplete coverage of the thoracic compartment in the ce-T1w sequence occurred more often in the automated WB-MRI workflow (P = 0.008) for reader 2. No other significant differences in the study volume coverage were found.
In conclusion, the automated WB-MRI scanner workflow showed a significant reduction of the examination time and the user interaction compared with the manual WB-MRI workflow. Image quality and the coverage of the study volume were comparable in both groups.
From the *Institute of Diagnostic and Interventional Radiology, University Hospital Zürich;
†University of Zürich, Zürich, Switzerland;
‡Oncology Application Predevelopment, Siemens Healthcare GmbH, Erlangen, Germany;
§Siemens Healthcare AG, Zürich; and
∥Radiologie, Spital Thurgau, Kantonsspital Munsterlingen, Münsterlingen, Switzerland.
Received for publication January 24, 2018; and accepted for publication, after revision, February 22, 2018.
Conflict of Interest: Bernd Kuehn, PhD, and Berthold Kiefer, PhD, are employees of Siemens Healthcare GmbH, Erlangen, Germany. Markus Klarhöfer, PhD, is an employee of Siemens Healthcare AG, Zürich. All other authors declare that they have no conflict of interest.
Correspondence to: Caecilia S. Reiner, MD, Institute of Interventional and Diagnostic Radiology, University Hospital Zurich and University of Zurich, Raemistrasse 100, 8091 Zürich, Switzerland. E-mail: email@example.com.