Caecilia S Reiner1,2, Bernd Kuehn3, Daniel Nanz2, Berthold Kiefer3, and Gustav Andreisek1,2
1Institute of Diagnostic and Interventional Radiology, University Hospital Zurich, Zurich, Switzerland, 2University of Zurich, Zurich, Switzerland, 3Oncology Application Predevelopment, Siemens Healthcare, Erlangen, Germany
Synopsis
An increasing demand for time-efficient and standardized MRI drives the
development of automated MRI-workflows. The purpose of this study was to
evaluate the feasibility of such a novel automated scanner-workflow for
multi-station MRI. The scanner automatically detects the body regions selected
and sequences are automatically adjusted to patient’s size and breath-holding
capacity to generate optimal image quality. In 20 patients scanned with a
multi-station protocol (chest, abdomen, and/or pelvis), image quality was good
to excellent and complete body region coverage was achieved in 95% of patients.
This nearly “single-button” automated multi-station MRI could open new
possibilities in the diagnostic process.Background and Purpose:
Time- and cost-efficiency are among the major challenges in clinical
magnetic resonance imaging (MRI), mainly driven by the shortening of
reimbursement in most health care systems. At the same time, there is an
increasing overall demand for higher quality of MRI exams regarding the
comparability of exams, i.e. important for primary and/or follow-up studies in
oncologic patients. To address these challenges, several vendors and
researchers are developing automated scanner workflows for clinical MRI
systems. The hypothesis is that these workflows allow a standardized and
time-efficient use and provide a robust image quality at only little user
interaction. The purpose of this study was to evaluate the feasibility of a
novel automated scanner workflow for multi-station MRI such as chest, abdomen
and pelvis, gauge the reliability of the algorithm with regard to completeness
of anatomical coverage, and assess overall image quality and severity of
imaging artifacts.
Material and Methods
Patient cohort:
Twenty oncologic patients (10 women, 10 men; mean age, 49 ys, range,
20-75 ys) were examined on a 3 T MRI scanner (MAGNETOM Skyra, Siemens
Healthcare) using a Whole-Body Dot Engine
supporting an automated multi-station scanner workflow. Combined exams
of chest, abdomen and/or pelvis were performed in 13 patients for oncologic
follow-up, in three for primary staging, and in four for tumor screening.
Image acquisition:
The Whole-Body Dot Engine automatically detects landmarks like lung
apex, lung recessus, diaphragm, liver apex, iliac bone on a fast low resolution
whole body scout, which is acquired during moving table. Based on this the body
regions selected for scanning, namely chest, abdomen and/or pelvis are
segmented (Figure 1). Sequence parameters are automatically adjusted in order
to ensure proper coverage of the body regions of interest. Additionally, the
Whole-Body Dot Engine uses an anticipated patient’s breath-hold capacity to
automatically adjust the imaging protocols in body regions where breathhold is
required generating optimal image quality. In this study, the breath-hold
capacity was set to 20 sec. The protocol included coronal and transverse
T2-weighted single-shot turbo-spin-echo (HASTE) (TR/TE, 1230/92 ms; flip angle,
160°; slice thickness (ST)/spacing, 5/1mm; matrix, 256x256), transverse
single-shot diffusion-weighted echo-planar imaging with slice-specific shim
optimization (EPI-DWI, iShim (1)) (TR/TE, 6100/56ms; flip angle, 90°;
ST/spacing, 5/1mm; matrix 128x128), and
transverse T1-weighted pre- and post-contrast 3D spoiled gradient-echo 2-point
Dixon (VIBE) (TR/TE, 4.27/1.28ms; flip angle, 12°; ST 3mm; matrix 320x192) pre-
and post-contrast acquisitions (delay:
chest 35 sec, abdomen 70 sec, pelvis 90 sec after injection of 0.1mmol/kg
bodyweight gadoterate meglumine, Dotarem, Guerbet). The cranio-caudal coverage
per block was adjusted to 400 mm.
Image evaluation:
The scans were evaluated for overall image quality (IQ) (5=excellent,
4=good, 3=moderate, 2=poor, 1=non-diagnostic) and artifacts (5=no artifacts,
4=mild artifacts, 3=moderate artifacts, 2=severe artifacts, 1=non-diagnostic)
on a 5-point scale by a board-certified abdominal radiologist with 8 years of
experience. The image acquisition time was noted as well as whether the
coverage of the targeted body region was complete. Descriptive statistics were
performed.
Results
In 16 patients, abdomen and pelvis were scanned, in three patients
chest, abdomen and pelvis, and in one patient only chest. Overall in all but
one patient (19 of 20, 95%), the selected body regions were covered completely
by the automated algorithm, with exception of the DWI, which did not cover the
sub-diaphragmatic part of the liver in four patients (4 of 20, 20%). The mean
score for overall IQ was 4.6 ± 0.49 standard deviation (SD) and for artifacts
was 4.2 ± 0.57 SD. In two patients (2 of 20, 10%), mild respiratory motion
artifacts were observed on T1-weighted post-contrast images (Figure 2). The
mean examination time was 26.3±3.4min for chest, abdomen and pelvis, 21.2 ± 3.4
min for abdomen and pelvis, and 14.1min for chest only.
Conclusion
MR scanning with the automated Whole-Body Dot Engine results in good to
excellent image quality within a reasonable total examination time with only
small patient dependent variations. A nearly “single-button protocol” for
standardized fast, reproducible, and automated workflow of chest, abdomen, and
pelvis, could open new possibilities in the diagnostic process. However,
further comparison studies with traditional manual scan modes need to be
performed to support our preliminary experience.
Acknowledgements
NoneReferences
1.
Alto
Stemmer et. al.: Combination of integrated slice-specific dynamic shimming and
pixel-wise unwarping of residual EPI distortions. Proc. Intl. Soc. Mag. Reson.
Med. 23 (2015) #3729