3D Chemical Exchange Saturation Transfer MR Imaging of In Vivo Human Knee
Cartilage Glycosaminoglycan at 7T
Guruprasad Krishnamoorthy1,2 , Puneet Bagga1, Hari Hariharan1, Ravinder Reddy1
1Center
for Magnetic Resonance and Optical imaging, Department of Radiology, University of Pennsylvania
2School of Biomedical Engineering, Science and Health Systems, Drexel university
Materials and Methods
Abstract
Osteoarthritis (OA) is a degenerative disease affecting molecular composition of

Five healthy volunteers of age 32.8 ± 14 yrs underwent MRI at whole body 7T scanner (Siemens
cartilage of Knee. Early detection of OA is critical for preventing the progression and reversing
Healthcare, Erlangen, Germany) using a 28-channel Knee Coil (Quality Electrodynamics, OH, USA).
the course of the disease when the cost of the treatment is low and treatment success rate is
Three volunteers were scanned twice to determine reproducibility
much higher. Early stage of OA is associated with the loss of Glycosaminoglycan which are side

Results
A
1
B
2
1
2
A new 3D gagCEST sequence has been developed as shown in the figure below. The preparation
20
block consists of a CEST preparation. Readout is done with very low flip (~5°) angle spoiled
available non-invasive methods to measure PG have limitations in either specificity, sensitivity or
Gradient echo segments (~600 per shot with a GRAPPA factor of 2 along phase encoding direction)
spatial resolution. A recently developed method known as Chemical Exchange Saturation
to obtain rapid 3D images
Pulse
Cluster
method is highly specific to Glycosaminoglycan (GAG), current implementation of this method is
16
Readout
(Cartesian 3D GRE)
CEST Preparation
14
T1 Recovery Delay
based on single slice, time intensive and not conducive for routine evaluation of patients. In this
Duration
0–5s
5–8s
0-3s
study, a novel 3D / multi-slice gagCEST imaging technique is developed to image PG of human
knee cartilages in a practically achievable scan time with reproducible results at 7T MRI
12
10
gagCEST%
Transfer imaging of Glycosaminoglycan (gagCEST) images PG non-invasively. While this
18
gagCEST%
chains of Proteoglycan (PG) molecules from the Extracellular Matrix of cartilage. Currently
8

Elliptical center encoding strategy was used for the phase-slice encode plane. Using this view
6
ordering and ‘burst mode’ shots strategy, high throughput rate (>80%) for readout is achieved.
Background
Briefly, gagCEST technique selectively saturates the exchangeable hydroxyl protons of GAG by

4
Imaging parameters
the application of radio frequency (RF) pulse; the saturated protons of GAG exchange with
protons of bulk water leads to a reduced of bulk water signal, which is concentration dependent

140 mm FOV with 0.6X0.6X3 mm3 resolution for 234X234X16 matrix size.
2

TR/TE = 7.5/3 msec with 5° flip angle, 2 shots with 8sec shot TR , GRAPPA factor = 2
0
Figure:
B1 GRE => 30° and 60 ° flip angle method, without GRAPPA

A. 3D Multi-slice gagCEST maps of cartilages of a healthy young volunteer

WASSR B0map => 200 msec duration with 20 Hz power for 0:0.1:0.8 ppm offsets

gagCEST => 500 msec duration for 0.2:0.1:1.8 ppm offsets (1ppm center for gag)
B. gagCEST maps of 65-year old subject with knee pain showing reduced patellar
cartilage thickness and gagCEST values.
1. Axial orientation showing patellar cartilage
2. Coronal orientation showing Femoral (on top) and Tibial cartilages
Results (contd.)
Note: The gagCEST maps are overlaid on anatomical image and cropped
A
Sherry, Woods, 2008
Conclusion
B
Figure: gagCEST values of healthy
Medial
Where, Msat( ±Δω) are B0 corrected CEST images acquired at ±1 ppm from water resonance.
Mcntl can either be M0 i.e., -20ppm or Msat(- Δω) ie., -1ppm
Lateral
20
% gagCEST
Msat −∆ω − Msat(+∆ω)
gagCESTasym = 100 ∗
Mcntl
Medial
6
cartilages of different compartments
4
of knee
4
2
A. M-ve Normalized
0
0
16
12
8
Patellar
Femoral
Tibial
Patellar
A
Ling et al, Proceedings of the National Academy of Science 2008;105(7):2266-2270
Singh A et al, Magnetic Resonance in Medicine. 2012;68(2):588-594
Krusche-Mandl I et al, Osteoarthritis and Cartilage. 5// 2012;20(5):357-363
Varma G et al, Magnetic Resonance in Medicine. 2012;68(4):1228-1233
Singh A et al, Magnetic Resonance in Medicine. 2013;69(3):818-824
Kogan F et al, Current Radiology Reports. 2013/06/01 2013;1(2):102-114.
Liu G et al, NMR in Biomedicine. 2013;26(7):810-828.
Wei W et al, Magnetic Resonance Imaging. 1// 2014;32(1):41-47.
Retest - % gagCEST
1.
2.
3.
4.
5.
6.
7.
8.
Femoral

We have developed a highly reproducible MRI technique to quantify GAG at a
practically achievable scan time

Figure: Intra-Subject reproducibility
8
To the best of our knowledge, this is the first study reporting 3D multi-slice
gagCEST of various compartments of human knee at 7T
B. M0 Normalized
Tibial
B
20
References:
volunteers calculated from
Lateral
8

There has been a lot of interests shown by doctors and pharmaceutical industry to
diagnose the health of cartilage non-invasively to determine patient response to a
of the proposed technique.
16
ICC = 0.949*
ICC = 0.948*
6
particular drug under development
A. M-ve Normalized
12
B. M0 Normalized
4

Work is in progress to evaluate more healthy subjects and patients with Knee pain
8
Note: Intra-class correlation(ICC)
2
4
above 0.9 shows that the technique
0
0
0
4
8
12
16
Test - % gagCEST
20
0
2
4
6
Test - % gagCEST
8
is highly reproducible
Acknowledgements: This work is supported by the National Institute
of Biomedical Imaging and Bioengineering under award number P41-EB015893