Joint Annual Meeting ISMRM-ESMRMB ~ 19-25 May 2007

Unsolved Problems and Unmet Needs in Magnetic Resonance

Unsolved Problems and Unmet Needs in MR:
Results from a survey of ISMRM Study Groups, December 2005



Late in 2005, the ISMRM study groups were asked to compile lists of key unsolved problems and unmet needs in their areas of interest.  Here are the results:


MR Engineering



1.   Wireless or optical transmission of the MR signal out of the bore, to facilitate the use of large receiver coil arrays.


2.   What number of channels is optimal/required for both receive (Parallel Imaging) and transmit (Transmit SENSE) .


3.   SAR: how to control SAR at high fields (> 3T), how to know where the hot spots are in vivo, how to monitor the SAR in the human body in real time.


Magnets, Shim and Gradient Coils

4.   Length of magnet/bore: how short can it be?  What are the limits in openness?


5.   Actively shielded > 7 T whole body magnets.


6.   Dynamic shimming: development of shim coils that can be used in dynamic shimming (strong, low inductance, actively shielded) and quantification of the advantages to be gained via dynamic shimming.


7.   Reduction of acoustic noise to due switched magnetic field gradients, especially at high field.


8.   Methods for imaging without the use of pulsed field gradients.



9.   Methods for Precise QA of the MRI scanner in terms of SNR/ stability/ distortion/ artifacts and differences between MR methods.  Currently, the lack of such methods limit the pooled use of MR data generated from different scanners/vendors.


10.  Practical methods for measuring the electrical properties of tissues using MRI.


11.  Methods for increasing patient throughput  - can we produce an order-of-magnitude improvement in any way (parallel imaging, higher field, dockable tables, etc.) in the time needed to image (leading to a reduction in scanning costs).



12.  A discussion board to draw an active group of engineers together for dissemination of knowledge and experience, rather than retainment in individual labs.


High-Field Systems and Applications


I. Hardware

1.   Solutions for the transmit B1 inhomogeniety at high field due to wavelength/dielectric effects.


2.   Development of standardized method for SAR monitoring and setting SAR limits for multiple local transmit coils.


3.   Development of transmit SENSE at ultra-high field, along with development of algorithms for determining the phase and amplitude of the RF for each driven element in a coil, such as a TEM coil.


4.   Acoustic noise reduction at high field.


5.   Low noise, high fidelity, and powerful gradient amplifiers.


6.   Development of an actively shielded 7 T whole-body sized magnet.


II. Applications

7.   Fast CSI pulse sequences.


8.   Pulse sequences for clinical applications of hetero-nucleus imaging (such as 3Li, 31P, 17O, etc.)


9.   Detection of molecular probes that can pass through the blood-brain barrier and selectively bind to specific cells or tissues.


10.  A method for identifying areal boundaries via high resolution MRI (to allow definition of functional areas on individual subjects).


III. Misc.

11.  Long-term health consequences of exposure to very high magnetic field.


12.  Methods to remove magnetic susceptibility-induced image distortion.


Diffusion & Perfusion

The Top Twenty -- If only we could:

1.   “Definitively quantify the contributions of the intra- and extra-cellular components to the diffusion signal” [Score = 30; Number of votes = 9]

2.   “Measure brain perfusion in acute stroke patients accurately enough to be useful”
[Score = 27; Number of votes = 7]

3.   “Have realistic phantoms for diffusion imaging”
[Score = 26; Number of votes = 11]

4.   ”Quantify connectivity between different regions, separated by different distances, reliably and consistently”
[Score = 25; Number of votes = 8]

5.   “Have an accepted and practical gold standard for tract tracing in the human brain”
[Score = 23; Number of votes = 7]

6.   “Reliably measure the vascular territories of individual arteries”
[Score = 21; Number of votes = 6]

7.   “Better understand the biophysical nature of diffusion MR signal, in order to optimize diffusion experiments more effectively”
[Score = 20; Number of votes = 4]

8.   “Have a definitive, reproducible and easy way of calibrating ASL experiments with respect to M0 (of tissue or blood, depending on the model) in order to get quantitative CBF values”
[Score = 18; Number of votes = 4]

9.   “Use diffusion-derived measures, other than the mean diffusivity, in a clinical manner”
[Score = 16; Number of votes = 5]

10. “Have standard post-processing software (including motion correction) for ASL integrated onto a clinical scanner”
[Score = 16; Number of votes = 5]

11.  “Perform meaningful group comparisons on low dimensionality diffusion data (scalar invariants of the tensor), even if we don’t understand the biophysical mechanisms underlying them” / “Perform meaningful voxel-based comparisons of DTI data”
[Score = 16; Number of votes = 4]

12.  “Reliably quantify the dependence of diffusion on diffusion time, to identify different tissue types or geometrical features”
[Score = 15; Number of votes = 5]

13.  “Have realistic phantoms for perfusion imaging”
[Score = 15; Number of votes = 5]

14.  “Use DTI to reliably discriminate tumor infiltration from bland (tumorfree) edema”
[Score = 14; Number of votes = 5]

15.  “Use arterial spin labeling to measure perfusion in white matter”
[Score = 14; Number of votes = 3]

16.  "Establish the definitive biophysical mechanism underlying the dependence of ADC on the b-value in the brain".
[Score = 12; Number of votes = 4]

17.  “Resolve the topological ambiguity in diffusion displacement profiles (cross, kiss, twist, bend)”
[Score = 11; Number of votes = 4]

18.  “Reach a common consensus, once and for all, on the number of directions needed for DTI, and the b-  values needed (i.e., do we need b < 600 s/mm2 or b> 3000 s/mm2)”
[Score = 11; Number of votes = 4]

19.  “Identify the cellular correlates of changes in diffusion anisotropy in white matter”
[Score = 11; Number of votes = 5]

20.  “Know whether FA abnormalities in specific fibers correlate with fMRI abnormalities in gray matter (i.e., use DTI to predict the outcome of fMRI experiments)” / “Identify the functional basis and the functional meaning of DTI”
[Score = 10; Number of votes = 5]


Dynamic MR

1) A hand-held MRI scanner. The NMR mouse exists and is capable of measuring spectra
    at very small depths. However, for hand held medical imaging we would need to
a. Create a reasonably homogeneous B0 field penetrating the body to organ depth.
b. Overcome B0 field inhomogeneities.
c. Create reasonably linear B0 gradients
d. Compensate for the residual non-linearity of the gradients.
e. Have MRI sequences with only adiabatic pulses
f.  Have enough B1 penetration for the adiabatic pulse without exceeding SAR limits.

2) Create “The mother of all sequences” (so named by Paul Bottomley), a truly high resolution
    3D pulse sequence that in a reasonably short scan time measures spin density, T1 and T2.
    Then use this information to generate any slice at any oblique angle with any T1 or T2 weighted
    contrast. This would replace the series of scout images, oblique scouts, T1 and T2 weighted images
    that usually make examinations long.

3) Measure real time changes in spin density at a time scale shorter than T1. Example measure
    real time changes in phosphorous metabolites or in intracellular sodium in vivo in muscle or other
    excitable tissue.

4) Measure changes in T1 at a time scale shorter than T1.





1.      Applications of Parallel Imaging in MSK MRI

2.      New coils for 3T and 7T: phased array, small coils

3.      Kinematic and loaded imaging of joints and spine

4.      Novel sequences to image cartilage including SSFP and uTE

5.      Automated segmentation and parameter mapping software (morphology, T1, T2, T1rho)

6.      Spectroscopy for MSK MRI - muscle, spine, cartilage (1H, 31P, Na)

7.      Fat saturation for low field - robust Dixon imaging

8.      New hardware: gradient inserts and dedicated extremity 3T systems

9.      Fast T1 and T1rho mapping sequences

10.    DWI and DTI imaging for muscle and nerves

11.    Marrow Imaging



White Matter


1.      How to distinguish between demyelination, inflammation and axonal loss?


2.      Early diagnosis of Alzheimers (before occurrence of atrophy)


3.      Increased specificity for diagnosis of Multiple Sclerosis in patient with WM lesions


4.      Possibility to establish glioma grading with high confidelity


5.      Expand the MRI capabilities for brain white matter to high resolution imaging of the spine columns


6.      Diagnose White Matter Pathology Definitively using MRI.


MR Flow & Motion Quantitation


1.      How to achieve robust, accurate, high-temporal-and-spatial-resolution flow measurements in,
         for example, the right and circumflex coronary arteries, which move a lot.


2.      We need a well-defined, complex flow phantom for comparison/calibration of scanners.


3.      How to have first-time, every-time measurements of flow with likely errors less than 10%

         in PATIENTS in vessels down to 5mm diameter?


4.      What consistutes a clinical report of an MR flow study? an MR cardiac motion study?


5.      When will MR velocimetry match ultrasound?


6.      Spiral, EPI, segmented gradient echo, FVI, etc, etc, etc, do we have the right tools for
         every situation? and do we really need them all in clinic and in research settings?


7.      Limits of temporal, spatial and velocity resolution - how low and how high can we go
         (in flow and tagging)?


8.      How well does phase mapping work with accelerated imaging techniques?


9.      How do I know if a flow study was done well?


10.     4D velocity mapping - how fast? how good?