Metabolic Concentrations and Ratios of Brain Tissue

1
Metabolic Concentrations and Ratios of Brain
Tissue Amarjeet Bhullar, Lars Ewell, Tim McDaniel
and Baldassarre Stea Department of Radiation
Oncology, The University of Arizona 2008 AHSC
Frontiers in Biomedical Research Poster Forum
October 15, 2008
Introduction Proton magnetic resonance
spectroscopy gives us in vivo spectrum, which
provides metabolites concentration and their
ratios on brain tissue. This information can
be useful to distinguish between healthy and
abnormal tissues and thus can help to follow-up
the disease. Due to artifact, MRS signal
contains undesirable baseline and noise, making
it difficult to calculate the absolute
concentrations. In this study, the data were
recorded on GE Signa MRI (3T) and transferred to
a work station (sun, GE Healthcare) for offline
post-processing using Functool 2004 software. 
The signal intensity of various metabolite peaks
was evaluated using integrals of each peak as a
measure of its concentration (intensity). The
integration limits of each respective peak were
manually defined before Functool's calculation.
In this calculation, Functool generates
metabolite concentration and directly their
ratios. It is also possible to calculate the
metabolite ratio from Functool's generated
concentrations. The directly generated Functool
ratios and ratios calculated from Functool
generated metabolite concentrations are not
significantly different from our calculation.
Quantitative analysis was also done by using
linear combination model. Data Acquisitions and
Methods Both phantom and patient exams were
recorded on 3T GE scanner. The following
parameters were used for all 2D CSI spectrums a
point resolved spectroscopy sequence (PRESS)
TR/TE, 1000/144 field of view, 18 cm matrix, 8
x 8 slice thickness, 10 mm. Choline (Cho),
creatine (Cr), and N-acetylaspartate (NAA)
concentrations were derived and the following
metabolite ratios were calculated Cho/Cr and
Cho/NAA. Metabolite concentrations and ratios
were calculated using (a) manually adjusted peak
boundaries by Functool (b) Gaussian function
overlapped on metabolite peak with baseline
correction (c) Gaussian function overlapped on
metabolite peak with baselinenoise correction.

Mathematical Model Gaussian Function
Results and Outlook
Calculation
Plots of ADC values for two patients. Lesions
extended over four MRI slices. The plot on the
left shows ADC values for four examinations with
low baseline intensity. The right plot shows
three exams for which the baseline intensity was
over 3000.
Conclusion Our theoretical model is helpful to
determining the Cho/Cr and Cho/NAA ratios to
chasing the disease. These ratios may be best
numeric discriminators for brain tissue.
Plots showing the dependency of calculated ADC
values on location. Plots are for normal tissue
versus slice number measured superior to
inferior. The left hand plot shows baseline
intensity values for three different patients
measured in areas of normal tissue. The right
hand plot shows the resulting ADC dependency on
slice location for normal tissue.