Physica Medica: European Journal of Medical Physics - Volume 32 - Supplement 1 - February 2016

135 / 146

Introduction: Resting-state functional connectivity (fcMRI) represents a
novel fMRI approach that allows detection of temporal correlations in spon-
taneous BOLD signal oscillations while subjects rest quietly in the scanner.
Under resting conditions the brain is engaged in spontaneous activity that
causes a low frequencies (<0.1 Hz) BOLD signal fluctuations. Functional con-
nectivity (FC) can be defined as the synchrony of neural activity among
spatially distant regions.
Transcranial direct current stimulation (tDCS) is a non-invasive brain stim-
ulation technique that is known to modulate cortical activity and FC among
brain regions, as measured by functional magnetic resonance imaging.
This study is aimed at measuring the variation of functional connectivity
between cortical brain regions after tDCS along time.
Materials and Methods: For this purpose we enrolled 20 healthy right-
handed subjects. All subjects underwent 4 sessions RS-fMRI (10’ each, TR
2’’, 300 volumes, 1.5 T scanner): 2 immediately before and 2 after 20’ tDCS
over left M1. 10 of them received real (anodal) tDCS, 10 received sham stim-
ulation. We analyzed FC between left and right M1 with two different
statistical analyses: Seed-based correlation analysis (SCA) and the tempo-
ral concatenation group ICA (TC-GICA).
Results: Seed-based correlation analysis showed a significant decrease of
FC during the first fMRI acquisition immediately after anodal tDCS stim-
ulation (p = 0.005) that reaches back to baseline during the last fMRI session.
This behavior was not found in subjects who underwent sham stimula-
tion (p = 0.12).
The temporal concatenation group ICA showed that immediately after anodal
stimulation the average value of voxels decreases significantly (p < 0.05)
whereas there is no significant decrease in the case of sham tDCS stimulation.
Conclusions: Our results show that anodal tDCS is able to induce connec-
tivity changes within motor network, that is, reversible in a period lasting
between 10’ and 20’ after stimulation.
E.440
LONG TERM SIGNAL TO NOISE RATIO ANALYSIS OF MAGNETIC
RESONANCE IMAGES FROM MULTICOIL ARRAY
L. Mascaro A. Duina L. Rossi C. Pinardi M. Maddalo F. Dus
R. Moretti a Medical Physics Unit, Hospital Spedali Civili, Brescia, Italy;
b Professional School of Radiation Health Care Technologist, University of Brescia,
Brescia, Italy; c School of Medical Physics, University of Milan, Milan, Italy;
d Neuroimaging Lab, Medical and Surgical Specialties, Radiological Sciences and
Public Health, University of Brescia, Brescia, Italy; e Department of Medical and
Surgical Specialties, Radiological Sciences and Public Health, University of
Brescia, Brescia, Italy
Introduction: Image signal to noise ratio (SNR) is one of the most sensi-
tive parameters in MR, however the use of standard evaluation protocols
developed for single channel coils can be misleading for multicoil arrays.
Moreover, when performing an acceptance test, a statistical approach is
needed in order to define normalization ranges for routine evaluations. This
study presents the results of SNR evaluations, repeated for 6 months with
a variable time scheme, on head coil images, both for single coil and final
reconstructed images, in order to define normality ranges for system
performance.
Materials and Methods: A homogeneous spherical phantom was scanned
on a 1.5 T system with a 20 channel head coil, using a SE standard acqui-
sition protocol and a SOS reconstruction. 10 images were acquired each
time in order to evaluate short term reproducibility. Daily measurements
were performed for the first two weeks, then acquisitions were gradually
reduced up to a fortnightly measure after 6 months. Image analysis was
performed with NEMA SNR evaluation protocols.
Results: Single measure reproducibility in the combined images was
good in many cases, with the worse results for axial scanning direction
(CV 0.10% to 20%), while SNR had higher fluctuations for single coil
images (7.8%–23.7%). Long term analysis showed no significant trend of
the parameter except for one coil that had a continuous decrease of SNR
(from 250 to less than 50) after 3 months, due to a signal drop off, that was
not identified with the SNR evaluation of the final reconstructed images.
For all the analyzed case, an acceptability range was calculated on a
2σ basis.
Conclusion: Our analysis aimed to define acceptability ranges for SNR of
single coil elements and combined images, by means of short and long term
repeated measurements. Other than the definition of ranges of normal per-
formance, our results showed that reconstructed images are less sensitive
to system failure compared to single coil images.
E.441
MAPPING OF THE RESPONSE OF A 16 ELEMENT PHASED-ARRAY COIL
USED FOR BREAST MRS
G. Meliadò *, C. Cavedon, M.G. Giri, L. Camera, I. Baglio, S. Montemezzi.
Azienda Ospedaliera Universitaria Integrata Verona, Verona, Italy
Introduction: In MR spectroscopy, it is crucial to recognize the reasons of
inconsistent quantification of metabolite concentration due to inhomoge-
neities in the vicinity of the volume of interest (VOI) or to suboptimal coil
geometry. The purpose of this work was to assess the relative response of
a breast coil as a function of the position within the coil volume.
Materials and methods: On a 3 T MRI scanner equipped with a dedi-
cated 16 element phased-array coil and a couch to accommodate the patient
in prone position, a series of acquisitions was performed changing the VOI
position within a phantom consisting of a plastic bag (vol. = 1 l) contain-
ing a solution of choline chloride (1 mmol/l).
The characteristics of the acquisition sequence were the following: VOI se-
lection = PRESS; Samples = 1024; BW = 2000 Hz, TE = 135ms, TR = 3000ms,
VOI 10 mm3.
The comparison was assessed through the following figures of merit: peak
areas of choline (tCho) and non-suppressed water (H2O); ratio of peak areas
(tCho/H2O). The VOI position inside the phantom was described by the co-
ordinates of the machine reference frame: patient prone, feet first;
coordinates (AP, RL, FH) in mm (magnet center: 0, 0, 0); positive coordi-
nates increasing towards posterior (P), left (L) and superior (H) directions.
Results: For each figure of merit, results were normalized to 100 at the center
(−55, 95, 0) of the left coil housing: the tCho/H2O ratio measured at all po-
sitions had a substantially constant trend (mean= 104, sd = 7%). tCho (107,
27%) and H2O (101, 24%) showed a more accentuated variation. In partic-
ular, tCho assumed the value 55 for the most posterior position we tested
in the left housing of the coil (−10, 95, 0) and 67 in medial position (-55,
55, 0), while H2O measure 61 in both locations.
Conclusions: As the tCho absolute values result to be four orders of mag-
nitude lower than H2O, it might be difficult to detect tCho at positions where
the variation is greater with respect to the coil center.
E.442
DIFFUSION KURTOSIS IMAGING IN HEAD AND NECK AND BRAIN TUMOR:
A FEASIBILITY STUDY
S. Minosse A. Vidiri F. Piludu S. Marz a Post Graduated School of
Medical Physics, University La Sapienza, Rome, Italy; b Radiology and Diagnostic
Imaging Department, Regina Elena National Cancer Institute, Rome, Italy;
c Medical Physics Laboratory, Regina Elena National Cancer Institute, Rome, Italy
Introduction: The aim of this work was to investigate the feasibility of DKI
(diffusion kurtosis imaging) in patients affected by brain or head and neck
(HN) cancer. The DKI technique allows quantifying the non-Gaussian be-
havior of water diffusion in biological tissues, as a consequence of the
interaction of water molecules with underlying micro-structures (cell
membranes).
Materials and Methods: Eleven patients affected by HN tumor (5 pts) or
brain tumor (6 pts) underwent a MR examination at 3 T, including DKI at
ultrahigh b-values up to 2500 s/mm2. The two parameters Dk (diffusion
coefficient) and K (diffusional kurtosis) of the DKI model were derived
using a home-made code. Conventional ADC (apparent diffusion coeffi-
cient) and D (pure diffusion coefficient) were also derived using the same
DKI data set. Parametric maps of Dk, K, ADC and D were generated. The
value of the residual sum of squares (RSS) of the best solution was calcu-
lated and used to assess the quality of the fit at the voxel level. The lesion
was manually contoured by an expert radiologist to define a region of
e130 Abstracts/Physica Medica 32 (2016) e124–e134