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

130 / 146

Furthermore, the use of the SNRL parameter allows comparing the per-
formance of different scanners in terms of SNR values regardless of technical
specifications of each manufacturer.
Reference
[1]
E.423
RENAL FUNCTION ASSESSMENT BY TEXTURE ANALYSIS OF R2 MAP
M. Biondi L. Pellicci A. Bogi L.N. Mazzon E. Vanz G.M. Belmont
G. De Otto S.F. Carbone A. Guasti F. Banci Buonamic a Department
of Medical Physics, University Hospital of Siena, Siena, Italy; b Department of
Medical, Surgical and Neurosciences, University of Siena, Siena, Italy;
c Department of Diagnostic Imaging, University Hospital of Siena, Siena, Italy;
d Department of Medical Physics, U.S.L. 7, Siena, Italy
Introduction: Texture analysis is a new method for digital images inves-
tigation which provides a measure of intralesional heterogeneity. First order
statistics include mean, standard deviation, skewness, kurtosis, uniformi-
ty, and entropy of the gray level histogram. The close relationship between
the functional magnetic resonance imaging and renal function is well known
in literature. In light of these backgrounds, the purpose of our work is to
evaluate texture analysis of R2 maps (TAR2m) in the assessment of renal
function.
Materials and Methods: Axial multi-echo FGRE sequence on upper
abdomen was acquired in 11 patients with renal parenchymal diseases
at different renal function stages; 7 subjects without renal disease
were used as controls. Serum Creatinine (sCr) of all the subjects was ob-
tained and eGFR was calculated by MDRD formula. A hand-made ROI on
central slice R2 map was used to sample renal parenchyma, including
renal cortex and medulla, and finally mean, median, kurtosis, skewness,
and density were calculated by using an open source texture analysis
software.
Results: sCr and skewness were found to have a significant relationship
(p < 0.05). Significant differences were found between stage 1 and 2 for
density (p = 0.04) and nearly significant between stage 2 and 3 for skew-
ness (p = 0.07).
Conclusions: TAR2m of kidney does not seem to be able to stratify renal
impairment exception for the skewness and density. These latter param-
eters seem to be very promising and further studies on a large population
are needed to best estimate the present preliminary data.
E.424
RECTAL CANCER TEXTURE ANALYSIS APPLIED ON ADC MAPS IN
RESPONSE ASSESSMENT TO NEOADJUVANT THERAPY
M. Biondi A. Bog L.N. Mazzon E. Vanz G.M. Belmonte G. De Otto
R. Martini E. Foderà S.F. Carbone L. Volterrani A. Guasti ,
F. Banci Buonamici a Department of Medical Physics, University Hospital of
Siena, Siena, Italy; b Department of Medical, Surgical and Neurosciences,
University of Siena, Siena, Italy; c Department of Diagnostic Imaging, University
Hospital of Siena, Siena, Italy; d Department of Medical Physics, U.S.L. 7, Siena,
Italy
Introduction: The aim of this work was to evaluate feasibility of texture
analysis of ADC maps in the assessment of response in neoadjuvant
therapy of rectal cancer. With digital images texture analysis (TA) it is
possible to measure ROI heterogeneity. This method consists in the eval-
uation of some gray level histogram parameters: mean (M), standard
deviation (SD), skewness (SK), kurtosis (K), uniformity (U) and entropy
(E).
Materials and Methods: In this retrospective observational study, ten
patients affected by rectal cancer underwent to MR imaging before
and after neoadjuvant chemo-radiation therapy (CRT); in all the
cases, post-surgical tumor regression grade (TRG) was obtained. ADC
maps were calculated by diffusion-weighted imaging (b-values
0–800 s/mm2).
Image analysis was made with a homemade ImageJ macro. Two blinded
readers manually segmented hypontense areas corresponding to tumor ADC
restriction. TA parameters for selected ROIs were evaluated applying a
Laplacian of Gaussian bandpass filter between 0.5 and 2.5 to highlight dif-
ferent spatial scales.
ANOVA was performed to evaluate differences among responder (R, TRG1-
2) and non responder (NR, TRG 3) patients and longitudinal changes. Inter-
reader variability was assessed by ICC.
Results: Three patients were NR (TRG 3), while seven were considered R
(TRG 1-2). M, E and U showed an ICC value > 0.75 while SK and K had an
ICC value < 0.50.
Among R and NR group, SD value was not significantly lower in R pa-
tients in pre-CRT ADC maps (p = 0.05); however, filter application seems
to differentiate despite the small population considered. M, E and U values
differed significantly after CRT (p < 0.05). Post-CRT changes were signifi-
cant in R patients for M value (p < 0.01) while SD, E and U changed
significantly both in R and NR patients.
Conclusions: TA seems to be a promising approach, however, S
and K showed a high inter-reader variability and further studies on
large population are needed even with ADC maps with better spatial
resolution.
E.424 bis
A NEW REFINED ACOUSTIC AND THERMAL COUPLING MODEL FOR
TEMPERATURE RISE IN MR-GUIDED HIFU
C. Borrazzo M. Carnì E. Di Castr S. Pozzi B. Caccia G. Borasi
A. Napoli a Department of Molecular Medicine, Sapienza University of Rome,
Rome, Italy; b Medical Physics unit, Policlinico Umberto I, Rome, Italy;
c Department of Technology and Health, Istituto Superiore di Sanità and INFN,
Rome, Italy; d Bicocca University of Milan, Milan, Italy; e Department of
Radiological Sciences, Oncology and Pathology, Sapienza University of Rome,
Rome, Italy
Introduction: The success of MRgFUS therapy relies on the accuracy of
thermal mapping of sonication, which is obtained with guided MR. Un-
fortunately, inhomogeneities in the medium of propagation can cause
significant distortion of the ultrasound beam, resulting in changes in
focal-zone amplitude, location and shape. An adequate description of
absorption, diffraction and nonlinear phenomena can be necessary to
attenuate the limits of the technique. The aim of this study is to evaluate
a quantitative comparison between MRgFUS experimentally measured
thermal fields and that obtained using finite element method with Monte
Carlo (MC).
Materials and Methods: Essentially, the method employs MC integration
to evaluate the solution of the nonlinear Khokhlov–Zabolotskaya–Kuznetsov
(KZK) equation. The method can be used for complicated geometries, and
it is well suited to parallelization. The method is validated against exper-
imental temperature measurements on a homemade phantom gel, using
fluorotopic thermometer.
Results: The results of modeling obtained by both codes are compared
with each other and with known experimental data, and are found to
be in a good agreement. The analytical temperature solution used for
temperature-based parameter estimations assumes a radial Gaussian heating
pattern and that axial conduction and perfusion effect are negligible.
The results demonstrated that the nonlinear model absorption should be
taken into account in the evaluation of temperature rise for materials
sonicated.
Conclusion: The method is well suited to be used in applications where
flexibility and rapid computation time are crucial, in particular clinical HIFU
treatment planning. The development of optimized HIFU treatments is useful
to control the ablation of the target tissues and to improve patient safety.
It also potentially reduces the overall treatment duration and exposure to
non-target tissues thanks to a better understanding of acoustic and thermal
wave propagation.
e125 Abstracts/Physica Medica 32 (2016) e124–e134