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

96 / 146

Ingegneria, Modena, Italy; f University of L’Aquila, Scuola Specializzazione Fisica
Medica, L’Aquila, Italy; g University of Perugia, Dip. Medicina, Perugia, Italy;
h University of Perugia, Dip. Fisica e Geologia, Perugia, Italy
Introduction: Personnel real time dosimetry has grown in the past years
as an important request to help medical staff to optimize their radiation
exposure during interventional radiology and hemodynamics proce-
dures. Often the absorbed dose could be reduced if appropriate modifications
of the implementation of procedures could be introduced, based on evi-
dence from collected dose-rate data correlated with medical staff position
with respect to the patient. Also real time dose monitoring could help
Medical Physicist in their day-by-day dosimetry measurements, allowing
automatic data recording and hence simplification of procedures. The INFN
RAPID project has developed a real-time dosimeter based on a new sensor,
capable of dose-rate measurement with a frequency of few Hz, worn by
medical staff during real procedures and calibrated on a certified X-ray beam.
Material and Methods: The RAPID wireless prototype dosimeter device has
as sensor element a CMOS VGA imager (3.5 × 2.5 mm2) sensitive from 10 keV
up to 100 keV X-ray photons. The prototype, with a dimension of
10 × 5 × 2 cm3 can be worn either on the chest than on the arm of the op-
erator. The wireless transmission could use both the EU and USA frequencies
reserved for medical devices. The device has been tested both in labora-
tory than in operating room during procedures. The control dosimetry during
all measurements has been done using TLD dosimeters.
Results: The prototype works continuously for more than 8 hours without
changing batteries, it has shown a linear response to dose-.rate up to
0.5 mGy/s, and a dose measurement uncertainty in the range of 5% for all
kind of interventional radiation procedures.
Conclusions: The RAPID real-time wireless dosimeter has been devel-
oped and tested and it could be now be used to study problems like the
ratio of absorbed dose due to fluorography respect to fluoroscopy during
a given type of procedure, and/or to start collecting data to optimize the
dose absorbed by medical personnel.
B.308
MEDICAL DEVICE EVALUATION FOR OPERATORS’ RADIATION
PROTECTION IN ANGIOGRAPHIC PROCEDURES OF CHRONIC TOTAL
OCCLUSION
R. Soavi *, P. Berardi, L. Baroni, G. Nobile, P. Sangiorgio, L. Pagan. Ospedale
Maggiore, AUSL Bologna, Bologna, Italy
Introduction: The purpose of this work was to evaluate the effectiveness
of using a particular medical device, the RADPAD drape with femoral access,
for the protection of the operators from the radiation scattered by the patient
during the interventional procedures of chronic total occlusion (CTO) of the
coronary artery, which entail long exposure times and high doses to patient
and operators.
Materials and Methods: A phantom consisting of chest, abdomen and some
PMMA slabs, to simulate the lower limbs, was used and scattered dose rates
were measured in some significant points, corresponding to the position
taken by the first and second operator during CTO procedures.
Measurements were performed on a dynamic flat panel angiographic system
Philips Allura FD20, in both fluoroscopy and fluorography modalities, at the
two most frequently used C-arm angles in clinical practice: LAO = 45° –
CRAN = 27° and RAO = 30° – CRAN = 35°.
Dose rates were compared, for each measurement point, under the fol-
lowing conditions: (a) in absence of the drape, (b) in presence of the RADPAD
drape, placed on the side of the abdomen phantom (as usual in clinical prac-
tice), and (c) with the drape rotated by 180°, to shield the radiation scattered
by the PMMA, i.e. by the side of the limb.
Results: The measured drape attenuation to the primary radiation was 76%
at 70 kV, according to the company specifications.
The shadow produced by the drape allows to shield operators’ arms and
hands: the attenuation of the scattered X-ray beam depends both on the
C-arm tilt of the angiographic system and on the drape position on the
phantom, varying from 30% to 75%.
Conclusions: In CTO procedures the drape, along with other radiation pro-
tection devices, is helpful in reducing the exposure near the operators’ hands,
especially if it is positioned on the patient limb side rather than the abdomen
one. Its use, however, shall be evaluated taking into consideration the high
cost involved.
B.309
COMPARISON AMONG THREE DIFFERENT APPROACHES FOR LOW
CONTRAST DETECTABILITY EVALUATION IN DIGITAL RADIOGRAPHY
C. Spadavecchia R. Villa C. Pasqual N. Paruccini A. Cresp a Scuola
di Specializzazione in Fisica Medica, UNIMI, Milano, Italy; b S.C. Fisica Sanitaria,
A.O. San Gerardo, Monza, Italy
Introduction: Contrast detail curves are commonly used to assess low con-
trast detectability (LCD), which refers to the ability of imaging small objects
having low contrast. The curve displays the minimum diameter detected
(‘threshold’) for several inserts of different contrast.
Materials and Methods: This work aimed at investigating three different
approaches for LCD assessment: (1) an automated one, based on the readout
of CDRAD images performed by the software package; a significance level
to define the threshold for detectability is selected by the user. (2) A sub-
jective one, involving human perception and decision criteria: 4 observers
identified the just visible details for each detail diameter of the CDRAD
image. (3) A statistical approach based on Chao’s method, considering the
distribution of mean values for many ROIs on a uniform region and as-
suming a normal distribution.
Two different devices for digital radiology have been considered: Acselerate
Fujifilm and DigitalDiagnost Philips in RQA5 exposure conditions.
For comparisons among different methods and systems, the Inverse Image
Quality Figure (IQFinv = 1/SUMi(diameteri*threshold_contrasti ) was
calculated.
Results: The best agreement between the statistical method and the CDRAD
objective analysis was found when significance levels between 1E-4 and
1E-5 were set. As for the visual evaluation, the most suitable values were
between 1E-8 and 1E-9 providing accordance with the CDRAD auto-
mated readout.
Relative performances of the three methods showed good agreement among
them.
Conclusions: the statistical approach proved to be less time-consuming,
independent on inter- and intra-observers variability and repeatable, thus
being a valid method for image quality (or LCD) assessment.
B.310
LOW CONTRAST DETECTABILITY IN DIGITAL MAMMOGRAPHY:
COMPARISON BETWEEN TWO SYSTEMS THROUGH A STATISTICAL
METHOD
C. Spadavecchia R. Villa C. Pasqual N. Paruccini A. Cresp a Scuola
di Specializzazione in Fisica Medica, UNIMI, Milano, Italy; b S.C. Fisica Sanitaria,
A.O. San Gerardo, Monza, Italy
Introduction: Contrast detail curves or LCD are usually determined with
the Artins CDMAM phantom, compatible with the European Guidelines rec-
ommendations (4thedition).
This study proposed a method allowing to overcome the disadvantage of
obtaining the contrast detail curve related to a unique thickness through
16 acquisitions of the CDMAM phantom (as reported in the European
recommendations).
Materials and Methods: LCD was estimated applying Chao’s statistical
method and using a dedicated phantom: an acetate sheet equipped with
a central uniform aluminium region (3 × 3cm2, thickness 0.49 mm) and an
aluminium step wedge (increasing thickness: 0.2–1 mm) for linear con-
version from pixel values to millimetres of aluminium.
Different slabs of PMMA (thickness: 2–7 cm) were positioned over the
phantom providing attenuation and scatter. Two digital mammography
systems were used to produce images of the phantom in full automatic tech-
nique: Siemens Mammomat Novation and Sectra Microdose Philips.
To provide a quantifiedmeasurement of image quality, the Inverse Image Quality
Figure was calculated: IQFinv = 1/SUMi(diameteri*threshold_contrasti), where
diameter and contrast refer to the contrast detail curve.
e91 Abstracts/Physica Medica 32 (2016) e71–e96