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

77 / 146

Results: Image noise decreased (up to 50% as compared to FBP) when in-
creasing the percentage of ASIR reconstruction (behavior more relevant for
higher spatial frequencies). Only for low tube load (<56 mAs) and lower
contrast objects (with respect to phantom background – i.e. PMP, Delrin,
LDPE and Polystyrene) acquisitions, MTF analysis showed that, increasing
the ASIR percentage, the reconstructed images were characterized by an
appreciable reduction in spatial resolution, when compared to FBP-
reconstructed images.
Conclusion: When compared to FBP, ASIR allows a relevant noise reduc-
tion without appreciably affecting image quality, except for very low dose
and contrast acquisitions.
B.240
DOSIMETRIC AND QUALITATIVE EVALUATION OF CT LONGITUDINAL
AUTOMATIC TUBE CURRENT MODULATION USING IN-HOUSE PHANTOMS
D. Becci *, L. Gentile, W. Sartor. Medical Physics Unit, ASL CN1, Cuneo, Italy
Introduction: We have developed three different phantom designs to
perform several test methods for the purpose to study automatic tube
current modulation (ATCM) in CT.
Materials and Methods: The Z-axis dose modulation (Z-DOM without
ACS) present in the routine chest scan protocol of a Philips Brilliance 16
CT scanner was investigated. The first phantom used in this study, Adapted
Body Phantom (ABP), was obtained adapting a standard compact body/
head phantom pulling out half of the head module. ABP presents three
sections: hole, body and head respectively. The second, Recycled Step
Phantom (RSP), comprises three circular sections of different dimensions.
SP has a hole in the center to allow for the insertion of a pencil dose
probe. The third, Funnel Filled Phantom (FFP), was manufactured by filling
a large plastic funnel with water, which was then hermetically sealed.
Point dose measurements were performed for ABP and RSP at the center
in each one of the three different sections that compose both phantoms
using RTI CT Dose Profiler. The dose saving (DSV) was evaluated, with
and without Z-DOM comparing with the same values extracted from
Dicom header. ABP, RSP and FFP were scanned for measurements of noise
(SD of CT number).
Results: Maximum DSV for ABP, RSP and FFP was respectively 29%, 41% (cal-
culated) and 58% (extracted). The mean noise within images using Z-DOM
is worse than those without. ABP and RSP present a different average noise
among sections changing with set mAs/slice. These differences are less no-
ticeable when we consider total average noise. Also the system almost seems
not to reduce mAs/slice in the Hole section with respect to the Body section
in the ABP. In FFP however the noise remains constant with mAs/slice
set.
Conclusions: The system shows a real reduction in dose while maintain-
ing an adequate image quality. It seems less adequate in saving dose in
hollow regions scanned.
B.241
MONITORING OF DATA OF EXPOSURE OF COMPUTED TOMOGRAPHY
EXAMINATIONS BY ANALYSIS OF RADIATION DOSE STRUCTURED
REPORTS
G. Belli F. D’Eli A.M. Cimmino A.S. Curion C. Gasperi F. Zenone
C. Ciccotosto a U.O.C. Fisica Sanitaria, AUSL8, Arezzo, Italy; b U.O.C.
Radiodiagnostica, AUSL8, Arezzo, Italy
Introduction: The statistical analysis of the data of exposure is needed to
choose the best strategy to optimize the dose administered to patients.
We present goals that can be achieved with software developed in
house for monitoring the radiation dose indexes in computed tomogra-
phy studies.
Materials and Methods: In our hospital, 6 MSCT scanners (1× 64,32,4, 2×
16 slices) send the Radiation Dose Structured Report (RDSR) together with
images to the PACS.
A program, developed in C++, is used for automatic retrieval of the RDSR
on a daily basis, and possibly of a previous period.
Exposure data as kVp, mA, rotation time, beamwidth, pitch, CTDIvol, partial/
cumulative DLP and other descriptive parameters of the studies are
automatically extracted from the RDSR and finally stored in a relational da-
tabase of doses or exported as CSV/HTML formats.
First, controls on the accuracy and consistency of the acquisition settings
in relation to the examination protocols are performed.
Then, the statistical analysis of radiation dose indexes for the most common
examinations it is performed.
For both steps, SQL queries are executed on database.
Results: They have been retrieved more than 85,000 RDSRs.
Accuracy/consistency tests:
• CT 64 slices: beam collimation resulted wrong for all spiral acquisitions;
• mA and CTDIvol were recorded erroneously for a study by a factor ten
compared to the true values retrieved from the images.
The cross-analysis between Study Description (SD), Target and IEC CTDI
phantom showed a mismatch for about 20% of the anatomical regions
scanned relying on the exclusive use of the SD.
Optimization of the image–quality dose ratio by changing noise-related ac-
quisition parameters:Analysis of RDSR showed a reduction for CTDI-DLP
(75thpercentiles) of 18%–11% for chest/mdc, 20%–13% for abdomen/mdc.
Conclusions: Software that allows executing queries on aggregate data from
RDSR can be a useful tool to monitor acquisition protocols and dose index
involved in radiological examinations.
B.242
A NOVEL IMAGEJ MACRO FOR CATPHAN LOW CONTRAST MODULE
ANALYSIS: PRELIMINARY RESULTS
M. Biondi A. Bogi L.N. Mazzoni E. Vanz G.M. Belmont G. De Ott
R. Martini S.F. Carbone L. Guerrini A.E. La Rocca D. Guerrieri
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: In the last decades the number CT scans performed per year
has dramatically increased, so that protocol optimization, preferably based
on objective methods for image analysis, is nowadays mandatory. While
for high contrast resolution evaluation MTF could be used, there is not a
commonly accepted method for low contrast resolution evaluation. Catphan
phantoms are widely used in CT quality assurance. In this work an ImageJ
macro for CTP515 low contrast insert quantitative analysis has been
developed.
Material and Methods: CTP515 module contains 3 groups (contrast level
0.3%, 0.5% and 1%) of 9 cylindrical rods with diameters between 2 and
15 mm. Images were acquired with a CT scanner (GE, Discovery 750 CT),
varying tube voltage (kV = 100, 120, 140) and noise index (NI = 8, 12). Each
image was first filtered with the human visual-response function peaked
at four cycles per degree. Choosing a “signal known exactly/background
known exactly” approach ROIs was placed automatically. Test statistics
related to signal and to background were build using 69 images, than a dis-
crimination index d’ related to PC in a 2-AFC was calculated for each detail.
Minimum detectable detail size (MDDS) for each contrast level was evalu-
ated with a fit to the PC values obtained using 75% as threshold. Finally 3
readers scored each image so that macro results could be compared to
human perception.
Results: Macro results showed that image quality is correlated more with
NI than to kV value. MDDS for automated method were 1.65 mm (1%),
3.4 mm (0.5%) and 4 mm (0.3%). Human observer study showed inter-
reader variability and similar trends were found: MDDS were 3 mm (1%),
5 mm (0.5%) and 7 mm (0.3%).
Conclusions: The proposed ImageJ macro can be used for low contrast ob-
jective analysis in CT. Thanks to the ease of use and to the wide spread of
Catphan phantom it could be an useful tool in optimization works. Further
works on error estimation and on correlation with human observer results
are needed.
e72 Abstracts/Physica Medica 32 (2016) e71–e96