Evaluation of the Imaging Properties of a CT Scanner with the

Adaptive Statistical Iterative Reconstruction Algorithm

Noise, Contrast and Spatial Resolution Properties of CT Images Reconstructed at

Different Blending Levels

Patrizio Barca

1,3

, Marco Giannelli

, Maria Evelina Fantacci

1,3

and Davide Caramella

Department of Physics, University of Pisa, Pisa, Italy

Unit of Medical Physics, Pisa University Hospital “Azienda Ospedaliero-Universitaria Pisana”, Pisa, Italy

INFN, Pisa Section, Pisa, Italy

Department of Radiology, Pisa University Hospital “Azienda Ospedaliero-Universitaria Pisana”, Pisa, Italy

Keywords: CT Iterative Reconstruction, ASIR, Image Quality, NPS, CNR, MTF.

Abstract: X-ray Computed Tomography (CT) is an essential imaging technique for different diagnostic and

therapeutic tasks. However, ionizing radiation from CT scanners represents the largest source of medical

exposure for the population of industrialized countries. In order to reduce CT dose during patient

examination, iterative reconstruction algorithms have been developed to help existing dose reduction

methods. In this paper, we studied the image quality performance of a 64-slice CT scanner (Optima CT660,

GE Healthcare, Waukesha, WI, USA) that implements both the conventional filtered back-projection (FBP)

and the Adaptive Statistical Iterative Reconstruction (ASIR, GE Healthcare, Waukesha, WI, USA)

algorithm. In order to compare the performance of these two reconstruction technologies, CT images of the

Catphan

504 phantom were reconstructed using both conventional FBP and ASIR with different

percentages of reconstruction from 20% to 100%. Noise level, noise power spectrum (NPS), contrast-to-

noise ratio (CNR) and modulation transfer function (MTF) were estimated for different values of the main

radiation exposure parameters (i.e. mAs, kVp, pitch and slice thickness) and contrast objects. We found that,

as compared to conventional FBP, noise/CNR decreases/increases non-linearly up to 50%/100% when

increasing the ASIR blending level of reconstruction. Furthermore, ASIR modifies the NPS curve shape (i.e.

the noise texture). The MTF for ASIR-reconstructed images depended on both tube load and contrast level,

whereas MTF of FBP-reconstructed images did not. For lower tube load and contrast level, ASIR offered

lower performance as compared to conventional FBP in terms of reduced spatial resolution and MTF

decreased with increasing ASIR blending level of reconstruction.

1 INTRODUCTION

In the last years, concerns about ionizing radiations

exposure due to computed tomography (CT)

technology have lead to develop strategies to

optimize CT procedures (tube current modulation,

automatic exposure control, advances in detection

technology etc.) (McNitt-Gray, 2002). A promising

approach for dose reduction is represented by the

improvement of image reconstruction algorithms,

that, in contrast with FBP algorithm, take into

account a model of the imaging system to describe

the different physical aspects of image acquisitions.

This may offer the opportunity to reduce image

noise with respect to FBP reconstructions. Because

of the strict correlation between radiation exposure

and image noise, IR algorithms can be employed

when CT data at reduced tube load product and/or

tube potential (i.e. at reduced radiation exposure) are

acquired (Willemink et al., 2013; Beister et al.,

2012). Therefore, IR algorithms can be used with

reduced radiation exposure without significantly

affecting the diagnostic image quality with respect to

conventional FBP (Willemink et al., 2013; Beister et

al., 2012).

In this study, our attention was focused on a 64-

slice CT scanner (Optima CT660, GE Helathcare,

Waukesha, WI, USA) which implements both the

conventional FBP algorithm and the Adaptive Statis-

200

Barca P., Giannelli M., Fantacci M. and Caramella D.

Evaluation of the Imaging Properties of a CT Scanner with the Adaptive Statistical Iterative Reconstruction Algorithm - Noise, Contrast and Spatial Resolution Properties of CT Images

Reconstructed at Different Blending Levels.

DOI: 10.5220/0006240802000206

In Proceedings of the 10th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2017), pages 200-206

ISBN: 978-989-758-216-5

tical Iterative Reconstruction (ASIR, GE Healthcare,

Waukesha, WI, USA) algorithm (Willemink et al.,

2013; Beister et al., 2012; Argaud, 2009).

ASIR works on raw data space modelling the

fluctuations in the projection measurement and the

noise characteristics of the scanned object (Argaud,

2009). Furthermore, ASIR offers the possibility of

blending with FBP at various levels, from 0%

(conventional FBP) to 100% ("pure" IR).

Previous works have shown that the noise

reduction in the ASIR-reconstructed images is

accompanied by changes in the noise texture as

compared to FBP reconstruction, and the spatial

resolution can vary as a function of dose and

contrast (Richard et al., 2012; Samei et al., 2015;

Miéville et al., 2013).

In our study, we quantitatively assessed noise

level, noise power spectrum (NPS), contrast-to-noise

ratio (CNR) and modulation transfer function (MTF)

using different ASIR blending levels of reconstru-

ction and a wide range of the main radiation

exposure parameters (i.e. tube load, tube potential,

pitch, slice thickness) values as well as different

contrast objects.

The paper is organised as follows: after this brief

introduction, we describe the materials and methods

in which we present the image acquisition protocols

and the adopted methodology for data analysis; in

the next section, we describe the results in terms of

the main image quality parameters analysed (noise

level, NPS, CNR and MTF); then, we dedicate a

section to the discussion of our results and finally we

suggest our conclusions.

2 MATERIALS AND METHODS

2.1 Scanner and Phantom Acquisition

Images of the Catphan

504 phantom (The Phantom

Laboratory, NY, USA) were acquired with a 64-slice

CT scanner (Optima CT660, GE Healthcare

Waukesha, WI, USA). This phantom is composed of

4 modules with cylindrical shape (internal diameter

of 15 cm). We employed the CTP486 module (a

homogeneous water-equivalent module) and the

CTP404 module (composed of many inserts of

different materials in a water-equivalent background

- nominal CT Hounsfield’s units (HU) of the inserts

are reported in the Catphan

504 Manual).

For noise analysis, images of the CTP486

module were acquired varying the main acquisition

parameters on a range of values as reported in Table

1. For spatial resolution evaluation, the CTP404

module was scanned across a range of acquisition

parameters as reported in Table 2. The contrast-to-

noise ratio (CNR) was computed from a subset of

the images used in spatial resolution analysis. All

images were reconstructed by using both

conventional FBP and ASIR with different blending

levels of reconstruction (20%, 40%, 60%, 80%,

100%).

2.2 Data Analysis

Image data analysis was performed using ImageJ

(Wayne Rasband, National Institute of Health, USA)

and OriginPro 9.0 (OriginLab Corporation, MA,

USA) software packages.

2.2.1 Noise

Noise properties of ASIR reconstructed images were

evaluated by measuring the standard deviation (SD)

of HU values on a circular region of interest (ROI)

(4.5 cm diameter) centered in the images. In

addition, the noise power spectrum (NPS) was

calculated from images acquired using a subset of

the exposure parameters in Table 1 (tube load 112

mAs, tube potential 120 kVp, slice thickness 2.5

mm, pitch 0.984). We computed the 3D NPS

(Siewerdsen et al., 2002; Verdun et al., 2015;

Friedman et al., 2013) considering an ensemble of

20 volumes of interest (VOIs) selected from 19

slices. From each VOI we calculated the 3D NPS

and then we made the ensemble average. In order to

obtain a radial representation of NPS the f

=0 plane

of the 3D NPS was selected and a radial average was

performed (Friedman et al., 2013).

2.2.2 Spatial Resolution

Spatial resolution properties of ASIR reconstructed

images were assessed through the calculation of the

modulation transfer function (MTF).

The MTF analysis was performed at different

radiation exposure and considering 6 different

inserts (air, PMP, LDPE, polystyrene, delrin and

teflon). We adopted the circular edge method

(Richard et al., 2012; Samei et al., 2015; Friedman et

al., 2013; Takenaga et al., 2015) to compute the

MTF. We acquired an ensemble 7 distinct images

with the same scanning parameters and the MTF

curves corresponding to the same insert were

averaged. The uncertainty of the MTF estimation

was obtained as the standard deviation of the above

7 measurements.

Evaluation of the Imaging Properties of a CT Scanner with the Adaptive Statistical Iterative Reconstruction Algorithm - Noise, Contrast and

Spatial Resolution Properties of CT Images Reconstructed at Different Blending Levels

201

2.2.3 CNR

The CNR was obtained from 7 repeated acquisitions,

using a subset of the acquisition parameters reported

in Table 2 (tube load 140 mAs, 84 mAs, 56 mAs and

28 mAs). In this analysis, we considered polystyrene

(low contrast), LDPE (medium contrast) and teflon

(high contrast) inserts. The CNR was estimated as

follows:

CNR=(HU

object

-Hu

bkg

)/√(σ

object

+ σ

bkg

) (1)

where HU

object

/σ

object

and HU

bkg

/σ

bkg

are the

mean/standard deviation of HU values in a circle

ROI in the insert and background region,

respectively. For each insert, the CNR values were

calculated from the mean value and its standard

deviation (uncertainty) across the repeated

acquisitions.

Table 1: Acquisition protocol/parameters for noise

analysis.

Acquisition mode: helical

Tube load (mAs): 28, 42, 56, 70, 84, 98, 112

Tube potential (KV): 80, 100, 120, 140

Tube rotation time (s): 0.7

Slice thickness (mm):

0.625, 1.25, 2.5, 3.75, 5,

7.5

Collimation along

longitudinal direction

(mm):

Pitch: 0.516, 0.984, 1.375

Number of reconstructed

slices

Table 2: Acquisition protocol/parameters for spatial

resolution and CNR analysis.

Acquisition mode: helical

Tube load (mAs):

28, 56, 84, 112, 140, 168,

196, 224

Tube potential (KV): 120

Tube rotation time (s): 0.7

Slice thickness (mm): 2.5

Collimation along

longitudinal direction

(mm):

Pitch: 0.984

Number of reconstructed

slices

3 RESULTS

3.1 Noise

Figure 1 shows noise with varying tube load, tube

potential, slice thickness and pitch for conventional

FBP algorithm and ASIR with different blending

levels of reconstruction (20%, 40%, 60%, 80%,

100%). Noise decreased non-linearly with the

increase of ASIR blending level of reconstruction as

well as with increasing tube potential/tube load/slice

thickness. On the other hand, noise increased with

the increase of pitch value.

NPS results are reported in figures 2 and 3. ASIR

algorithm acts as a low pass filter whose effect

increases with the increase of blending level of

reconstruction (Fig. 3).

3.2 Spatial Resolution

The spatial resolution results are reported in Figures

4 and 5. MTF of ASIR-reconstructed CT images

varied with the contrast level (Fig. 4), especially at

lower tube load. In particular, MTF decreased with

decreasing contrast level. We also verified that MTF

of FBP-reconstructed CT images was substantially

independent of the contrast level and tube load.

Furthermore, for ASIR-reconstructed CT images and

lower contrast level, MTF decreased with decreasing

tube load (Fig. 5 A). While for high contrast objects

or high mAs values ASIR preserves the spatial

resolution offered by FBP reconstruction , for lower

tube load and contrast level, MTF of ASIR-

reconstructed CT images was lower than MTF of

conventional FBP-reconstructed images, and the

first decreased with increasing blending level of

reconstruction (Fig. 5 B).

3.3 CNR

CNR results are reported in Table 3. CNR values

increased with increasing tube load. Also, for each

tube load value and insert (teflon, LDPE, polystyre-

ne), CNR values of ASIR-reconstructed CT images

were higher than CNR values of conventional FBP-

reconstructed CT images and increased non-linearly

with increasing blending level of reconstruction.

BIODEVICES 2017 - 10th International Conference on Biomedical Electronics and Devices

202

Table 3: CNR values (mean ± standard deviation) of teflon (high contrast)/ LDPE (medium contrast)/polystyrene (low

contrast) inserts of the Catphan-CTP404 module for FBP- and ASIR-recontructed CT images with different blending levels

(20%, 40%, 60%, 80%, 100%). Tube load ranges from 140 mAs to 28 mAs and acquisition parameters are reported in

Table 2.

Teflon LDPE Polystyrene

140 mAs

FBP 59.7 ± 3.9 14.8 ± 0.4 10.0 ± 0.6

ASIR 20% 65.8 ± 4.8 16.6 ± 0.4 11.2 ± 0.6

ASIR 40% 72.9 ± 5.8 18.8 ± 0.5 12.7 ± 0.8

ASIR 60% 81.1 ± 7.2 21.6 ± 0.7 14.6 ± 1.0

ASIR 80% 90.6 ± 9.0 24.9 ± 1.1 16.8 ± 1.2

ASIR 100% 100.5 ± 11.0 28.6 ± 1.6 19.2 ± 1.5

84 mAs

FBP 41.4 ± 3.6 11.1 ± 0.3 7.4 ± 0.5

ASIR 20% 52.7 ± 4.1 13.1 ± 0.4 8.8 ± 0.6

ASIR 40% 58.6 ± 5.7 14.8 ± 0.4 10.0 ± 0.7

ASIR 60% 65.5 ± 7.0 16.9 ± 0.8 11.5 ± 1.0

ASIR 80% 73.6 ± 8.8 19.4 ± 1.0 13.3 ± 1.3

ASIR 100% 82.5 ± 10.2 22.2 ± 1.5 15.4 ± 1.4

Teflon LDPE Polystyrene

56 mAs

FBP 36.4 ± 3.1 9.8 ± 0.3 6.1 ± 0.4

ASIR 20% 43.6 ± 4.4 11.1 ± 0.3 7.1 ± 0.6

ASIR 40% 48.5 ± 5.2 12.6 ± 0.4 8.0 ± 0.7

ASIR 60% 54.4 ± 6.9 14.4 ± 0.5 9.1 ± 0.9

ASIR 80% 61.2 ± 8.4 16.5 ± 0.9 10.5 ± 1.0

ASIR 100% 68.8 ± 9.8 19.0 ± 1.4 12.0 ± 1.3

28 mAs

FBP 26.8 ± 2.8 6.8 ± 0.3 3.4 ± 0.3

ASIR 20% 31.2 ± 3.0 .3 ± 0.3 4.9± 0.4

ASIR 40% 35.1 ± 4.3 8.3 ± 0.3 5.5 ± 0.4

ASIR 60% 39.8 ± 6.3 9.5 ± 0.5 6.3 ± 0.6

ASIR 80% 45.5 ± 7.0 10.9 ± 0.7 7.3 ± 0.8

ASIR 100% 51.9 ± 7.4 12.6 ± 0.7 8.4 ± 1.0

Figure 1: Noise (standard deviation) for conventional FBP algorithm and different ASIR blending levels of reconstruction

(20%, 40%, 60%, 80%, 100%) with varying tube potential (tube load 112 mAs, pitch 0.984, slice thickness 2.5 mm) (panel

A), tube load (tube potential 120 kVp, pitch 0.984, slice thickness 2.5 mm) (panel B), pitch (tube potential 120 kVp, tube

load 112 mAs, slice thickness 2.5 mm) (panel C) and slice thickness (tube potential 120 kVp, tube load 112 mAs, pitch

0.984) (panel D).

Evaluation of the Imaging Properties of a CT Scanner with the Adaptive Statistical Iterative Reconstruction Algorithm - Noise, Contrast and

Spatial Resolution Properties of CT Images Reconstructed at Different Blending Levels

203

Figure 2: Visualization of the 3D NPS with respect to the

plane fz = 0 for conventional FBP (A) and ASIR with a

blending level of reconstruction of 100% (B). (The

maximum spatial frequency was determined by applying

the Nyquist sampling criterion). Images were acquired at

120 kVp, 112 mAs, slice thickness 2.5 and pitch 0.984.

Figure 3: Example (same parameters as Fig. 2) of radial

NPS for conventional FBP algorithm and ASIR algorithm

with different blending levels of reconstruction.

Figure 4: MTF for ASIR-reconstructed CT images (blending level of reconstruction of 100%) and low (28 mAs) (panel

A)/high (224 mAs) (panel B) tube load, with varying contrast level (polystyrene, LDPE, delrin, PMP, teflon, air).

Acquisition parameters are reported in Table 2.

Figure 5: MTF for ASIR-reconstructed CT images (blending level of 100%) and low (polystyrene) contrast level, with

varying tube load (28 mAs, 56 mAs, 84 mAs, 112 mAs, 140 mAs, 168 mAs, 196 mAs, 224 mAs), (panel A). MTF for low

(28 mAs) tube load and low (polystyrene) contrast CT images, with varying reconstruction methods (FBP (black) , ASIR

with 20% (red), 40% (blue), 60% (fuchsia), 80% (green) and 100% (violet) of blending level of reconstruction), (panel B).

Acquisition parameters are reported in Table 2.

BIODEVICES 2017 - 10th International Conference on Biomedical Electronics and Devices

204

4 DISCUSSION

In this quantitative phantom study, CT image quality

performance of two different reconstruction techno-

logies have been carefully evaluated.

We comprehensively characterised the physical

properties (in terms of several quality parameters

such as noise, NPS, MTF and CNR) of the ASIR-

reconstructed CT images using different blending

levels of reconstruction in a number of experimental

designs.

Our findings confirm the dose reduction potential

of ASIR (Richard et al., 2012; Samei et al., 2015;

Miéville et al., 2013; Smith, 2014 et al.; Brady et al.,

2012; McCollough et al., 2015; Yanagawa et al.,

2010). As compared to conventional FBP,

noise/CNR decreases (Fig. 1)/increases (Table 3) up

to 50%/100% when using ASIR. Also, ASIR does

not modify the typical noise dependence on the

acquisition parameters. Furthermore, the noise and

CNR vary non-linearly with the ASIR blending level

of reconstruction. In addition, NPS analysis (Fig. 3)

shows that ASIR acts as a low-pass filter and

modifies noise texture: for ASIR-reconstructed CT

images, the frequency of the maximum of the NPS

curve shifts non-linearly toward lower frequencies

with increasing blending level of reconstruction, in

agreement with the results of previous studies

(Samei et al., 2015; Miéville et al., 2013).

We assessed the spatial resolution by estimating

the MTF at different contrasts and exposure values.

We found that, unlike conventional FBP, the MTF

decreases with decreasing contrast and tube load. It

should be noted that, for lower contrast and tube

load, the MTF of ASIR-reconstructed images is

lower than that of FBP-reconstructed images, and

decreases with increasing blending level of

reconstruction.

Recent studies have shown that these effects may

affect low-contrast resolution and thus may

influence the performance of automatic contour

detection software (Precht et al., 2016).

When compared to conventional FBP reconstru-

ction, ASIR allows for an improvement of image

quality in terms of reduced noise and increased CNR,

and hence a potential dose reduction in CT imaging

can be obtained while preserving diagnostic

capabilities. However, ASIR can modify noise

texture as well as affect spatial resolution at low

contrast and radiation exposure. For these reasons,

the optimal ASIR blending level of reconstruction

(i.e. the best trade-off between image quality and

dose reduction) should be assessed for each specific

application through quantitative as well as subjective

analysis.

Because of the noise reduction and CNR

increment offered by ASIR, CT examinations can be

performed at reduced radiation exposure levels.

However, in order to avoid potential effects of losses

in spatial resolution, which are inherent to ASIR and

may reduce the diagnostic value of CT images, the

optimal blending level of reconstruction should be

assessed for each specific clinical application.

In this regard, CT follow-up examinations (Lim

et al., 2016; Precht et al., 2016) and screening

programs could benefit of this new reconstruction

technology. In particular, the use of ASIR in CT

screening programs – aimed at detecting small

contrast lesions with low dose – should be carefully

evaluated.

5 CONCLUSIONS

A relevant noise reduction and CNR increment in

CT images are achieved with the ASIR algorithm

with respect to the conventional FBP reconstruction

in different experimental designs. For this reason,

the iterative reconstruction approach represents an

effective method for optimizing dose in CT imaging.

However, for low dose and low contrast acquisitions

(typical for instance of screening programs) ASIR

can provide lower performance, in terms of reduced

spatial resolution capabilities, as compared to

conventional FBP, and its use, along with the choice

of the optimal blending level of reconstruction,

should therefore be carefully evaluated.

Moreover, this work lays the basis for further

studies on CT imaging with ASIR. In particular, our

recent interests are focused on the performance of a

Computed Aided Detection (CAD) system with

ASIR-reconstructed clinical images of the lung. In

fact, the CAD system has been developed by taking

into account the FBP-related appearance of the

images and therefore, an investigation on the CAD

response to the ASIR-reconstructed images could be

of considerable interest.

ACKNOWLEDGEMENTS

We would like to thank Prof. Duccio Volterrani,

Prof. Antonio Claudio Traino and Dr. Davide

Giustini for supporting this work.

Evaluation of the Imaging Properties of a CT Scanner with the Adaptive Statistical Iterative Reconstruction Algorithm - Noise, Contrast and

Spatial Resolution Properties of CT Images Reconstructed at Different Blending Levels

205

REFERENCES

McNitt-Gray M. F., 2002. AAPM/RSNA Physics tutorial

for residents: Topics in CT. Radiation dose in CT.

Radiographics; 22(6):1541–53.

Willemink M. J., de Jong P. A., Leiner T., de Heer L. M.,

Nievelstein RA, Budde R. P., Schilham AM., 2013.

Iterative reconstruction techniques for computed

tomography Part 1: Technical principles. European

Radiology; 23:1623–31.

Beister M., Kolditz D., Kalender W. A., 2012. Iterative

reconstruction methods in X-ray CT. Physica Medica;

28(2):94-108.

Argaud C. ASIR: a new reconstruction technique to lower

dose without compromise, ASIR whitepaper. GE

Healthcare (Waukesha, WI, USA).

Richard S., Husarik D. B., Yadava G., Murphy S. N.,

Samei E., 2012. Towards task-based assessment of CT

performance: system and object MTF across different

reconstruction algorithms. Medical Physics; 39(7):

4115-22.

Samei E. and Richard S. Assessment of the dose reduction

potential of a model-based iterative reconstruction

algorithm using a task-based performance metrology.

Medical Physics 2015; 42(1):314-23.

Miéville F. A., Gudinchet F., Brunelle F., Bochud F. O.,

Verdun F. R., 2013. Iterative reconstruction methods

in two different MDCT scanners: Physical metrics and

4-alternative forced-choice detectability experiments –

A phantom approach. Physica Medica; 2:99-110.

Siewerdsen J. H., Cunningham I. A., Jaffray D. A., 2002.

A framework for noise-power spectrum analysis of

multidimensional images. Medical Physics;

29(11):2655-71.

Verdun F. R., Racine D., Ott J. G., Tapiovaara M. J., Toroi

P., Bochud F. O., Veldkamp W. J., Schegerer A.,

Bouwman R. W., Giron I. H., Marshall N. W.,

Edyvean S., 2015. Image quality in CT: From physical

measurements to model observers. Physica Medica;

31(8):823-43.

Friedman S. N., Fung G. S. K., Siewerdsen J. H., Tsui B.

M. W., 2013. A simple approach to measure computed

tomography (CT) modulation transfer function (MTF)

and noise-power spectrum (NPS) using the American

College of Radiology (ACR) accreditation phantom.

Medical Physics; 40(5):051907.

Takenaga T, Katsuragawa S, Goto M, Hatemura M,

Uchiyama Y, Shiraishi J., 2015. Modulation transfer

function measurement of CT images by use of a

circular edge method with a logistic curve-fitting

technique. Radiological Physics and Technology;

8:53–59.

Smith E. A., Dillman J. R., Goodsitt M. M., Christodoulou

E. G., Keshavarzi N., Strouse P. J., 2014. Model-based

iterative reconstruction: Effect on patient radiation

dose and image quality in pediatric body CT.

Radiology; 270(2):526-34.

Brady S. L., Yee B. S., Kaufman R. A., 2012. Characteri-

zation of adaptive statistical iterative reconstruction

algorithm for dose reduction in CT: A pediatric

oncology perspective. Medical Physics; 39(9):5520-31.

McCollough C. H., Yu L., Kofler J. M., Leng S., Zhang Y.,

Li Z., Carter R. E., 2015.

Degradation of CT low-

contrast spatial resolution due to the use of iterative

reconstruction and reduced dose levels. Radiology;

276(2):499-506.

Yanagawa M., Honda O., Yoshida S., Kikuyama A., Inoue

A., Sumikawa H., Koyama M., Tomiyama N., 2010.

Adaptive statistical iterative reconstruction technique

for pulmonary CT: image quality of the cadaveric lung

on standard- and reduced-dose CT. Academic

Radiology; 17(10):1259-66.

Lim H., Chung M. J., Shin K. E., Hwang H. S., Lee K. S.,

2016. The impact of Iterative Reconstruction in Low-

Dose Computed Tomography on the Evaluation of

Diffuse Interstitial Lung Disease. Korean Journal of

Radiology; 17(6):950-60.

Precht H., Kitslaar P. H., Broersen A., Dijkstra J., Gerke

O., Thygesen J, Egstrup K, Lambrechtsen J., 2016.

Influence of Adaptive Statistical Iterative

Reconstruction on coronary plaque analysis in corona-

ry computed tomography angiography. Journal of

Cardiovascular Computed Tomography; 10(6):507-16.

BIODEVICES 2017 - 10th International Conference on Biomedical Electronics and Devices

206