Task-based phantom evaluation of cardiac catheterization imaging modes

Physica Medica 46 (2018) 114–123

Contents lists available at ScienceDirect

Physica Medica journal homepage: www.elsevier.com/locate/ejmp

Original paper

Task-based phantom evaluation of cardiac catheterization imaging modes a,⁎

a

b

Lukmanda Evan Lubis , Leonard Airell Craig , Hilde Bosmans , Djarwani Soeharso Soejoko a b

T

a

Department of Physics, Faculty of Mathematics and Natural Sciences, Universitas Indonesia, Depok 16424, Indonesia Medical Imaging Research Center, Catholic University of Leuven, Leuven 3000, Belgium

A R T I C L E I N F O

A B S T R A C T

Keywords: Cardiac catheterization Dose Figure of merit Image quality Optimization

This study aimed to quantify the dose and quality of the preprogrammed imaging modes on two cardiac angiography devices (Philips Allura FD10 Clarity and Allura FD10) using a task-specific in-house phantom, and to discuss the appropriateness of the pre-programmed settings. A Figure of Merit (FOM), defined as the squared Signal Difference to Noise Ratio (SDNR) divided by Entrance Surface Air Kerma (ESAK), was calculated for phantom inserts with different sizes and concentrations of iodine, as well as tin foils. For the Allura FD10 Clarity device, the low dose fluoroscopic mode was found to be very dose efficient, while the available ciné modes should only be used for cases with high demand for contrast and temporal resolutions. For both devices, the basic beam spectrum of the low dose fluoroscopic mode should be explored for use on other imaging modes. Ciné modes for the Allura FD10 device differ only by their spatial resolution characteristics and have almost identical dose per frame. This study also found that tin may not be a suitable replacement for iodine for research purposes due to mismatching SDNR. The number of recommendations formulated for these two devices suggests that comparative dose and image quality tests of all routinely used imaging modes should be an obligatory part of the physicists’ acceptance testing.

1. Introduction Cardiac catheterization procedures are now a worldwide standard of practice for cardiac diseases [1,2]. The involvement of X-ray imaging for these complex procedures is potentially associated with deterministic effects such as skin damage, hair loss and severe necrosis above specific dose thresholds [3–8]. In addition, as with any other use of Xrays, there is a stochastic risk for neoplasm formation that increases with increasing exposure [4]. Dose monitoring is recommended for all patient groups, from children to obese patients and it is necessary to optimize the procedures. The required quality should be produced with a dose that is as low as reasonably achievable [4,9–11]. The patient’s detriment from cardiac procedures is mainly governed by the tube current-exposure time product (mAs), tube voltage (kVp), filtration, patient attenuation, and the body part examined. Today, the exposure parameters are controlled automatically for both fluoroscopy and ciné acquisition modes by an Automatic Exposure Rate Control (AERC) unit. Understanding the settings of such units plays a key role in determining the optimal imaging modes for given groups of patients and allows subsequent proper in-room training and feedback. Following the European and International Basic Safety Standards [12,13], the medical physics expert should be involved in these tasks, but detailed procedures have not been introduced, and experience is therefore limited. In present study, image quality was studied through the measurement of

⁎

signal from a dedicated, newly developed contrast-detail test object. Detectability of the inserts is influenced by several components of the imaging chain, including beam quality and dose rate. The specific quantitative parameter employed was Signal Difference-to-Noise Ratio (SDNR), calculated from iodinated contrast agents provided in a blood vessel shaped insert. Radiation detriment was expressed in terms of entrance skin dose (ESD) [14,15]. The Entrance Surface Air Kerma (ESAK) is related to the ESD through the back-scatter factor, which does not vary greatly for cardiac procedures, and is utilized in this study. This Figure of Merit was first introduced by Zamenhof et al. [16]. Aim of the study was to investigate the AERC modes on two angiographic X-ray systems used for adult cardiac procedures with the dedicated phantom, and plan, if needed, further optimization or teaching actions. 2. Materials and methods The study was performed on two Philips Allura FD10 devices (Philips Medical System, Best, the Netherlands), both in operation in the Invasive Cardiology Department, Pertamina Central Hospital, Jakarta. The first device was an Allura FD10 with a maximum flat detector size of 20 cm × 20 cm. The maximum nominal voltage applied to the MRC 200 0508 ROT-GS 1003 X-ray tube was 125 kVp with a maximum output of 1250 mA at 80 kV. The tube had a permanent filtration of 2.5 mm Al with additional filters of 0.1 mm Cu, 0.4 mm Cu, and 1.0 mm Al and two

Corresponding author. E-mail address: [email protected] (L.E. Lubis).

https://doi.org/10.1016/j.ejmp.2018.02.002 Received 12 October 2016; Received in revised form 19 December 2017; Accepted 1 February 2018 Available online 05 February 2018 1120-1797/ © 2018 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

Fig. 1. Arrangement of X-ray tube, phantom and image flat-panel detectors (FPD) (left) (not to scale), and real arrangement (right). (IC = ionization chamber, PMMA = poly methyl methacrylate).

Table 1 lists the exposure parameters for each mode for the 21.5 cm-thick object. Tube filtration was fixed for all pre-programmed modes. To accommodate the size of the in-house phantom, only the full FOV (25 cm) was used in the study. Magnification modes were not tested. The mean energy of the beam was calculated using the method described by Boone and Seibert (1997) [23], assuming a 3 mm of half-value layer for each beam.

nominal focal spot sizes of 0.5 mm and 0.8 mm. The device was equipped with Clarity software for dose reduction and image quality enhancement [17–19]. The second device was an older Allura FD10 with identical technical details to the above but without the Clarity software in place. Prior to the study, performance testing of both devices had been performed by the X-ray Compliance Test Service Division of the Medical Physics and Biophysics Laboratory of Universitas Indonesia, using the technical test procedures of the Indonesian Nuclear Energy Regulatory Agency. Additional tests in compliance with AAPM Report No. 70 were also performed [20]. A 5 cm thick PMMA phantom was constructed in house (Fig. 1), and contains tubes with different internal diameters of 1 mm, 2 mm, 4 mm, 6 mm, and 8 mm. They were filled with 300 mg/cc iodine contrast agents that were diluted to blood plasma concentrations of 14%, 16%, 18%, and 20%. The concentrations were determined as in our previous work [9], and had been optimized to represent typical injections in adult patients. The samples were prepared at the Biophysics Laboratory, Universitas Indonesia. At the mid-plane of the phantom, we put tin foils of 99.98% purity, a thickness of 127 μm and sizes as the iodine tubes. Tin was selected for its similarity with iodine in terms of K-edge [21].

2.3. Dose estimation and image quality analysis A Radcal®10X6-6 with 6 cc of active volume ionization chamber (IC) was used to measure the ESAK at the surface of the phantom. Ionization chamber was positioned to include backscatter [24]. This dosimeter does not influence the settings of the AERC. The ESAK values from all exposure modes were recorded for subsequent calculation of the FOM. The images of all exposure modes (at least three images for each imaging mode) were recorded in DICOM format and analysed using ImageJ software. Among the series of images produced per exposure, one frame was randomly selected for image quality assessment utilizing SDNR measurements. The SDNR is based on measurement of the mean pixel value and its standard deviation for rectangular regions of interest (ROIs) for each vessel diameter and tin foil, using corresponding areas of background. Typical positions of the ROIs are shown in Fig. 2. SDNR was calculated using Eq. (2), in which NB is the mean pixel value of the background ROI, located adjacent to the simulated vessels and the tin foil, and SDB is the standard deviation of the same ROI. The parameter NO is the object mean pixel value of the ROI located inside the simulated vessel or in the tin foil and SDO is the standard deviation of the same ROI [16,25–27].

2.1. Experimental setup The measurement geometry was set to represent the clinical PA projection (Fig. 1). The measured source-to-image distance (SID) was 117 cm, in accordance with typical clinical settings. For the same reasons, the anti-scatter grid was used during the entire work. The distance between the tube focus and dosimeter was 80 cm. Stacks of PMMA slabs of each 0.5 cm thick and the in-house phantom were positioned in a sandwich-like manner to reach the posterior-anterior (PA) thickness of a typical Indonesian adult chest, which is 21.5 cm, based on a study by Manuaba [22]. A total of 8 cm and 8.5 cm PMMA slabs were placed on and below the phantom, respectively. An air gap of about 15 cm was kept between the top of the phantom and the detector to simulate space for anesthetic devices in clinical condition.

SDNR =

NB−NO (SDO2 + SDB2 ) 2

FOM =

SDNR2 ESAK

(1)

(2)

At least three repetitions were made for both image quality and dose measurements and average values were retrieved. Figure of Merit values were calculated for all measurements along with the standard deviations. The attributed error is addressed as the combined uncertainty utilizing the standard deviation of the repeated measurements performed for each condition and dose measurement instrument uncertainty (shown in Figs. 3–7 as solid error bars).

2.2. Imaging parameters The exposures were made using both devices with the ‘adult cardio’ patient selection protocol using the AERC for fluoroscopic and ciné acquisition modes. There are three dose modes for each device in fluoroscopy, while in ciné a number of modes with various pulse-rates are selectable. 115

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

Table 1 Exposure parameters for both devices and each AERC mode of the ‘Adult Cardio’ protocol using a 21.5 cm thick phantom. Device

AERC Modes

Tube filtration

Tube voltage (kVp) min – max (mean)

Tube current (mA) min – max (mean)

Mean energy (keV)

Frame rate (fps)

Detector dose (nGy/frame)

Allura FD10 Clarity

Fluoro setting I (low dose mode) Fluoro setting II (medium dose mode) Fluoro setting III (normal dose mode) Ciné low Ciné normal Ciné boost Ciné coronary

0.4 mmCu + 1.0 mmAl 0.4 mmCu + 1.0 mmAl 0.1 mmCu 0.4 mmCu + 1.0 mmAl 0.1 mmCu + 1.0 mmAl No additional filter 0.4 mmCu + 1.0 mmAl

73.4–73.5 72.5–72.7 71.1–71.2 69.5–69.6 67.4–67.6 68.1–68.3 69.4–69.5

7 (7) 12 (12) 11 (11) 561–566 461–471 493–502 566–572

53.3 52.1 47.1 51.4 45.9 42.8 51.4

25 25 25 15 15 15 25

11.1 17.1 41.5 88.2 184.4 237.3 86.5

Fluoro setting I (low dose mode) Fluoro setting II (normal dose mode) Fluoro setting III (high dose mode) Ciné 15 fps Ciné 30 fps

0.4 mmCu + 1.0 mmAl 0.1 mmCu + 1.0 mmAl 0.1 mmCu + 1.0 mmAl No additional filter No additional filter

83.0 (83.0) 72.0–73.0 (72.3) 73.0 (73.0) 69.0 (69.0) 69.0 (69.0)

9.0 (9.0) 57.5 12.0 (12.0) 48.2 16.0 (16.0) 48.5 528.0–535.0 (531.7) 43.2 519.0–531.0 (524.3) 43.2

15 15 15 15 30

27.6 37.3 49.2 210.2 204.7

Allura FD10

(73.4) (72.6) (71.1) (69.5) (67.5) (68.1) (69.4)

(562.7) (464.0) (496.3) (568.4)

3. Results

3.2. Image quality (SDNR) assessment

The compliance test performed on both X-ray systems indicated that both devices were safe for use and comply with our national regulations. On all imaging modes enlisted in Table 1, the ESAK, SDNR, and FOM have been determined for each phantom vessel diameter and each iodine concentration.

Table 3 (Appendix) shows mean SDNR values for fluoroscopic modes (low, medium and normal dose) on the Allura FD10 Clarity device and Table 4 shows the same data for Allura FD10. The data demonstrate that results from the smallest objects (i.e., objects with 1 mm and 2 mm sizes) were sometimes deviating from trends such as ‘increasing SDNR for increasing contrast and object size’. This was attributed to their low SDNR values. Proper trends were observed above 18% iodine contrast agents. For the Allura FD10, the calculated SDNRs were as expected, with the low and high fluoroscopy mode yielding the lowest and highest SDNRs, respectively.

3.1. Dose measurement Entrance Surface Air Kerma measurements for all fluoroscopic and ciné modes on Allura FD10 Clarity and Allura FD10 systems are shown in Table 2, with the fluoroscopic dose rates for both devices significantly lower than the ciné dose rates. Fig. 3 illustrates these data graphically. Specifically for the Allura FD10 Clarity device, the fluoro setting II was about 2.20 μGy per frame higher (54.6%) than the fluoro setting I. In contrast, the fluoro setting II on the Allura FD10 device gave nearly twice the dose delivered by the fluoro setting I. Both systems use different dose increase steps. Two ciné modes of Allura FD 10 Clarity, i.e., 15 fps ‘low’ dose and 25 fps ‘coronary’, have similar exposure factors in terms of filtration, tube current, and voltage. Their respective ESAK rates (per frame) are not significantly different. The 15 fps ‘boost’ mode delivered a 160.7% higher dose than the 15 fps ‘normal’ dose, induced in part by the absence of aluminum filter on the 15 fps ‘boost’ mode. Similarly, the Allura FD10 device did not have any beam quality modulation for the different ciné acquisition modes. On the two ciné modes, the 15 fps mode delivered 154.0 ± 2.8 μGy per frame, while its counterpart 30 fps gave 150.0 ± 2.1 μGy per frame.

3.3. Figure of Merit (FOM) calculation The recorded SDNR and ESAK values were taken into account for FOM calculation using Eq. (2). Calculation results for the Allura FD10 Clarity system are presented in Table 3 and Table 5 for fluoroscopy and acquisition (ciné) modes respectively, while results for Allura FD10 system under fluoroscopic and ciné modes are displayed in Tables 4 and 6. 4. Discussions The purpose of this study was to assess the commonly used, preprogrammed exposure modes available on two angiographic X-ray systems using a task-based approach. For this purpose, we had designed, fabricated and used an in-house cardiac phantom with objects representing iodinated vessels of various iodine concentrations, as well as tin foils to represent 100% iodine concentration.

Fig. 2. Example images showing tin objects (left) and artificial, simulated vessels (right) with ROI positions for mean pixel value and standard deviation in the object and background. Rectangular ROIs of 20 mm × 1 mm were used for every object and background.

116

Physica Medica 46 (2018) 114–123

L.E. Lubis et al. 200

20 18

160

Fluoroscopic modes

fp s 30

N O R M AL fp s

15

15

fp s

H ig h

N or m

LO W

0

al

0

Lo w

20

N or m al

2

M ed iu m

40

Lo w

4

fp s

60

15

6

80

fp s

8

100

25

10

120

BO O ST

12

140

fp s

14

15

ESAK rate (PGy/frame)

16

ESAK rate (PGy/frame)

Allura Clarity Allura

180

Allura Clarity Allura

Cine modes

Fig. 3. Results of ESAK measurements for fluoroscopic and ciné modes. Dose rates in fluoroscopy, usually expressed in µGy/min, have been recalculated towards µGy/frame.

(setting III) delivering a significantly higher ESAK rate associated with the use of different filtration materials. Higher SDNR values at the lower dose level may also be due to other post processing. It has to be verified whether that would have any detrimental impact on temporal resolution. As seen in Table 4, there is an observable difference between the SDNR from ‘low’ dose fluoroscopy mode (setting I) and the other fluoroscopy modes (i.e. the ‘normal’ and ‘high’ dose modes; settings II and III) on the Philips Allura

4.1. Fluoroscopy The SDNR calculations for fluoroscopic modes in Allura FD10 Clarity (Table 3) demonstrated that fluoroscopy setting I yielded typically higher SDNR values in comparison with other modes, notwithstanding the fact that the mode delivered the lowest ESAK rate. The remaining modes, namely settings II and III, gave virtually similar SDNRs, with the normal dose

5

5 16%

4

4

3

3

SDNR

SDNR

14%

2 Setting I (low) Setting II (normal) Setting III (high)

1

2 Setting I (low) Setting II (normal) Setting III (high)

1

0

0 1

2

4 6 Object size (mm)

8

1

5

2

4 6 Object size (mm)

8

5 20%

4

4

3

3

SDNR

SDNR

18%

2

Setting I (low) Setting II (normal) Setting III (high)

1

2 Setting I (low) Setting II (normal) Setting III (high)

1

0

0 1

2

4

6

1

8

2

4

6

8

Object size (mm)

Object size (mm)

Fig. 4. SDNR calculations for fluoroscopic modes on the Allura FD10 device for various object sizes containing 14%, 16%, 18%, and 20% iodine concentration.

117

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

(a)

Allura (setting I) Allura (setting II) Allura (setting III) Allura Clarity (setting I) Allura Clarity (setting II) Allura Clarity (setting III)

4,0

2,0

(b)

Allura (setting I) Allura (setting II) Allura (setting III) Allura Clarity (setting I) Allura Clarity (setting II) Allura Clarity (setting III)

FOM (PGy-1)

SDNR

3,0

2,0

1,0

1,5

1,0

0,5

0,0

0,0

1

2

4

6

8

1

Object size (mm)

2

4

6

8

Object size (mm)

Fig. 5. Comparison of (a) SDNRs and (b) FOMs calculated for various object sizes and iodine concentration of 20% for Philips Allura FD10 and Philips Allura FD10 Clarity devices. In both devices, fluoroscopic flavor 1 presents higher FOM due to lower dose.

FD10 device, where the ‘low’ dose mode (setting I) has, as expected, the lowest SDNR. For the Philips Allura FD10 device, overlapping error bars (Fig. 4) in most data points between ‘normal’ and ‘high’ dose modes (i.e., settings II and III) indicate that changing between the two modes would have little to no effect on the image contrast. This is explained by identical filtrations for these imaging modes, namely 0.1 mmCu + 1.0 mmAl, the same tube voltage and a small change in tube load only, from 12 mA for normal (setting II) mode to 16 mA for the high dose (setting III) mode. The ‘low’ dose mode (setting I), on the other hand, provided images with typically lower contrast, as a consequence of the chosen parameters (extra filtration of 0.4 mm of Cu and 1 mm Al, 83 kV and a lower dose level). Lower current will provide less photon fluence to the detector and higher tube voltage presents a photon beam with a mean energy of up to 57 keV which is unmatched with iodine’s K-edge (around 33 keV) [21]. For most iodine concentrations and object sizes, the ‘low’ dose fluoroscopy mode of the Philips Allura FD10 Clarity (setting I) yielded the greatest FOM, particularly due to the low ESAK, a logical consequence of using copper filtration. FOM calculations aim to show optimal working points in terms of SDNR per unit ESAK. Some results are at first unexpected (Table 1): the ‘low’ dose mode FOM tends to excel over the FOMs of the remaining three dose modes. This finding could be used to adjust the higher dose towards the same beam qualities as in the low dose mode. This would have to be discussed with the manufacturer, as other limitations may apply such as total tube output. Today, in practice, the clinical use of the ‘low’ dose fluoroscopy mode on the Philips Allura FD10 (setting I) should be encouraged when the object of interest is relatively small (≤2 mm in size), and when the image contrast is not of paramount interest or decision-afflicting nature. Additionally, the use of ‘normal’ dose mode (setting II) over ‘high’ dose mode (setting III) for all clinical use could also be recommended. The use of ‘high’ dose fluoroscopy mode on the Allura FD10 device without Clarity feature should be discouraged. Different dose levels with the beam quality similar to the ‘low’ dose mode should be provided. It is also of interest to compare Philips Allura FD10 and Philips Allura FD10 Clarity devices. We had assumed at first that since the systems have identical hardware, the use of SDNR across the systems is valid. It is, however, possible that differences in image processing are not properly included in our present phantom and SDNR calculations. In addition, SDNR represents only a large area contrast. It does not include a measure of sharpness. As far as we know, there are no current international guidelines that assist in the testing of image processing part during the acceptance testing of intervention X-ray units by medical physicists. Most optimization work in radiological imaging uses also SDNR rather than more complicated measures. The method explored in a recent study by Dehairs et al. is

8

99.98% Tin 20% Iodine

SDNR

6

4

2

0 A

B

C

D

E

F

Fluoroscopic imaging modes Fig. 6. Comparison of SDNRs calculated for 8 mm sized vessel simulations with 20% iodine concentration with those of 99.98% tin under fluoroscopic modes of the Philips Allura FD10 Clarity (A ‘low’, B ‘medium’, C ‘normal’) and Philips Allura FD10 (D ‘low’, E ‘normal’, F ‘high’). 14

99.98% Tin 20% Iodine

12

SDNR

10

8

6

4

2

0 A

B

C

D

E

F

Acquisition (cine) imaging modes Fig. 7. Comparison of SDNR calculated from 8 mm sized object with 20% iodine concentration with those of 99.98% tin under acquisition (ciné) modes of Philips Allura FD10 Clarity (A, B, C, D) and Philips Allura FD10 (E, F). The modes are namely; (A) Ciné Low, (B) Ciné Normal, (C) Ciné Boost, (D) Ciné coronary, (E) Ciné 15 fps, and (F) Ciné 30 fps.

118

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

provide similar SDNRs, dose per frame, and FOM. The ‘ciné coronary’ mode delivers higher dose rate when expressed per second, by around 67%. Therefore, its use should be exclusively limited for fast moving objects that demand higher temporal resolution. Since only two ciné acquisition modes were available for adult patient protocols on the Allura FD10 device, making proper choices is currently not difficult. As a preliminary notice, the use of both ciné modes should be encouraged only after adjustments are made to include copper filtration. In their current state, the two ciné modes do not show any difference for the tests performed with our phantom. This can be explained from Table 1, where, other than the pulse rates, no distinction was observed on the radiologic parameters. This fact has been observed before [9]. However, while the doses per frame are similar, it should be noted that the dose increases when expressed per unit time, as expected from the pulse rate. While this is logical, it may not be known to the operators during routine practice. It is an example of a measurement result that should be communicated to the operators in the room. Moreover, as Table 6 has shown, the two modes have very similar ESAK rates when expressed per frame, and the selection of acquisition (ciné) mode should be based on the required temporal resolution. That is, the ‘ciné 15 fps’ mode should be employed when high temporal resolution is not clinically mandatory. Additionally, from Tables 5 and 6, it is shown that some degree of knowledge on the size of the objects to be imaged is required. For present settings, the SDNR values of smaller objects (2 mm or less in size) tended to give inconsistent values. It has to be verified whether these values are lower than those on other systems and whether any clinical consequences have been observed.

Table 2 ESAK rate measurements for various imaging modes. Device

Allura FD10 Clarity

Allura FD10

Imaging modes

ESAK rate µGy/s

µGy/frame

Fluoro setting I (low dose mode) Fluoro setting II (medium dose mode) Fluoro setting III (normal dose mode) Ciné 15 fps low Ciné 15 fps normal Ciné 15 fps boost Ciné 25 fps coronary

100.03 ± 0.19

4.00 ± 0.01

155.0 ± 1.2

6.20 ± 0.04

375.0 ± 6.9

15.0 ± 0.2

478.5 ± 9.0 1000 ± 17 2608 ± 26 782 ± 4

31.9 ± 0.6 66.7 ± 1.1 173.9 ± 1.8 31.3 ± 0.2

Fluoro setting I (low dose mode) Fluoro setting II (normal dose mode) Fluoro setting III (high dose mode) Ciné 15 fps Ciné 30 fps

101.9 ± 0.4

6.79 ± 0.03

202.5 ± 5.7

13.5 ± 0.4

267.0 ± 9.3

17.8 ± 0.6

2310 ± 42 4500 ± 62

154.0 ± 2.8 150.0 ± 2.1

promising in this regard, but not applicable without access to unprocessed data [28]. Present study used realistic test objects, yet did not test the effects of vessel motion. Fig. 5 shows the SDNRs for the Philips Allura FD10 and Philips Allura FD10 Clarity units for 20% iodine contrast agents over all object sizes. The concentration of 20% was chosen as the most probable concentration used clinically, making it a suitable candidate test insert for general purpose testing of angiography and catheterization devices. Our phantom tests demonstrated that for fluoroscopy mode I there was a difference of SDNR between the two devices, where the Allura FD10 Clarity device tended to produce images with slightly lower SDNR (on average 27% lower) at lower ESAK values (41%), except for larger objects. There was no significant difference between fluoroscopy modes II and III for any of the two devices.

4.3. Tin detectability As the use of tin as a representative for blood vessel containing iodinebased contrast agents failed to show similar results to SDNRs for paediatric cases [9] but was suggested as a promising test material in the study by Gislasson et al. [1], only a simple comparison was performed in this work to test its suitability for use in adult cardiac catheterization studies. Comparisons were made between 8 mm sized vessel type inserts containing 20% iodine contrast objects and 99.98% tin sheets (127 μm thick) with corresponding width. It is shown (Figs. 6 and 7) that the SDNR values of tin show different behavior for different conditions when compared to the values from the vessel simulations. Smaller vessels and less iodine concentration are expected to differ even more, due to increased impact of background scatter. Therefore, the use of tin to represent iodine-based contrast agent in adult cardiac catheterization optimization studies requires more investigations.

4.2. Acquisition (ciné) For the Allura FD10 device, ciné images showed very similar SDNR values. This is because the modes differed only by the frame rate, and our method does not accommodate this aspect. Analyzing the ciné modes (Table 5), it is shown that the Allura FD10 Clarity ‘ciné boost’ mode excels over all other modes for all object sizes and iodine concentrations present in the phantom. The calculated SDNR values of the ‘ciné low’ and ‘ciné coronary’ can be considered identical. Table 1 indicated that the difference between these modes was due to the pulse rate. Since the present analysis is performed in single frames and with a static phantom, the advantages of faster rates or frame averaging in terms of temporal resolution are not included. The high SDNR values of the ‘ciné boost’ mode is due to higher dose, absence of filtration, and the estimated beam energy being relatively closer to the k-edge of iodine. From a FOM point of view on the Allura FD10 Clarity (Table 6), the three modes (ciné low, ciné normal, and ciné coronary) demonstrated indistinguishable values. The ‘ciné boost’ mode, on the other hand, showed typically lower FOM due to higher dose per frame. While the SDNR was highest, our tests suggest that the setting is intrinsically less efficient. It is an interesting finding that deserves further discussion with the manufacturer and could lead to a reprogramming of the ‘ciné boost’ mode. When contrast resolution is of high interest during a procedure (e.g. for roadmaps), it may be favorable to employ the ‘ciné normal’ mode. At present, for the visibility of the distinct materials imaged within the patient (e.g. cardiac stents), higher contrast images may be needed, and the ‘ciné boost’ mode could then be used. However, due to the low FOM and the fact that no copper filtration was present on the ‘boost’ mode, the use of this mode should be discouraged for routine procedures. Finally, ‘ciné low’ and ‘ciné coronary’

4.4. Limitations of study In the clinical setting, other factors not considered in this study may impact the AERC, either directly or indirectly. Such factors include the field of view (FoV) and the beam geometry that may change during a single procedure. This study, however, addressed only the ideal PA imaging projection with a single FoV. Other tools should be developed to allow a more complete checking within an acceptable time frame. A more complete test protocol would include the testing of other angulations and, as mentioned before, temporal resolution. While the system has two focal spot sizes, there was no indicator on the operator’s display of the focal spot in use. The automation of tube voltage and filtration for these modes did not allow any systematic study of either dose or image quality measure. The present study can therefore only discuss the available choices, but cannot find the optimal overall settings for complex clinical use. This will hopefully be solved in the future when the NEMA XR-27 standard becomes available. More generally, it has to be acknowledged that, while our study suggests an optimization process in terms of SDNR and EAKR, there is currently no hard proof of any direct relationship between the FOM and clinical imaging performance. 119

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

In ciné imaging modes on the Allura FD10 Clarity, the ‘ciné boost’ mode has lower FOM due to a high SDNR but significantly higher dose. Its use should be discouraged until adjustment is made to include copper filtration. The 30 fps ciné mode on the Allura FD10 should only be applied in cases with high demand for temporal resolution. The use of tin to represent iodine-based contrast agent in adult cardiac catheterization optimization studies requires more investigations before it can be recommended. While both the phantom and the evaluation method have limitations (such as not testing for the effects of moving vessels), the tests showed that many features of the system could be optimized or at least discussed with the manufacturer; the number of practical hints towards the users of the system was unexpectedly high. The additional value of the work is that the many results underline the need for medical physicists to develop a specific test object for the catheterization laboratory.

Another limitation is related to the fact that only one thickness of chest size was used to represent the Indonesian adult patient chest but this choice is supported by a specific study [22]. The use of only PMMA material in the phantom may lead to an overestimation in the actual ESAK since PMMA simulates harder tissues [29] with 11.3% higher density compared to soft tissue [21]. An introduction of aluminum sheets to represent the spine is considered for a next study and will make the phantom more anthropomorphic. The inclusion of some common clinical gantry angulations could also be considered. An additional limitation of this study is the static nature of the phantom, which is unable to demonstrate the effect of patient or vessel movement and the associated impact of pulse duration. It is also not clear whether the FOM includes image processing effects correctly, as it does not involve any frame averaging. Finally, we would recommend to consider including a test object to allow the separate evaluation of noise and contrast in future phantoms. 5. Conclusion

Acknowledgements

The present study shows that medical physicists can help in the optimization of clinically important interventional procedures. Many findings from this work have a direct impact on cardiac procedures. The ‘low dose’ fluoroscopy mode (setting I) on the Allura FD10 Clarity device is very efficient (high FOM). On the contrary, on the Allura FD10, the ‘low dose’ mode (setting I) is only advisable when lower SDNR is tolerable. On both devices, the use of the ‘high dose’ mode (setting III) should be avoided since there is a tendency to deliver more radiation dose with little to no contribution in comparison to lower dose settings. Fluoroscopy modes II and III, on both devices, require optimization.

This study was entirely funded by University of Indonesia Initial Research Grant 2015 (Hibah Riset Awal Universitas Indonesia 2015, contract no. 1578/UN2.R12/HKP.05.00/2015). The authors appreciated the help from Dr. Donald McLean for the advice on the manuscript, and the support from the head of Invasive Cardiology Department, Pertamina Central Hospital as well as from Rendra D. Sugandi and Dea Ryangga for technical assistance and valuable medical physics expertise.

Appendix See Tables 3–6. Table 3 Calculated SDNR and FOM of fluoroscopic exposures on the Allura FD10 Clarity device for various contrast concentrations and object sizes. Fluoroscopy setting

Setting I (low dose mode)

Object size (mm)

1 2 4 6 8

Setting II (medium dose mode)

1 2 4 6 8

Setting III (normal dose mode)

1 2 4 6 8

SDNR (FOM) for fluoroscopy on the Allura FD10 Clarity device 14%

16%

18%

20%

Tin

0.38 ± 0.16 (0.04 ± 0.03) 0.92 ± 0.09 (0.21 ± 0.04) 1.45 ± 0.11 (0.53 ± 0.08) 1.49 ± 0.07 (0.56 ± 0.05) 2.22 ± 0.11 (1.24 ± 0.12)

0.68 ± 0.10 (0.12 ± 0.03) 1.07 ± 0.16 (0.29 ± 0.08) 1.65 ± 0.07 (0.69 ± 0.06) 1.88 ± 0.18 (0.89 ± 0.16) 2.68 ± 0.22 (1.80 ± 0.30)

0.72 ± 0.14 (0.14 ± 0.03) 0.82 ± 0.12 (0.17 ± 0.03) 1.92 ± 0.07 (0.92 ± 0.07) 2.07 ± 0.14 (1.07 ± 0.15) 2.89 ± 0.23 (2.10 ± 0.34)

0.97 ± 0.12 (0.24 ± 0.06) 1.22 ± 0.32 (0.39 ± 0.16) 2.20 ± 0.09 (1.16 ± 0.11) 2.37 ± 0.11 (0.60 ± 0.09) 2.72 ± 0.27 (1.86 ± 0.38)

4.19 ± 0.61 (4.47 ± 1.34) 4.27 ± 0.18 (4.56 ± 0.38) 4.10 ± 0.16 (4.20 ± 0.32) 4.24 ± 0.08 (4.49 ± 0.17) 3.81 ± 0.62 (3.70 ± 1.18)

0.17 ± 0.12 (0.01 ± 0.01) 0.40 ± 0.07 (0.03 ± 0.01) 1.00 ± 0.10 (0.16 ± 0.03) 1.04 ± 0.08 (0.17 ± 0.03) 1.76 ± 0.13 (0.50 ± 0.08)

0.05 ± 0.04 (0.001 ± 0.001) 0.52 ± 0.08 (0.04 ± 0.01) 1.11 ± 0.15 (0.20 ± 0.05) 1.43 ± 0.10 (0.33 ± 0.05) 2.40 ± 0.29 (0.94 ± 0.23)

0.27 ± 0.01 (0.01 ± 0.001) 0.22 ± 0.01 (0.01 ± 0.001) 1.39 ± 0.12 (0.31 ± 0.05) 1.65 ± 0.08 (0.44 ± 0.04) 2.43 ± 0.16 (0.95 ± 0.13)

0.32 ± 0.15 (0.02 ± 0.02) 0.84 ± 0.12 (0.11 ± 0.03) 1.73 ± 0.10 (0.48 ± 0.06) 1.79 ± 0.16 (0.57 ± 0.09) 2.51 ± 0.25 (1.02 ± 0.20)

4.52 ± 0.54 (3.34 ± 0.78) 4.20 ± 0.23 (2.85 ± 0.32) 3.77 ± 0.46 (2.32 ± 0.52) 3.80 ± 0.45 (2.35 ± 0.50) 4.09 ± 0.15 (2.70 ± 0.20)

0.72 ± 0.24 (0.04 ± 0.03) 0.16 ± 0.15 (0.03 ± 0.01) 0.92 ± 0.09 (0.06 ± 0.01) 0.93 ± 0.10 (0.06 ± 0.01) 1.74 ± 0.77 (0.31 ± 0.01)

0.41 ± 0.06 (0.01 ± 0.01) 0.39 ± 0.13 (0.01 ± 0.01) 0.95 ± 0.16 (0.06 ± 0.02) 1.59 ± 0.15 (0.17 ± 0.03) 2.88 ± 0.06 (0.55 ± 0.02)

0.37 ± 0.22 (0.01 ± 0.01) 0.25 ± 0.09 (0.01 ± 0.01) 1.50 ± 0.11 (0.15 ± 0.02) 1.87 ± 0.05 (0.23 ± 0.01) 2.60 ± 0.28 (0.45 ± 0.10)

0.28 ± 0.12 (0.01 ± 0.01) 0.84 ± 0.09 (0.07 ± 0.01) 1.79 ± 0.06 (0.23 ± 0.01) 2.31 ± 0.15 (0.35 ± 0.02) 2.54 ± 0.19 (0.49 ± 0.13)

3.53 ± 0.98 (0.88 ± 0.47) 5.10 ± 0.39 (1.74 ± 0.26) 5.03 ± 0.26 (1.69 ± 0.17) 4.96 ± 0.18 (1.64 ± 0.12) 5.21 ± 0.10 (1.81 ± 0.07)

120

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

Table 4 Calculated SDNR and FOM of fluoroscopic exposures on the Allura FD10 device for various contrast concentrations and object sizes. Fluoroscopy setting

Object size (mm)

Setting I (low dose mode)

1 2 4 6 8

Setting II (normal dose mode)

1 2 4 6 8

Setting III (high dose mode)

1 2 4 6 8

SDNR (FOM) for fluoroscopy on the Allura FD10 device 14%

16%

18%

20%

Tin

0.22 ± 0.14 (0.01 ± 0.01) 0.90 ± 0.07 (0.12 ± 0.02) 1.69 ± 0.01 (0.42 ± 0.01) 1.56 ± 0.08 (0.36 ± 0.04) 2.53 ± 0.09 (0.94 ± 0.07)

0.56 ± 0.42 (0.06 ± 0.07) 1.34 ± 0.13 (0.27 ± 0.05) 1.75 ± 0.09 (0.45 ± 0.05) 1.86 ± 0.15 (0.51 ± 0.09) 2.90 ± 0.07 (1.24 ± 0.06)

0.45 ± 0.37 (0.04 ± 0.04) 0.93 ± 0.44 (0.15 ± 0.11) 2.17 ± 0.06 (0.69 ± 0.04) 1.98 ± 0.19 (0.58 ± 0.11) 2.27 ± 0.32 (0.77 ± 0.22)

1.18 ± 0.20 (0.21 ± 0.07) 1.44 ± 0.18 (0.31 ± 0.07) 2.81 ± 0.10 (1.26 ± 0.07) 2.82 ± 0.15 (0.95 ± 0.11) 2.82 ± 0.41 (1.28 ± 0.17)

5.83 ± 0.57 (5.03 ± 1.00) 4.92 ± 0.12 (3.57 ± 0.17) 4.65 ± 0.13 (3.19 ± 0.18) 4.64 ± 0.10 (3.17 ± 0.14) 5.00 ± 0.11 (3.69 ± 0.16)

0.55 ± 0.41 (0.03 ± 0.03) 1.69 ± 0.37 (0.22 ± 0.09) 2.35 ± 0.05 (0.41 ± 0.02) 2.22 ± 0.31 (0.37 ± 0.10) 3.57 ± 0.20 (0.95 ± 0.15)

0.92 ± 0.41 (0.07 ± 0.06) 1.84 ± 0.14 (0.25 ± 0.04) 2.61 ± 0.22 (0.51 ± 0.09) 2.81 ± 0.12 (0.59 ± 0.05) 4.26 ± 0.22 (1.35 ± 0.14)

0.89 ± 0.10 (0.06 ± 0.01) 1.23 ± 0.35 (0.12 ± 0.06) 2.96 ± 0.10 (0.65 ± 0.05) 2.96 ± 0.34 (0.66 ± 0.14) 3.08 ± 0.20 (0.70 ± 0.09)

1.54 ± 0.21 (0.18 ± 0.05) 2.17 ± 0.13 (0.35 ± 0.04) 3.71 ± 0.04 (0.98 ± 0.03) 3.76 ± 0.27 (1.11 ± 0.17) 3.99 ± 0.37 (1.02 ± 0.21)

6.61 ± 0.55 (3.25 ± 0.55) 6.89 ± 0.11 (3.52 ± 0.11) 6.69 ± 0.06 (3.32 ± 0.06) 6.17 ± 0.30 (2.82 ± 0.28) 7.09 ± 0.08 (3.72 ± 0.08)

1.16 ± 0.40 (0.08 ± 0.05) 1.89 ± 0.49 (0.21 ± 0.09) 2.53 ± 0.28 (0.36 ± 0.08) 2.58 ± 0.13 (0.37 ± 0.04) 2.96 ± 0.21 (0.71 ± 0.06)

1.43 ± 0.07 (0.12 ± 0.01) 2.07 ± 0.16 (0.24 ± 0.04) 2.83 ± 0.09 (0.45 ± 0.03) 2.93 ± 0.23 (0.48 ± 0.07) 4.12 ± 0.23 (0.95 ± 0.11)

1.43 ± 0.09 (0.12 ± 0.02) 1.63 ± 0.15 (0.15 ± 0.03) 3.09 ± 0.02 (0.54 ± 0.01) 3.01 ± 0.41 (0.51 ± 0.14) 3.17 ± 0.14 (0.57 ± 0.05)

1.76 ± 0.38 (0.18 ± 0.07) 2.32 ± 0.27 (0.30 ± 0.07) 3.49 ± 0.05 (0.71 ± 0.02) 3.73 ± 0.19 (0.73 ± 0.05) 3.85 ± 0.23 (0.69 ± 0.10)

6.39 ± 0.24 (2.30 ± 0.17) 6.26 ± 0.23 (2.20 ± 0.16) 6.19 ± 0.24 (2.15 ± 0.17) 5.94 ± 0.15 (1.98 ± 0.10) 5.99 ± 0.30 (2.02 ± 0.20)

Table 5 Calculated SDNR and FOM of ciné modes on the Philips Allura FD10 Clarity device for various contrast concentrations and object sizes. Ciné mode

Ciné low

Object size (mm)

1 2 4 6 8

Ciné normal

1 2 4 6 8

SDNR (FOM) for ciné mode on the Allura FD10 Clarity device 14%

16%

18%

20%

Tin

0.29 ± 0.15 (0.03 ± 0.02) 0.55 ± 0.05 (0.01 ± 0.01) 1.42 ± 0.09 (0.06 ± 0.01) 1.65 ± 0.06 (0.08 ± 0.01) 2.69 ± 0.07 (0.23 ± 0.01)

0.37 ± 0.13 (0.05 ± 0.03) 0.88 ± 0.25 (0.03 ± 0.01) 1.31 ± 0.04 (0.05 ± 0.01) 2.05 ± 0.07 (0.13 ± 0.01) 3.44 ± 0.09 (0.37 ± 0.02)

0.32 ± 0.38 (0.01 ± 0.01) 0.30 ± 0.15 (0.03 ± 0.01) 1.99 ± 0.09 (0.12 ± 0.01) 2.39 ± 0.05 (0.18 ± 0.01) 3.45 ± 0.26 (0.38 ± 0.05)

0.74 ± 0.04 (0.02 ± 0.01) 1.33 ± 0.24 (0.06 ± 0.02) 2.34 ± 0.02 (0.17 ± 0.03) 2.74 ± 0.18 (0.24 ± 0.03) 3.57 ± 0.22 (0.40 ± 0.05)

4.37 ± 0.27 (0.60 ± 0.07) 5.57 ± 0.18 (0.97 ± 0.06) 5.38 ± 0.69 (0.92 ± 0.22) 5.97 ± 0.17 (1.12 ± 0.06) 6.05 ± 0.35 (1.15 ± 0.13)

0.30 ± 0.23 (0.02 ± 0.01) 0.87 ± 0.14 (0.01 ± 0.01) 2.21 ± 0.14 (0.07 ± 0.01) 2.48 ± 0.26 (0.09 ± 0.02) 3.86 ± 0.21 (0.22 ± 0.03)

0.37 ± 0.18 (0.02 ± 0.01) 1.24 ± 0.28 (0.02 ± 0.01) 2.25 ± 0.32 (0.08 ± 0.02) 3.17 ± 0.18 (0.15 ± 0.02) 4.88 ± 0.17 (0.36 ± 0.02)

0.68 ± 0.25 (0.01 ± 0.01) 0.74 ± 0.14 (0.01 ± 0.01) 3.05 ± 0.18 (0.14 ± 0.02) 3.77 ± 0.21 (0.21 ± 0.02) 4.87 ± 0.51 (0.36 ± 0.08)

0.87 ± 0.28 (0.01 ± 0.01) 2.05 ± 0.14 (0.06 ± 0.01) 3.46 ± 0.21 (0.18 ± 0.02) 4.03 ± 0.24 (0.24 ± 0.03) 4.81 ± 0.80 (0.35 ± 0.11)

6.54 ± 1.86 (0.68 ± 0.32) 7.68 ± 0.41 (0.89 ± 0.10) 7.79 ± 0.61 (0.91 ± 0.14) 7.26 ± 0.25 (0.79 ± 0.05) 7.69 ± 0.01 (0.89 ± 0.01) (continued on next page)

121

Physica Medica 46 (2018) 114–123

L.E. Lubis et al.

Table 5 (continued) Ciné mode

Object size (mm)

Ciné boost

1 2 4 6 8

Ciné coronary

1 2 4 6 8

SDNR (FOM) for ciné mode on the Allura FD10 Clarity device 14%

16%

18%

20%

Tin

0.48 ± 0.36 (0.02 ± 0.01) 1.09 ± 0.06 (0.07 ± 0.01) 2.81 ± 0.23 (0.05 ± 0.01) 2.99 ± 0.43 (0.05 ± 0.01) 5.22 ± 0.14 (0.16 ± 0.01)

0.91 ± 0.11 (0.17 ± 0.04) 1.54 ± 0.14 (0.01 ± 0.01) 2.77 ± 0.56 (0.05 ± 0.02) 3.99 ± 0.19 (0.09 ± 0.01) 6.00 ± 0.29 (0.21 ± 0.02)

0.78 ± 0.37 (0.01 ± 0.01) 0.82 ± 0.072 (0.01 ± 0.01) 3.70 ± 0.28 (0.08 ± 0.01) 4.79 ± 0.03 (0.13 ± 0.01) 5.45 ± 0.62 (0.17 ± 0.04)

0.84 ± 0.06 (0.01 ± 0.01) 3.72 ± 0.18 (0.08 ± 0.01) 4.78 ± 0.22 (0.13 ± 0.01) 5.19 ± 0.71 (0.16 ± 0.04) 5.74 ± 0.24 (0.19 ± 0.02)

7.18 ± 1.51 (0.31 ± 0.13) 9.56 ± 0.76 (0.53 ± 0.08) 8.96 ± 0.15 (0.46 ± 0.02) 9.28 ± 0.50 (0.50 ± 0.05) 9.70 ± 0.01 (0.54 ± 0.01)

0.09 ± 0.02 (0.01 ± 0.01) 0.51 ± 0.15 (0.01 ± 0.01) 1.45 ± 0.05 (0.07 ± 0.01) 1.45 ± 0.13 (0.07 ± 0.01) 2.65 ± 0.12 (0.22 ± 0.02)

0.28 ± 0.14 (0.01 ± 0.01) 0.79 ± 0.18 (0.02 ± 0.01) 1.53 ± 0.21 (0.08 ± 0.02) 2.08 ± 0.12 (0.14 ± 0.02) 2.82 ± 1.26 (0.35 ± 0.02)

0.55 ± 0.31 (0.01 ± 0.01) 0.31 ± 0.19 (0.01 ± 0.01) 1.98 ± 0.18 (0.13 ± 0.02) 2.31 ± 0.13 (0.17 ± 0.02) 3.42 ± 0.45 (0.38 ± 0.10)

0.54 ± 0.22 (0.01 ± 0.01) 1.28 ± 0.14 (0.05 ± 0.01) 2.50 ± 0.11 (0.20 ± 0.02) 2.67 ± 0.11 (0.23 ± 0.01) 3.32 ± 0.61 (0.36 ± 0.14)

5.46 ± 0.64 (0.96 ± 0.22) 5.63 ± 0.24 (1.01 ± 0.09) 5.78 ± 0.22 (1.07 ± 0.08) 5.84 ± 0.37 (1.09 ± 0.13) 6.07 ± 0.25 (1.18 ± 0.10)

Table 6 Calculated SDNR and FOM for ciné modes on the Allura FD10 device for various contrast concentrations and object sizes. Ciné mode

Ciné 15 fps

Object size (mm)

1 2 4 6 8

Ciné 30 fps

1 2 4 6 8

Mean SDNR (Mean FOM) for ciné mode on the Allura FD10 device 14%

16%

18%

20%

Tin

0.86 ± 0.02 (0.05 ± 0.01) 2.20 ± 0.80 (0.03 ± 0.02) 3.84 ± 0.17 (0.10 ± 0.01) 3.46 ± 0.23 (0.08 ± 0.01) 5.62 ± 0.11 (0.21 ± 0.01)

1.47 ± 0.66 (0.02 ± 0.01) 2.85 ± 0.05 (0.05 ± 0.01) 4.07 ± 0.57 (0.11 ± 0.03) 4.54 ± 0.15 (0.13 ± 0.01) 7.37 ± 0.31 (0.35 ± 0.03)

0.63 ± 0.24 (0.03 ± 0.02) 1.70 ± 0.54 (0.02 ± 0.01) 5.05 ± 0.20 (0.17 ± 0.01) 4.46 ± 0.45 (0.13 ± 0.03) 5.22 ± 0.71 (0.18 ± 0.05)

1.73 ± 0.59 (0.02 ± 0.01) 3.41 ± 0.37 (0.08 ± 0.02) 6.02 ± 0.17 (0.24 ± 0.01) 6.04 ± 0.18 (0.24 ± 0.01) 5.86 ± 0.33 (0.22 ± 0.02)

10.45 ± 0.20 (0.71 ± 0.03) 11.63 ± 0.75 (0.88 ± 0.11) 12.26 ± 0.38 (0.98 ± 0.06) 11.87 ± 0.37 (0.92 ± 0.06) 11.77 ± 0.58 (0.90 ± 0.09)

0.35 ± 0.13 (0.01 ± 0.01) 1.89 ± 0.72 (0.03 ± 0.02) 3.58 ± 0.24 (0.09 ± 0.01) 3.39 ± 0.38 (0.08 ± 0.02) 5.73 ± 0.41 (0.22 ± 0.03)

1.06 ± 0.23 (0.01 ± 0.03) 2.60 ± 0.05 (0.05 ± 0.02) 3.84 ± 0.56 (0.10 ± 0.03) 4.42 ± 0.35 (0.13 ± 0.02) 7.14 ± 0.14 (0.34 ± 0.01)

1.08 ± 0.20 (0.01 ± 0.03) 1.42 ± 0.52 (0.01 ± 0.01) 4.89 ± 0.18 (0.16 ± 0.01) 4.27 ± 0.73 (0.12 ± 0.04) 4.82 ± 0.14 (0.16 ± 0.01)

1.15 ± 0.84 (0.01 ± 0.02) 3.10 ± 0.30 (0.06 ± 0.01) 6.24 ± 0.28 (0.26 ± 0.02) 6.20 ± 0.05 (0.26 ± 0.01) 5.90 ± 0.74 (0.23 ± 0.06)

10.01 ± 4.47 (0.76 ± 0.67) 12.17 ± 0.61 (0.99 ± 0.10) 12.00 ± 0.56 (0.96 ± 0.09) 11.83 ± 0.36 (0.93 ± 0.06) 12.58 ± 0.63 (1.06 ± 0.11)

References [8] [1] Gislason AJ, Davies AG, Cowen AR. Dose optimization in pediatric cardiac X-ray imaging. Med Phys 2010;37:5258–69. [2] Balter S. Methods for measuring fluoroscopic skin dose. Pediatr Radiol 2006;36(Suppl 2):136–40. http://dx.doi.org/10.1007/s00247-006-0193-3. [3] Geise RA, O’Dea TJ. Radiation dose in interventional fluoroscopic procedures. Appl Radiat Isot 1999;50:173–84. [4] National Council on Radiation Protection and Measurement. Report No. 168 – Radiation dose management for fluoroscopically-guided interventional medical procedures. Bethesda: National Council on Radiation Protection and Measurement; 2010. [5] Health effects of exposure to low levels of ionizing radiation (BEIR VII Phase 2). Washington DC: National Academy Press; 2006. [6] Nickoloff EL. Acceptance testing and performance monitoring of interventional radiology systems. In: Balter S, Chan RC, Shope TB, editors. Med. Phys. Monogr. No. 28 Intravasc. Brachytherapy, Fluoroscopically Guid. Interv. Madison: Medical Physics Publishing; 2002. [7] Chu RYL, Thomas G, Maqbool F. Skin entrance radiation dose in an interventional

[9]

[10] [11] [12]

[13]

122

radiology procedure. Health Phys 2006;91:41–6. http://dx.doi.org/10.1097/01.HP. 0000198784.94210.1f. Kato M, Chida K, Moritake T, Koguchi Y, Sato T, Oosaka H, et al. Fundamental study on the characteristics of a radiophotoluminescence glass dosemeter with no energy compensation filter for measuring patient entrance doses in cardiac interventional procedures. Radiat Prot Dosimetry 2014;162:224–9. http://dx.doi.org/10.1093/ rpd/nct300. Lubis LE, Bayuadi I, Pawiro SA, Ng K-H, Bosmans H, Soejoko DS. Optimization of dose and image quality of paediatric cardiac catheterization procedure. Phys Med 2015;31:659–68. http://dx.doi.org/10.1016/j.ejmp.2015.05.011. Gislason-Lee AJ, McMillan C, Cowen AR, Davies AG. Dose optimization in cardiac X-ray imaging. Med Phys 2013;40:91911. http://dx.doi.org/10.1118/1.4818016. Mahesh M. Fluoroscopy: patient radiation exposure issues. Radiographics 2001;21:1033–45. http://dx.doi.org/10.1148/radiographics.21.4.g01jl271033. European Commission. Radiation protection no. 174: European guideline on medical physics expert. Luxembourg: Publications Office of the European Union; 2014. http://dx.doi.org/10.2833/18393. International Atomic Energy Agency. IAEA Safety Standards General Safety Requirements Part 3 No. GSR Part 3 radiation protection and safety of radiation sources. International Basic Safety Standards; 2014.

Physica Medica 46 (2018) 114–123

L.E. Lubis et al. [14] Giles ER, Murphy PH. Measuring skin dose with radiochromic dosimetry film in the cardiac catheterization laboratory. Health Phys 2002;82:875–80. [15] Rivolta A, Emanuelli S, Tessarin C, Bresciani S, Genovese MG, Rogge D, et al. Method of patient dose evaluation in the angiographic and interventional radiology procedures. Radiol Med 2005;110:689–98. [16] Zamenhof RG. The optimization of signal detectability in digital fluoroscopy. Med Phys 1982;9:688–94. http://dx.doi.org/10.1118/1.595130. [17] Ten Cate T, van Wely M, Gehlmann H, Mauti M, Camaro C, Reifart N, et al. Novel Xray image noise reduction technology reduces patient radiation dose while maintaining image quality in coronary angiography. Neth Heart J 2015;23:525–30. http://dx.doi.org/10.1007/s12471-015-0742-1. [18] Nakamura S, Kobayashi T, Funatsu A, Okada T, Mauti M, Waizumi Y, et al. Patient radiation dose reduction using an X-ray imaging noise reduction technology for cardiac angiography and intervention. Heart Vessels 2016;31:655–63. http://dx. doi.org/10.1007/s00380-015-0667-z. [19] Eloot L, Thierens H, Taeymans Y, Drieghe B, De Pooter J, Van Peteghem S, et al. Novel X-ray imaging technology enables significant patient dose reduction in interventional cardiology while maintaining diagnostic image quality. Catheter Cardiovasc Interv 2015;86:E205–12. http://dx.doi.org/10.1002/ccd.25913. [20] American Association of Physicists in Medicine. AAPM report no. 70: cardiac catheterization equipment performance. Madison: Medical Physics Publishing; 2001. [21] Hubbel JH, Seltzer SM. Tables of X-ray mass attenuation coefficients and mass energy-absorption coefficients from 1 keV to 20 MeV for elements Z = 1 to 92 and 48 additional substances of dosimetric interest 1996, http://www.nist.gov/pml/

data/xraycoef/index.cfm [accessed June 20, 2014]. [22] Manuaba IB. Measurement of entrance surface dose (ESD) by some methods for chest examinations in computed radiography (CR). University of Indonesia; 2010. [23] Boone JM, Seibert JA. An accurate method for computer-generating tungsten anode X-ray spectra from 30 to 140 kV. Med Phys 1997;24:1661–70. http://dx.doi.org/10. 1118/1.597953. [24] International Atomic Energy Agency. IAEA technical report series no. 457: dosimetry in diagnostic radiology: an international code of practice. Vienna: International Atomic Energy Agency; 2007. [25] Vano E, Geiger B, Schreiner A, Back C, Beissel J. Dynamic flat panel detector versus image intensifier in cardiac imaging: dose and image quality. Phys Med Biol 2005;50:5731–42. http://dx.doi.org/10.1088/0031-9155/50/23/022. [26] Vano E, Ubeda C, Leyton F, Miranda P. Radiation dose and image quality for paediatric interventional cardiology. Phys Med Biol 2008;53:4049–62. http://dx. doi.org/10.1088/0031-9155/53/15/003. [27] Gagne RM, Boswell JS, Myers KJ. Signal detectability in digital radiography: spatial domain figures of merit. Med Phys 2003;30:2180–93. [28] Dehairs M, Bosmans H, Desmet W, Marshall NW. Evaluation of automatic dose rate control for flat panel imaging using a spatial frequency domain figure of merit. Phys Med Biol 2017;62:6610–30. http://dx.doi.org/10.1088/1361-6560/aa7a9d. [29] International Atomic Energy Agency. IAEA human health series no. 24: dosimetry in diagnostic radiology for paediatric patients. Vienna: International Atomic Energy Agency; 2013.

123