Comparative dosimetry of radiography device, MSCT device and two CBCT devices in the elbow region

Abstract The aim of the study was to estimate and to compare effective doses in the elbow region resulting from four different x‐ray imaging modalities. Absorbed organ doses were measured using 11 metal oxide field effect transistor (MOSFET) dosimeters that were placed in a custom‐made anthropomorphic elbow RANDO phantom. Examinations were performed using Shimadzu FH‐21 HR radiography device, Siemens Sensation Open 24‐slice MSCT‐device, NewTom 5G CBCT device, and Planmed Verity CBCT device, and the effective doses were calculated according to ICRP 103 recommendations. The effective dose for the conventional radiographic device was 1.5 µSv. The effective dose for the NewTom 5G CBCT ranged between 2.0 and 6.7 µSv, for the Planmed Verity CBCT device 2.6 µSv and for the Siemens Sensation MSCT device 37.4 µSv. Compared with conventional 2D radiography, this study demonstrated a 1.4–4.6 fold increase in effective dose for CBCT and 25‐fold dose for standard MSCT protocols. When compared with 3D CBCT protocols, the study showed a 6‐19 fold increase in effective dose using a standard MSCT protocol. CBCT devices offer a feasible low‐dose alternative for elbow 3D imaging when compared to MSCT.

the effective dose to the brain and to the bone marrow dose in the calvaria. 10,11 In recent years, cone beam computed tomography (CBCT) technology has found a new application for imaging extremities, 12 thus offering a versatile alternative to MSCT devices. In clinical settings, CBCT images are particularly useful in detecting subtle and non-displaced radial head fractures and for staging of intra-articular fractures with entrapment of fragments within the elbow joint. 13 The key benefits of CBCT technology are that they perform high-resolution imaging of the extremities using less time than radiographs and multi-detector computed tomography (MDCT) and lower radiation dose than MDCT. [14][15][16] Furthermore, CBCT images of extremities (arms/legs) can be undertaken without irradiating other body parts.
Regardless of the x-ray examination modality, radiation dose is always present, and the diagnostic benefits of all examinations should be weighed against the radiation risk that they induce. To the best of the authors' knowledge, there are currently no contemporaneous studies involving the x-ray radiation risk in the elbow region.
Therefore, the aims of this study were to assess and compare the organ and effective doses in the elbow resulting from conventional radiography, MSCT, and two CBCT devices using manufacturer-recommended elbow protocols. In order to perform effective and absorbed dose comparisons, the field of view (FOV) of the conventional radiographic device and the MSCT scanner were chosen to match the FOV (13 cm × 16 cm) of the Planmed Verity CBCT scanner. The Planmed Verity radiation doses were assessed using the novel "Ultra Low Dose" (ULD) protocol, the NewTom 5G "Standard Scan" setting radiation dose was measured using three different FOVs (12 cm × 8 cm, 15 cm × 12 cm, 18 cm × 16 cm), and for the "HiRes" setting, the dose was measured using a 12 cm × 8 cm FOV. Since the diagnosis of elbow fractures is commonly based on AP and LAT projections, the sum of the doses obtained using the 2D projections served as a benchmark for the dose comparisons.

2.A | X-ray devices
The exposure settings used in this study were based on a publication by Huang et al. 14 Table 1). The NewTom 5G device, however, uses a fixed 110 kVp tube voltage and automatically adjusted mA value based on two scout images of the region of interest (Table 1). The source-todetector distance remained unchanged during the scanning procedure. However, the cone beam angle of the NewTom 5G scanner varied between 10.3 (small FOV) and 15.4 degrees (large FOV) between the different FOVs. (Table 5.)

2.B | Phantom
All organ radiation dose measurements were performed using a specifically designed anthropomorphic RANDO adult arm phantom (Radiation Analogue Dosimetry System; The Phantom Laboratory, Salem, NY, USA). The phantom contains human bones that were encased within a soft tissue equivalent material to make the phantom match the radiation scattering and attenuation properties of the human elbow. The full arm phantom was sliced into 24 detachable layers that were numbered from 1 to 24 from the tip of the fingers to the upper arm (Fig. 1).
The layers were 25-mm thick and had a 1.5 cm × 1.5 cm grid of 5-mm diameter holes for the placement of MOSFET dosimeters.
Each dosimeter void was factory fitted with a soft tissue equivalent plug that allowed the correct positioning of 11 MOSFET dosimeters.

2.C.2 | MOSFET dosimeters
The absorbed radiation doses required for the effective dose calculations were measured using a mobile TN-RD-70-W20 MOSFET device. can be selected using a switch on the reader device for high or low bias voltage that subsequently provides high or low sensitivity, respectively. In this study, the high sensitivity setting was applied to attain the best accuracy and to minimize dose related inaccuracies. | 129 most radiosensitive organs (Table 2, Fig. 3). 19 The highest contributor to the effective dose in extremities is bone marrow. 16,20,21 Therefore, one dosimeter was placed within the bone (humerus, radius) in each layer to attain accurate dose detection and to provide a means to evaluate the dose uniformity amongst the different x-ray modalities. The remaining dosimeters were used to measure the skin, bone surface, muscle and lymphatic (cubital) nodes.

2.D | Equivalent dose
The equivalent or radiation weighted dose H T of the irradiated tissues T were calculated using the following equation 22 : In this equation, w R is the radiation weighting factor (w R = 1 Sv/Gy for x-rays), f i is the mass fraction of tissue T in phantom layer i, and D Ti is the average absorbed dose of tissue T in layer i. The summation was performed over all exposed layers (11 to 17). In this study, the mass fraction (f i ) denotes the coverage of each tissue in relative scale compared with the total tissue mass in the corresponding organ in the body (Table 3.) The studied organs were exposed by the incident and scattered radiation during the examinations. The dosimetric assessment was, however, performed using similar coverage regardless of the positioning of the beam to the phantom layers. 16 T A B L E 1 Exposure parameters of conventional radiography device, CBCT and MSCT scanners.  For NewTom 5G CBCT: mA is automatically adjusted by "SafeBeam™" acquisition technique.
F I G . 1. A schematic illustration of an anthropomorphic arm phantom with the exposed field of view.

2.E.1 | Bone marrow
Bone marrow (BM) is one of the largest and most radiosensitive tissues in the human body. According to a study by Hindorf et al., 23 the total bone marrow content represents 4% of the total body weight. However, the active radiosensitive red bone marrow (RBM) comprises only one-third of the total bone marrow weight. 24 In the present study, the bone marrow volume (cm 3 ) was assessed by measuring the cross section of all bones embedded in the phantom and multiplying them by the corresponding layer thickness (2.5 cm). The bone marrow volumes were summed for layers 10 to 17 to obtain the total bone marrow volume. The RBM weight was subsequently calculated as one-third of the BM volume multiplied by the density of red bone marrow. 25 Since bone marrow is one of the most radiosensitive organs, the RBM mass was also calculated based on data provided by Iwata et al. 26 and using the RBM distribution according to a recent publication by Machann et al. 27 According to both assessment methods the bone marrow in the elbow region represents 1.0% of the total bone marrow quantity in the human body.
Typical setup of an anthropomorphic elbow phantom placed into the bore of a CBCT device and the MOSFET readers.

2.E.2 | Bone surface
The bone surface in the elbow area was calculated using the bone surface-to-volume ratio according to ICRP Publication 70, 25 the percentage of total fresh bone weights defined in ICRP Publication 89 24 and the irradiated fractions of ulna, radius, and humerus based on the bone lengths according to a study by Schlenker et al. 28 The bone surface in the exposed region was estimated to represent 1.7% of the total body bone surface.

2.E.3 | Skin
The skin area was calculated in layers 11 to 17. Each layer perimeter was multiplied by the 2.5-cm phantom layer thickness and summed to attain the total skin area. The skin fraction in the elbow region was estimated by dividing the resulting value by the total skin area calculated using the Du Bois and Du Bois equation. 29 The fraction of skin in the exposed volume was equivalent to 2.9% of the whole-body skin.

2.E.4 | Remainder tissues
The effective dose contributions of the remainder tissues were calculated according to ICRP Publication 103 recommendations. 19 The elbow region contains only two remainder organs, namely muscles and lymphatic nodes, also known as cubital nodes.

2.E.5 | Muscles
The total muscle mass was assessed by summing the fraction of each muscle in the elbow region according to a study by Holzbaur et al. 30 The muscle mass fractions of each muscle were summed for layers between 11 and 17. The total body muscle mass in the elbow was calculated by dividing the result by the total 28 000 g body muscle mass. 31,32 It was estimated that muscles in the exposed region represent 0.9% of the total body muscle mass.

2.E.6 | Lymphatic nodes
The lymphatic node content in the elbow was evaluated using data published by Luscieti et al. 33 and the results of a whole-body lymphoscintigraphy examination performed at the Docrates Cancer Center in Helsinki, Finland. 16 According to the study by Luscieti et al., the adult cubital node cross section in the elbow region was comparable with the popliteal fossa cross section located in the knee area. Furthermore, based on the lymphoscintigraphy, it was estimated that the popliteal fossa in the knee area contains 5% of the total lymphatic nodes. Since the lymphatic node content (cubital nodes) in the elbow and knee were comparable, the same (5%) lymphatic node mass fraction was used.

2.F | Effective dose
The effective dose calculation was performed according to the International Commission of Radiological Protection (ICRP) guidelines. 19 Although the effective dose is not a physical quantity, it provides a useful means to assess the stochastic risk between different imaging techniques and protocols for unevenly distributed organ doses in the body. 21 The effective dose (E) was calculated from measured absorbed organ doses using the following equation: where w T is the weighting factor of tissue (T) and H T is the equivalent dose in tissue (T). All effective dose contributions were calculated using their specific fractions irradiated, weighting factors and adult organ dose compositions). 19,33 Furthermore, given that it may be difficult to compare the effective doses resulting from radiography, CBCT and MSCT devices due to the different exposure settings, their currentexposure timeproduct (mA * s) normalized effective doses (µSv/mAs) were calculated ( Table 4). The tissue mass fractions (f i ) and the ICRP Publication 103 weighting factors w T used in the calculations are presented in Table 3.

2.G | Uncertainty analysis
The repeated measurements according to a previous study by Koivisto et al. 16 The combined type B uncertainty was calculated from the quadratic summation of all estimated uncertainties.

3.C | CBCT devices
The effective dose acquired using NewTom 5G device (12 cm × 8 cm FOV) and "Standard Scan" setting was 2.0 µSv, and for the "HiRes"  Table 4. The absorbed organ doses (mGy) and their dose comparisons with the average dose value of the radiography device are presented in Table 5.
The anterior-posterior projections (cropped images) of the exposed volume using conventional radiographic, two CBCT, and MSCT devices are presented in Fig. 4.

3.D | Uncertainty, effective and organ doses
The type A standard uncertainty of the absorbed dose for all protocols varied between 15% and 48% in the 11 dosimeters. The uncertainties of point dose measurements were calculated as weighted sum of variances and included the statistical measurement error of ten repeated measurements according to a previous study, 34 dosimeter-and phantom position uncertainties, (10% and 10% respectively), x-ray source variation (5%) and cable irradiation uncertainties (1%). 35 The tissue dose uncertainty

| DISCUSSION
In this paper, the absorbed and effective doses in the elbow region were evaluated on one radiographic device, two CBCT devices and one MSCT device. Two-dimensional radiography typically consists of AP and LAT projections and is the most commonly used diagnostic imaging method in the elbow area. Therefore, the effective dose resulting from radiography projections was used as a benchmark for the dose comparison.

4.A | Effective dose
The concept of effective dose was first introduced in 1977 to assess stochastic health effects of radiation in medicine. 36 Since the introduction of effective dose, there has been an ongoing debate about its suitability for dose assessment due to the uncertainties in the conversion coefficients that are used to calculate the tissue weighting factors. 37 Regardless of the drawbacks, effective dose is, however, the only means used to assess and compare the risk of health detriment and was therefore chosen for this study. 38 However, according to Fisher et al. 39 the individual assessment of potential detriment should be based on organ or tissue absorbed radiation dose. Therefore, the measured absorbed organ doses of each protocol were also included in this study for dose comparison. The effective dose ratios (Table 4) and the absorbed organ dose ratios of the  dose attained using the 2D radiographic device. The effective dose recorded on the Planmed Verity CBCT device (2.6 µSv) was 1.8 times the effective dose measured on the 2D radiographic device.
When compared to the CBCT device, the effective dose attained on MSCT device was between 6 to 19 times the dose attained using the NewTom CBCT device and 14 times dose acquired on the Planmed Verity CBCT device.
The difference between MSCT effective dose and the average effective dose of the investigated CBCT protocols was 34.2 µSv.
Although this dose difference may seem negligible, being equivalent to only 1% (4 days) of the yearly background radiation dose (3.1 mSv), 40 it is still comparable to the effective dose resulting from two dental panoramic examinations. 41 In comparison, the effective dose difference between the radiography (1.5 µSv) and the CBCT dose on average (3.2 µSv) was 1.7 µSv, which is comparable to 5 h of background radiation.

4.B.2 | CBCT protocol comparison
When comparing the CBCT devices the effective doses acquired using the NewTom 5G "Standard Scan" setting were between 4% and 22% lower ( However, when compared to the (3.5 mAs) tube current -exposure product of the "Standard Scan" the "HiRes" mode uses 7.9 times higher (27.7 mAs) current-exposure product.
Since the dose is linearly dependent on the mAs-value, the effective dose of the "HiRes" mode should also be 7.9 higher than that of the "Standard Scan" mode. Surprisingly, the effective dose of the "HiRes" mode resulted in only a 3.4 times higher value than the "Standard Scan" mode. This finding is in good agreement with previous studies. 20 The low mAs value is an obvious exposure reduction benefit and subsequently results in a lower effective dose. However, the low mAs value could have a negative influence on the image quality since the contrast-to-noise-ratio (CNR) of the image is inversely proportional to the mAs 0.5 . 42 Furthermore, the CNR is known to be closely associated with the image quality. [43][44][45] In a previous study, Biswas et al. 46  In an earlier study, Cross et al. 48  One problem faced in the current study is that to date there is no general consensus on the quantity of red bone marrow in the arms. Cristy et al. 50 reported that red bone marrow is generally not found in the arms. On the contrary, a recent study by Karampinos et al. 51 reported that red bone marrow can be found in the ends of the long bones near the joints in healthy adults. Moreover, in a recent study, Machann et al. 34 reported on red and yellow bone marrow distributions in young children and adults. More specifically, there is an age-related change in the distribution of active marrow ranging from a high of as much as 5% in infancy and early childhood to as low as 0% in adults. Therefore, the 1% red bone marrow content used in this study elbow of the total RBM was chosen as this reasonably reflects maximum potential adult marrow content and is a conservative estimator for the risk calculation.
One difficulty when imaging elbow fractures using an MSCT device compared with a dedicated extremity CBCT device is the positioning of the elbow into the FOV without irradiating other body parts. If the arm would be placed adjacent to the patient's body, the internal organs would become exposed by the radiation. This would increase the effective dose markedly and cause radiation beam ter at detecting small bone and joint fractures when compared to radiographic devices. However, it must be noted that the increased detection rate of fractures resulted in higher radiation doses. [52][53][54] Future studies are needed to investigate the MSCT and CBCT iterative reconstruction possibilities to reduce the effective dose while maintaining good diagnostic image quality.

| CONCLUSION S
When compared with the conventional radiographic device, the standard MSCT protocol resulted in a 25-fold increase in effective dose.
The standard elbow protocols on the NewTom 5G and Planmed Verity CBCT devices resulted in a 0.7-to 2.4-fold increase in effective dose, respectively. The two CBCT devices offered 3D images of the elbow at a significantly lower dose than the MSCT device.