Ⅰ. Introduction
Bone Mineral Density (BMD) is the amount of bone mass in unit area and osteoporosis is bone metabolic disease which alters the bone composition resulting in BMD decrease[1,2]. There are many factors to osteoporosis and it is classified into primary osteoporosis and secondary osteoporosis. The primary osteoporosis is caused by either menopause or aging, however, it is ambiguous to specify because it progresses at almost the same time. Secondary osteoporosis is due to an exposure to a drug or disorder that interferes with the acquisition of the maximal bone mass or reduction of bone mass[3]. In addition, previous research suggests that various factors such as sex, smoking, drinking, and exercise, menopause, calcium, vitamin D and caffeine intake also affected osteoporosis[4-8].
According to the Health Insurance Review and Assessment Service, the number of patients with osteoporosis increased by 12.58% from 805,304 in 2013 to 906,631 in 2017. In 2017, the number of treatment of women in age of 50, the age women normally reach menopause, was significantly higher than that of 40, and the number of men in age of 40 was lower than that of women, but it increased significantly after the age of 50[9].
Osteoporosis and osteopenia are one of the main factors for the diagnosis of bone mineral density, and it is very important to prevent and treat fractures [10-11]. Bone mineral density (BMD) is measured with vrious means such as radiographic absorptiometry, dual energy X-ray absorptiometry (DXA), quantitative ultrasound (QUS), quantitative computed tomography (QCT). DXA displays the multiple level of energy intensities according to the distribution of bone density through two different kinds of energy X-rays passing through the human body. It is the most commonly used measurement device for diagnosis of osteoporosis for used to quantify the bone density and. DXA bone density measurement devices should be its capability to maintain the accuracy and reliability of t he t est[12] . However, s ince t he b one density measurement value requires certain level of periodic maintenance, there is a risk of misdiagnosis by inaccurate result of the test if it is not properly maintained[13]. Moreover, general-purpose phantoms are not easy to secure for with its expensive prices[14].
Therefore, the study developed a bone density phantom using 3D printer for efficient quality control for DXA device. In order to solve the cost problem which is the major issue of the conventional phantom diffusion, 3D printer has been employed to manufacture a cross calibration phantom with FDM 3D printer which is simple and costs less to manufacture, and evaluate the capability of Korean bone mineral density by linear regression analysis of DXA phantom measurements obtained from the Korea National Health and Nutrition Examination Survey (KNHANES).
Ⅱ. Materials and Methods
1. Evaluating Korean Bone Mass Density through The Korea National Health and Nutrition Examination Survey
In this study, the results of the National Health and Nutrition Survey were used to create Korean bone mineral density custom phantom. The National Health and Nutrition Examination Survey was conducted to ascertain the health and nutritional status of the people based on Article 16 of the National Health Promotion Act, the raw data from the first period of 1998 to the seventh periodical data of the year 2017 is open to the public[15]. The study was conducted with the raw data from the fourth period (2008 ~ 2009) and the fifth period (2010 ~ 2011) in which BMD was performed and 21,072 subjects who measured vertebral bone density were selected. The data of 16,906 persons were analyzed as shown in Table 1, except for those who have the double missing value and are younger than 20 years and those older than 80 years.
2. Equipment and materials
The phantom was printed with personal FDM 3D printer (3DP-210F, Cubicon, Republic of Korea) of Cubicon. The maximum printer size of the 3D printer is 150 × 150 × 150 mm (W × H × D), and the X and Y positioning accuracy is 12.5 μm, and the Z positioning precision is 2.5 μm.
PLA filaments in a diameter of 1.75 mm and 30% Cu-PLA filaments were used to print out the soft tissue model and the bone tissue model, and the phantom modeling and slicing has been conducted with 123D design (v2.2.14, Autodesk, USA) and Cubicreator (v2.5.3.2, Cubicon, Republic of Korea).
3. Phantom production with 3D printer
The study has been conducted with the size of models; the width and depth of the soft tissue model used for phantom production are shown in Fig. 1, 90 mm and 130 mm, and the height was 215 mm and 155 mm. The bone tissue model was placed in the center of the phantom, and it was produced in size of 40 × 30 × 25 mm (W × D × H) to fit into the soft tissue model. Table 2 shows the printing conditions of the soft tissue and bone tissue model. The 30% Cu-PLA used in the bone model was printed by infilling from at beginning 70% and increased by 5% up to 100% in order to evaluate the Hounsfield Unit (HU).
4. Evaluation the usability of the phantom
The 30% Cu-PLA was photographed with a 16-ch CT (Alexion, TOSHIBA) to obtain DICOM and the HU was evaluated using Image J 1.52a (National Institutes of Health, USA). Hologic’s Discovery device has was employed and measured 10 times respectively to compute the cross sectional analysis and evaluate the BMD of the National Health and Nutrition Examination Survey and the printed phantom model. Linear regression analysis was performed using SPSS 23.0 (SPSS Inc. Chicago, IL, USA) to determine the relevance of the measured values.
Ⅲ. Results
1. HU Evaluation by bone model infilling
In Fig. 2, HU of 30% Cu-PLA was measured three times to simulate the bone tissue of BMD phantom. As shown in Table 3, 1,011.60 ± 3.80 at 70%, 1,139.55 ± 9.08 at 75%, 1,288.39 ± 6.25 at 80%, 1,376.20 ± 16.33 at 85%, 1,497.51 ± 2.04 at 90%, and 1,624.52 ± 5.52 at 95%, and 1,772.21±11.74 at 100% respectively. The study was conducted with phantom produced with 80%, 85% and 90% infilling rate. In addition, PLA was measured three times with 100% internal filling to simulate soft tissue, and HU was 36.73 ± 0.89.
2. Phantom print
Fig. 3(a) is a model with 80%, 85%, and 90% infilling of 30% Cu-PLA, and Fig. 3(b) is a bone mass model stacked with 95 mm a soft tissue phantom top and bottom. Fig. 3(c) is a model with 65 mm a soft tissue phantom.
The total output of time and usage of 215 mm height phantom was about 61 hours 9 minutes for ABS and 3,039.83 g. The total output of time and usage of 155 mm height phantom was 44 hours 11 minutes and 2,160.37 g. 30% Cu-PLA was used for about 2 hours and 26 minutes, 97.27 g.
3. Phantom BMD measurement and linear regression analysis
The L1, L2, and L3 BMD values of phantoms produced with the 3D printer were measured to be 3.015±0.049 g/cm2, 3.353±0.047 g/cm2, and 3.631± 0.042 g/cm2 at 215 mm height and 2.973±0.094 g/cm2, 3.247±0.046 g/cm2, 3.405±0.064 g/cm2 at 155 mm height.
BMD values from KNHANES were 0.886±0.146 g/cm2, 0.930±0.157 g/cm2 and 0.958±0.161 g/cm2. Therefore, a linear regression analysis was conduct to examine the relevance with the phantom model. As a result in Fig. 4, R2 was 0.995 at 215 mm height and computed the equation to predict the existing result as Eq. (1), and R2 was 0.9994 at the height of 155 mm and computed the equation as Eq. (2).
As Table 4 shows, with Eq (1). the BMD values of L1, L2 and L3 were 0.887 ± 0.006 g/cm2, 0.927±0.006 g/cm2 and 0.960±0.005 g/cm2 respectively, and there are value difference from those of KNHANES which were 0.15% and -0.32% , And 0.16%. In addition, with Eq (2) applied, The BMD values of L1, L2 and L3 were 0.882±0.011 g/cm2, 0.914±0.005 g/cm2 and 0.933± 0.008 g/cm2, respectively, and the values difference were -0.42% and -1.67% % And -2.60% from those of KNHANES.
Ⅳ. Discussions
In order to reduce dose exposure of patient and obtain accurate diagnosis result, quality control and maintenance of diagnostic radiation generator should be performed periodically. In particular, bone density test equipment could be influenced by many factors, and it is far more important because inaccurate measurement would eventually affect the result value.
The International Society for Clinical Densitometry (ISCD) suggests to conduct cross-analysis of DXA equipment when conducting comparative analysis with BMD values of different facilities or when the same manufacturer or others equipment is introduced[16]. However, the cross-correction phantom of DXA equipment is poorly penetrated due to lack of awareness of the industry and the high cost[17]. Therefore, the study has developed a low-cost cross-calibration phantom using 3D printing technology.
The density of the PLA material is 1.24 g / cm3, which is similar to that of the human tissue. The previous study have used it to produce X-ray dosimetry phantoms for radiotherapy and phantom for diagnostic X-ray devices [18-20]. Moreover, it has the similar density to PMMA which is used for soft tissue and tissue equivalent phantom, and is used for soft tissue mimic material. 30% Cu-PLA whose HU is relatively larger was used as a bone morphogenetic material to effectively display the difference in the radiation transmittance. However, in order to use the phantom as a substitute for the Korean BMD value, a crossover analysis must be performed and the result is shown in Fig 4. In the linear regression analysis, R2 was 0.995 at 215 mm height, and R2 was 0.9994 at 155 mm height, and 215 mm and 155 mm height which displays the statistically significant with the mean value of Korean bone density values. It proves that the produced phantom can be utilized on DXA cross correction by applying Eq (1) and Eq (2) and 3D phantom with a thin 155 mm height soft tissue model had a stronger correlation.
This suggests that the periodical error measurement enable to calibrate the reading and continuous maintenance of the patient’s bone density value by calibrating the measurd value of same models and same BMD value measured from different facilities before and after the measurement. However, the study was performed with the discovery device of Hologic, and the incompatibility between heterogeneous BMD measurement devices should be improved in future studies.
Ⅴ. Conclusions
This study has produced a DXA cross correction phantom based on Korean BMD with 3D printer. A linear regression analysis of the bone density of the phantom produced by the 3D printer and the Korean bone density displayed a high correlation respectively, and it was slightly higher when using 155 mm height of soft tissue than of 255 mm height.
It would be possible to conduct proper quality control and maintenance and cross - analyze using DXA device, and it is expected to be used as fundamental data for various medical phantom manufacturing using Korean standard numerical value and 3D printer.
