sensors Article
Development of Portable Digital Radiography System with a Device for Monitoring X-ray Source-Detector Angle and Its Application in Chest Imaging Tae-Hoon Kim 1 , Dong-Woon Heo 1 , Chang-Won Jeong 1 , Jong-Hyun Ryu 1 , Hong Young Jun 1 , Seung-Jun Han 1 , Taeuk Ha 1 and Kwon-Ha Yoon 1,2, * 1
2
*
Imaging Science Research Center, Wonkwang University, 460 Iksandeaero, Iksan, Jeonbuk 54538, Korea; [email protected] (T.-H.K.); [email protected] (D.-W.H.); [email protected] (C.-W.J.); [email protected] (J.-H.R.); [email protected] (H.Y.J.); [email protected] (S.-J.H.); [email protected] (T.H.) Department of Radiology, Wonkwang University School of Medicine, 460 Iksandeaero, Iksan, Jeonbuk 54538, Korea Correspondence: [email protected]; Tel.: +82-63-859-1921
Academic Editors: Ioannis Kompatsiaris, Thanos G. Stavropoulos and Antonis Bikakis Received: 15 December 2016; Accepted: 2 March 2017; Published: 7 March 2017
Abstract: This study developed a device measuring the X-ray source-detector angle (SDA) and evaluated the imaging performance for diagnosing chest images. The SDA device consisted of Arduino, an accelerometer and gyro sensor, and a Bluetooth module. The SDA values were compared with the values of a digital angle meter. The performance of the portable digital radiography (PDR) was evaluated using the signal-to-noise (SNR), contrast-to-noise ratio (CNR), spatial resolution, distortion and entrance surface dose (ESD). According to different angle degrees, five anatomical landmarks were assessed using a five-point scale. The mean SNR and CNR were 182.47 and 141.43. The spatial resolution and ESD were 3.17 lp/mm (157 µm) and 0.266 mGy. The angle values of the SDA device were not significantly difference as compared to those of the digital angle meter. In chest imaging, the SNR and CNR values were not significantly different according to the different angle degrees. The visibility scores of the border of the heart, the fifth rib and the scapula showed significant differences according to different angles (p < 0.05), whereas the scores of the clavicle and first rib were not significant. It is noticeable that the increase in the SDA degree was consistent with the increases of the distortion and visibility score. The proposed PDR with a SDA device would be useful for application in the clinical radiography setting according to the standard radiography guidelines. Keywords: portable digital radiography (PDR) system; X-ray source-detector angle (SDA); radiography setting
1. Introduction Digital radiography (DR) systems are frequently used in medical and industrial applications such as dental X-ray imaging, the aerospace industry, and in security and nondestructive testing [1]. Advancements in DR technologies have allowed more opportunities in the medical field because of a number of useful capabilities such as real-time acquisition, real-time post-processing, wireless techniques, and electronic data archiving and retrieval [2,3]. Such advantages of the DR system enhanced portability, mobility and smart environments due to the decreased size, light weight and better image quality depending on the improvement of the detector performance, as well as the implementation of multi-functional software. Among these advantages, mobility is an increasingly important topic in radiographic image acquisition [4]. Portable DR (PDR) detectors are suitable for Sensors 2017, 17, 531; doi:10.3390/s17030531
www.mdpi.com/journal/sensors
Sensors 2017, 17, 531
2 of 12
important topic in radiographic image acquisition [4]. Portable DR (PDR) detectors are suitable for Sensors 2017, 17, 531 2 of 12 intensive care units, emergency departments, outpatient clinics and so on [5]. Recent DR systems have increased the use of flat-panel detectors (FPDs) because they offer a lower radiation dose while preserving imaging as contrasted with thin-film transistor (TFT) [6–8]. Therefore, it intensive care units,quality emergency departments, outpatient clinics and sodetectors on [5]. Recent DR systems is important newer DR systems are designed to solve unmet needs of radiologists, have increasedthat the use of flat-panel detectors (FPDs) because theythe offer a lower radiation dose while technologists, administrators and private conjunction the [6–8]. advantages in DR preserving imaging quality as contrasted withpractices thin-filmin transistor (TFT) with detectors Therefore, it is technologies. important that newer DR systems are designed to solve the unmet needs of radiologists, technologists, In clinical radiography, the X-ray distance (SDD)inand source-detector administrators and private practices in source–to-detector conjunction with the advantages DR X-ray technologies. angleIn(SDA) are important factors for acquiring appropriate image quality. According to standard clinical radiography, the X-ray source–to-detector distance (SDD) and X-ray source-detector clinical radiography guidelines [9,10], is recommended for such imaging positions as follows: SDD angle (SDA) are important factors foritacquiring appropriate image quality. According to standard of 180 cm, and SDA of 0° (around 5°, vertical incidence) for chest AP (anterior to posterior) and clinical radiography guidelines [9,10], it is recommended for such imaging positions as follows: SDDPA of (posterior to SDA anterior) cm SDDincidence) and 15°–20° side direction incidence) 180 cm, and of 0◦imaging; (around100 5◦ , vertical forSDA chest(head AP (anterior to posterior) and for PA cervical spine AP imaging; and 100 SDD and 5°–7° side direction incidence) for lateral (posterior to anterior) imaging; 100cm cm SDD and 15◦SDA –20◦ (head SDA (head side direction incidence) for ◦ ◦ knee imaging. Currently, most commercial PDR systems are well equipped as a tape measure for cervical spine AP imaging; and 100 cm SDD and 5 –7 SDA (head side direction incidence) for lateral SDD, whereas Currently, they are relatively less equipped a specific device for SDA to the knee imaging. most commercial PDR systems are well equipped as acompared tape measure for equipment for they SDDare [11]. In clinical conditions or compared extreme medical situations SDD, whereas relatively lesssituations, equippedunexpected a specific device for SDA to the equipment such as imaging for immobilized bedridden patients, post-stroke for SDD [11]. Inexperiments clinical situations, unexpected conditions or extreme medicalhemiplegic situations patients such as and severely injured for patients are often encountered [12]. From these viewpoints, a radiographic imaging experiments immobilized bedridden patients, post-stroke hemiplegic patients and severely imaging systemare automatically determining thethese SDDviewpoints, and SDAa radiographic can be helpful for clinical injured patients often encountered [12]. From imaging system radiographic However, are few studies focusing on radiographic SDA measurement clinical automatically settings. determining the SDDthere and SDA can be helpful for clinical settings.inHowever, radiography. there are few studies focusing on SDA measurement in clinical radiography. Therefore, thepurpose purposeofof study to develop PDR system including for Therefore, the thisthis study waswas to develop a PDRasystem including a device afordevice obtaining obtaining the to SDA andthe to assess theperformance imaging performance for the diagnosis of chest imaging. the SDA and assess imaging for the diagnosis of chest imaging. 2. Materials Materials and and Methods Methods 2. 2.1. Portable Portable Digital Digital Radiography Radiography System System 2.1. The PDR PDR main main body body consisted consisted of of an an X-ray X-ray source, source, aa FPD, FPD, and and an an integral integral PC PC (Figure (Figure 1A) 1A) and and its its The detailed structure of PDR is shown in Figure 1B. The X-ray source generated 40–110 kVp and 20–100 mA detailed structure of PDR is shown in Figure 1B. The X-ray source generated 40–110 kVp and 20–100 withwith a focal spot size 1.8 of mm of 20.4 of kg20.4 (PXP-100CA, Poskom,Poskom, Goyang,Goyang, Korea). We used mA a focal spotofsize 1.8and mmweight and weight kg (PXP-100CA, Korea). a high resolution amorphousamorphous silicon (a-Si)/CsI with PIN diode sensor-based We used a high resolution siliconthin-film (a-Si)/CsItransistor thin-film(TFT) transistor (TFT) with PIN diode FPD with 3072 × 3072 pixels (pixel pitch 140 × 140 µm), image size of 43 cm × 43 cm, and sensor-based FPD with 3072 × 3072 pixels (pixel pitch 140 × 140 μm), image size of 43 cm × weight 43 cm, of 11weight kg (FXRD-1717SA, Vieworks, Seoul, Korea). Image acquisition time of thetime FPDofwas 1 s. and of 11 kg (FXRD-1717SA, Vieworks, Seoul, Korea). Image acquisition the below FPD was The integrated PC for internal processing weighed 7.95 kg and used OS, an Intel Core below 1 s. The integrated PC for internal processing weighed 7.95 kg Windows and used 7Windows 7 OS, an i5 4300U with a 1.9 GHz CPU (POC-W242, Advantech, Taiwan). In order to increase flexibility and Intel Core i5 4300U with a 1.9 GHz CPU (POC-W242, Advantech, Taiwan). In order to increase mobility ofand PDR system,of X-ray implemented twoimplemented gas springs totwo smoothly move under a zero flexibility mobility PDRsource system, X-ray source gas springs to smoothly gravity state and PDR main body installed two main wheels and an auxiliary wheel. The main body of move under a zero gravity state and PDR main body installed two main wheels and an auxiliary the PDR system attached on a slot for safekeeping an independent detector. wheel. The main body of the PDR system attached on a slot for safekeeping an independent detector.
Figure 1. A simple diagram of portable digital radiography (PDR) (A), the detailed drawing of PDR Figure 1. A simple diagram of portable digital radiography (PDR) (A), the detailed drawing of PDR structure (B) and developed PDR system (C). structure (B) and developed PDR system (C).
Sensors 2017, 17, 531
3 of 12
2.2. Internal Processing of Portable Digital Radiography Sensors 2017, 17, 531PC is detachable type and received image data from the FPD. In order to 3 of 12 The integrated process received image data, the PDR system used a two-step correction algorithm as follows. In the first step, 2.2. Internal Processing of Portable Digital Radiography bad pixels on image data were detected with a 3 × 3 differentiation operator. When the differentiation The integrated PC is detachable type and received image data from theasFPD. In pixel. order to at a given pixel was greater than a threshold value, the pixel was classified a bad Inprocess the second received image data,ofthe usedreplaced a two-step as follows. In the first step, the intensity value thePDR badsystem pixel was by correction a bi-linearalgorithm interpolating function of adjacent step, bad pixels on image data were detected with a 3 × 3 differentiation operator. When the pixel values [13]. Then, the final image data were generated to Digital Imaging and Communications differentiation at a given pixel was greater than a threshold value, the pixel was classified as a bad in Medicine (DICOM) file format. The DICOM files were stored in a specific directory on the PC hard pixel. In the second step, the intensity value of the bad pixel was replaced by a bi-linear interpolating disk storage. function of adjacent pixel values [13]. Then, the final image data were generated to Digital Imaging
and Communications in Medicine (DICOM) file format. The DICOM files were stored in a specific directory on the PC hard disk storage.
2.3. Development and Performance Test of X-ray SDA Device
To determine the flip angle degree between an X-ray source and a detector, we developed 2.3. Development andthat Performance Testofof aX-ray SDA Device an X-ray SDA device consisted main board (Arduino Uno R3 board; ARDUINO, Ivrea, determine the flip angle degree between an X-ray source a detector, developed an Italy), a 6To degrees of freedom (DOF) inertial measurement unitand (IMU) shieldwe (embedded in the X-rayaccelerometer SDA device thatand consisted of a main board Uno R3 board; ARDUINO, Ivrea, Italy),and a ADXL345 the ITG-3200 gyro; (Arduino Sparkfun Electronics, Boulder, CO, USA), a 6 degrees of freedom (DOF) inertial measurement unit (IMU) shield (embedded in the ADXL345 Bluetooth module (HC-05 master, HC-06 slave; Shenzhen Rainbowsemi Electronics Co., Guangdong, and the Electronics, CO, USA), and a2),Bluetooth China)accelerometer [14]. This device wasITG-3200 attachedgyro; to theSparkfun X-ray source (deviceBoulder, 1) and detector (device respectively. module (HC-05 master, HC-06 slave; Shenzhen Rainbowsemi Electronics Co., Guangdong, China) The Arduino main board was the core processing component coded by the Sketch software tool. The 6 [14]. This device was attached to the X-ray source (device 1) and detector (device 2), respectively. DOF IMU shield included an accelerometer and gyro sensor, and transmitted the angle data to the The Arduino main board was the core processing component coded by the Sketch software tool. The Arduino through I2C communication after sensing the angle data the X-ray andtodetector. 6 DOF IMU shield included an accelerometer and gyro sensor, andfrom transmitted the source angle data the The 6 Arduino DOF is achieved by using a microelectromechanical to sense movement through I2C communication after sensing thesystem angle data fromtranslational the X-ray source and in three detector. perpendicular axes is(surge, heave, sway) and rotational movement about threetranslational perpendicular The 6 DOF achieved by using a microelectromechanical system to sense movement threeThe perpendicular (surge,toheave, sway) and rotational about three axes (roll, pitch, in yaw). 6 DOF IMUaxes designed provide motion, position,movement and navigational sensing perpendicular axes (roll, pitch, yaw). The 6 DOF IMU designed to provide motion, position, and from a durable single device over 6 DOF (six-dimensional motion variants) [14]. The Bluetooth module navigational sensing from core a durable single device over 6 DOF (six-dimensional motion variants) was connected to the Arduino processor. [14]. The Bluetooth module was connected to the Arduino core processor. Figure 2 shows the system block for processing procedures by the SDA device. The SDA value Figure 2 shows the system block for processing procedures by the SDA device. The SDA value was measured as following steps. The Arduinos of devices 1 and 2 received angle data through I2C was measured as following steps. The Arduinos of devices 1 and 2 received angle data through I2C communication from the 6 DOF IMU Shield. Then, the Arduinos calculated the angle values for device communication from the 6 DOF IMU Shield. Then, the Arduinos calculated the angle values for 1 (XS and YS 1) and device 2 (XD and Y applying an algorithm to the angle [15]. Bluetooth device (XS and YS) and device 2 D(X) Dby and YD) by applying an algorithm to thedata angle dataThe [15]. The module (HC-06) on device 2 received the angle values (X and Y ) via serial communication from the D D Bluetooth module (HC-06) on device 2 received the angle values (XD and YD) via serial Arduino, and transmitted theArduino, angle values (XD and Ythe thevalues Bluetooth module of device communication from the and transmitted angle (XD and YD) to(HC-05) the Bluetooth D ) to module (HC-05) of devicemodule 1 in real-time. TheofBluetooth (HC-05) device 1 transmitted 1 in real-time. The Bluetooth (HC-05) device 1 module transmitted theofangle values (XD andthe YD ) to angle values (XD and YD) to the Arduino via serial communication. Finally, the Arduino of device 1 the Arduino via serial communication. Finally, the Arduino of device 1 calculated the SDA difference difference YSDA )Yas follows: XSDA =. XS − XD, YSDA = YS − YD. (XSDAcalculated and YSDAthe ) asSDA follows: XSDA(X=SDA XSand −X , = Y − Y D SDA S D
Figure 2. System block of the device X-raysource-detector source-detector angle (SDA). Figure 2. System block of the devicefor forobtaining obtaining the the X-ray angle (SDA).
Sensors 2017, 17, 531 Sensors 2017, 17, 531
4 of 12 4 of 12
In order to test the performance of SDA device, we used a digital angle meter (DL-155V, STS, ◦ /80 ◦ –90 orderwith to test the performance of SDA device, used a ◦digital meter (DL-155V, Tokyo, In Japan) accuracies of 0.1%/degree (0◦ –10we ), andangle 0.2%/degree (10◦ –80◦STS, ) as a Tokyo, Japan) with accuracies of 0.1%/degree (0°–10°/80°–90°), and 0.2%/degree (10°–80°) as a standard reference device. Figure 3 showed the developed SDA device attaching on the PDR system standard reference device. Figure 3 showed the developed SDA device attaching on the PDR system and the initial setting for achieving the angle degrees using both the SDA device and digital angle and the initial setting for achieving the angle degrees using both the SDA device and digital angle meter. First, the ‘0.0◦ ’ value on the SDA device was accordant with a ‘0.0◦ ’ reference value on the meter. First, the ‘0.0°’ value on the SDA device was accordant with a ‘0.0°’ reference value on the digital angle meter (Figure 3C). Next, the angle value of the digital angle meter changed from 0 to digital angle meter (Figure 3C). Next, the angle value of the digital angle meter changed from 0 to 60 60 degrees in units of five degrees, and then, the angle value on the SDA device measured three times. degrees in units of five degrees, and then, the angle value on the SDA device measured three times. The measured angle degree using PDR with SDA device displayed on a LCD monitor (Figure 3D). The measured angle degree using PDR with SDA device displayed on a LCD monitor (Figure 3D). Finally, thethe angle difference and SDA SDAdevice deviceinineach eachangle angle degree was Finally, angle differencebetween betweendigital digital angle angle meter meter and degree was calculated as mean difference of both angle values. calculated as mean difference of both angle values.
Figure 3. Details of the device for obtaining the SDA, which was attached on an X-ray source (A) and Figure 3. Details of the device for obtaining the SDA, which was attached on an X-ray source (A) and a detector (B). Photography showed the initial setting for obtaining the angle degree using the SDA a detector (B). Photography showed the initial setting for obtaining the angle degree using the SDA device and digital angle meter (C). The ‘0.0°’ value on the SDA device was accordant with a ‘0.0°’ device and digital angle meter (C). The ‘0.0◦ ’ value on the SDA device was accordant with a ‘0.0◦ ’ reference value on the digital angle meter. The measured angle degree using PDR with SDA device reference value on the digital angle meter. The measured angle degree using PDR with SDA device displayed on a LCD monitor (D). displayed on a LCD monitor (D).
2.4. Measurements of Image Quality and Radiation Dose 2.4. Measurements of Image Quality and Radiation Dose The image quality of the PDR system was evaluated as signal-to-noise ratio (SNR), The image quality of the was evaluated as signal-to-noise ratio contrast-to-noise ratio (CNR), and PDR spatialsystem resolution using a bar phantom (X-ray test pattern type(SNR), 18; FUNK, Berlin, Germany). The SNR and CNR were calculated the ratio (X-ray of the lead bar (0.05type mm18; contrast-to-noise ratio (CNR), and spatial resolution using a barasphantom test pattern thick)Berlin, value to noise and The the ratio theCNR lead bar-air contrast toas noise, respectively [16].bar Mean SNR FUNK, Germany). SNRofand were calculated the ratio of the lead (0.05 mm and CNR values were obtained from six image sets using the developed PDR system. The thick) value to noise and the ratio of the lead bar-air contrast to noise, respectively [16]. Mean SNR
Sensors 2017, 17, 531 Sensors 2017, 17, 531
5 of 12 5 of 12
modulation transfer function (MTF) has been used to evaluate spatial resolution of imaging systems and CNR values were obtained from six image sets using the developed PDR system. The modulation [17]. This study was used a bar phantom to generate the MTF curve and was measured image transfer function (MTF) has been used to evaluate spatial resolution of imaging systems [17]. This study resolution at 10% on MTF curve. was used a bar phantom to generate the MTF curve and was measured image resolution at 10% on The radiation dose was calculated using the method described by the International Commission MTF curve. on Radiological Protection [18]. Entrance surface dose (ESD) is the absorbed dose including the The radiation dose was calculated using the method described by the International Commission contribution from backscatter [1]. The ESD measurement was performed using a dosimeter (Piranha, on Radiological Protection [18]. Entrance surface dose (ESD) is the absorbed dose including RTI Electronics, Mölndal, Sweden). This study was measured the ESD under the conditions of 80 the contribution from backscatter [1]. The ESD measurement was performed using a dosimeter kVp tube voltage, 4 mAs current, 100 ms and 1 m SDD. (Piranha, RTI Electronics, Mölndal, Sweden). This study was measured the ESD under the conditions of 80 kVp tube voltage, 4 mAs current, 100 ms and 1 m SDD. 2.5. Chest Radiography According to Different Angle Degrees 2.5. Chest According to Different ChestRadiography radiographs were performed sixAngle timesDegrees under the conditions of different angles using the developed system with device.six Totimes assessunder the image quality according to the angulation of Chest PDR radiographs were SDA performed the conditions of different angles using the the detector in the mediolateral plane of the patient, the angle value of the SDA changed from 0° developed PDR system with SDA device. To assess the image quality according to the angulation (medial line) to in 30°the in units of 10 degrees 4). For the of chest images, and CNR of the detector mediolateral plane (Figure of the patient, theanalysis angle value of the SDA SNR changed from according to different angle degrees were measured from 15 different points using the Alderson ◦ ◦ 0 (medial line) to 30 in units of 10 degrees (Figure 4). For the analysis of chest images, SNR and Radiation Therapy (ART) angle male degrees phantom (ART-200A). The 15 image distortion according to the CNR according to different were measured from different points using the Alderson changes angulation the deformity percent changes in the to sizes cardiac RadiationofTherapy (ART)was malemeasured phantom (ART-200A). Theasimage distortion according the of changes of transverse diameter (CTD), thoracic transverse diameter (TTD), and thoracic longitudinal diameter angulation was measured the deformity as percent changes in the sizes of cardiac transverse diameter (TLD). Also, two expert radiologists (moreand thanthoracic 10 years of experience) blindly evaluated eachexpert chest (CTD), thoracic transverse diameter (TTD), longitudinal diameter (TLD). Also, two image, and (more reached a 10 consensus regarding the anatomic landmarks [19].image, The five anatomica radiologists than years of experience) blindly evaluated each chest and reached landmarks were the border of the heart, clavicle, first rib, fifth rib and scapula. Each anatomic consensus regarding the anatomic landmarks [19]. The five anatomic landmarks were the border of the landmark on chest image data was analyzed according to the radiological diagnosis on a five-point heart, clavicle, first rib, fifth rib and scapula. Each anatomic landmark on chest image data was analyzed scale: 1, definitely seen; 2, probably seen; probably not seen; and2,5,probably definitelyseen; not according to the radiological diagnosis on3,a equivocal; five-point 4, scale: 1, definitely seen; seen. 3, equivocal; 4, probably not seen; and 5, definitely not seen.
Figure 4. 4. The The angulation angulation plane plane (A) (A) and and reference reference line line setting setting (B) (B) for for the of distortion Figure the measurement measurement of distortion on chest phantom: (a) cardiac transverse diameter (CTD); (b) thoracic transverse diameter (TTD); on chest phantom: (a) cardiac transverse diameter (CTD); (b) thoracic transverse diameter (TTD); and (c) (c) thoracic thoracic longitudinal longitudinal diameter diameter (TLD). and (TLD). L L and and R R indicated indicated left left side side and and right right side. side.
Statistical Analysis 2.6. Statistical analyses were performed performed using the Statistical Package for the Social Sciences All statistical analyses Coefficient of of variance variance (CV) (CV) of SNR and CNR was (SPSS version 20.0, Chicago, IL, USA) software. Coefficient calculated for the variability of measurements in the PDR system [20]. [20]. The angle difference between and digital angle meter) was analyzed with the independent two sample both angle anglemeters meters(SDA (SDAdevice device and digital angle meter) was analyzed with the independent two t-test. According to the different degrees, variation image qualities, distortion anddistortion visibility sample t-test. According to the angle different anglethe degrees, theinvariation in image qualities, scores were analyzed with the repeated-measures analysis of variance (rmANOVA) and(rmANOVA) Tukey’s post and visibility scores were analyzed with the repeated-measures analysis of variance hoc tests. Two-sided less than 0.05 wereless considered indicate statistical and Tukey’s post hocp-values tests. Two-sided p-values than 0.05towere considered to significance. indicate statistical significance.
Sensors 2017, 17, 531 Sensors 2017, 17, 531
6 of 12 6 of 12
3. 3. Results Results 3.1. 3.1. Development Development and and Performances Performances of of PDR PDR The the PDR PDR system system were were 723 723 mm mm (width) (width) × × 650 mm (depth) (depth) × × 1376 The size size and and weight weight of of the 650 mm 1376 mm mm (height) and approximately 60 kg, respectively (Figure 1). The PDR has a smooth flexibility (height) and approximately 60 kg, respectively (Figure 1). The PDR has a smooth flexibility of of the the X-ray anan adult. TheThe mean SNR andand CNR werewere 182.47 ± 6.75 X-ray source sourceand andcan canbe bemoved movedeasily easilybyby adult. mean SNR CNR 182.47 ±(CV: 6.75 3.7%) and 141.43 ± 6.08 (4.3%), The overall values the system less were than 5%. (CV: 3.7%) and 141.43 ± 6.08 respectively. (4.3%), respectively. TheCV overall CVinvalues in thewere system less The resolution at 10% MTF (157lp/mm μm) (Figure 5) and the ESD wasthe 0.266 mGy. thanspatial 5%. The spatial resolution at was 10%3.17 MTFlp/mm was 3.17 (157 µm) (Figure 5) and ESD was Therefore, PDR system is easily applicable for imaging inpatients such as 0.266 mGy.our Therefore, our PDR system is easily applicable forbedside imagingpatients bedsideor patients or inpatients immobilized bedridden patients and post-stroke hemiplegic patients. such as immobilized bedridden patients and post-stroke hemiplegic patients.
Figure 5. Bar phantom image obtained from the PDR to calculate the signal-to-noise ratio (SNR), Figure 5. Bar phantom image obtained from the PDR to calculate the signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR) and spatial resolution (A). Line-squares on the phantom images contrast-to-noise ratio (CNR) and spatial resolution (A). Line-squares on the phantom images indicated indicated theofregions forCNR SNRmeasurements. and CNR measurements. A modulation transfer function the regions interestof forinterest SNR and A modulation transfer function (MTF, 10%) (MTF, 10%) curve was used for resolution evaluation (B). The 10% MTF value is 3.17 lp/mm (157μm) curve was used for resolution evaluation (B). The 10% MTF value is 3.17 lp/mm (157µm) on the image on the image obtained fromobtained the PDR.from the PDR.
3.2. Performance of the SDA Device 3.2. Performance of the SDA Device Figure 3 shows the features of the SDA device on the PDR system. The device was attached to Figure 3 shows the features of the SDA device on the PDR system. The device was attached the X-ray source and detector, respectively, and its size was 55 mm (width) × 80 mm (depth) × 35 mm to the X-ray source and detector, respectively, and its size was 55 mm (width) × 80 mm (depth) × (height). The angle values obtained from the SDA device were displayed in real time on the LCD 35 mm (height). The angle values obtained from the SDA device were displayed in real time on the monitor (1 times/s). The averaged angle values of the SDA device and digital angle meter are LCD monitor (1 times/s). The averaged angle values of the SDA device and digital angle meter are summarized in Table 1, and the mean difference in both angle values was 0.71° ± 0.15°. There was no summarized in Table 1, and the mean difference in both angle values was 0.71◦ ± 0.15◦ . There was no significant angle difference between the digital angle meter and our SDA device (p > 0.05). significant angle difference between the digital angle meter and our SDA device (p > 0.05). 3.3. 3.3. Chest Chest Radiographic Radiographic Study Study According According to to Different Different Angles Angles for for Clinical Clinical Application Application Figure Figure 66 shows shows the the representative representative chest chest AP AP images images obtained obtained from from the the developed developed PDR PDR according according to different angles. Imaging quality, distortion and visibility scores are summarized in Table 2. SNR2. to different angles. Imaging quality, distortion and visibility scores are summarized in Table and CNR values were not significantly different from different the angle degrees (rmANOVA, p> SNR and CNR values were not significantly different from different the angle degrees (rmANOVA, 0.05). TheThe objective image qualities were notnot significantly different ininthis p > 0.05). objective image qualities were significantly different thisstudy; study;thus, thus,ititcould could be be considered to be applicable for appropriate angle degrees in clinical experiments. However, the sizes considered to be applicable for appropriate angle degrees in clinical experiments. However, the sizes and and deformity deformity (%) (%) in in the the CTD, CTD, TTD TTD and and TLD TLD increased increased according according to to the the increment increment of of angulation angulation (rmANOVA, (rmANOVA, pp
Development of Portable Digital Radiography System with a Device for Monitoring X-ray Source-Detector Angle and Its Application in Chest Imaging Tae-Hoon Kim 1 , Dong-Woon Heo 1 , Chang-Won Jeong 1 , Jong-Hyun Ryu 1 , Hong Young Jun 1 , Seung-Jun Han 1 , Taeuk Ha 1 and Kwon-Ha Yoon 1,2, * 1
2
*
Imaging Science Research Center, Wonkwang University, 460 Iksandeaero, Iksan, Jeonbuk 54538, Korea; [email protected] (T.-H.K.); [email protected] (D.-W.H.); [email protected] (C.-W.J.); [email protected] (J.-H.R.); [email protected] (H.Y.J.); [email protected] (S.-J.H.); [email protected] (T.H.) Department of Radiology, Wonkwang University School of Medicine, 460 Iksandeaero, Iksan, Jeonbuk 54538, Korea Correspondence: [email protected]; Tel.: +82-63-859-1921
Academic Editors: Ioannis Kompatsiaris, Thanos G. Stavropoulos and Antonis Bikakis Received: 15 December 2016; Accepted: 2 March 2017; Published: 7 March 2017
Abstract: This study developed a device measuring the X-ray source-detector angle (SDA) and evaluated the imaging performance for diagnosing chest images. The SDA device consisted of Arduino, an accelerometer and gyro sensor, and a Bluetooth module. The SDA values were compared with the values of a digital angle meter. The performance of the portable digital radiography (PDR) was evaluated using the signal-to-noise (SNR), contrast-to-noise ratio (CNR), spatial resolution, distortion and entrance surface dose (ESD). According to different angle degrees, five anatomical landmarks were assessed using a five-point scale. The mean SNR and CNR were 182.47 and 141.43. The spatial resolution and ESD were 3.17 lp/mm (157 µm) and 0.266 mGy. The angle values of the SDA device were not significantly difference as compared to those of the digital angle meter. In chest imaging, the SNR and CNR values were not significantly different according to the different angle degrees. The visibility scores of the border of the heart, the fifth rib and the scapula showed significant differences according to different angles (p < 0.05), whereas the scores of the clavicle and first rib were not significant. It is noticeable that the increase in the SDA degree was consistent with the increases of the distortion and visibility score. The proposed PDR with a SDA device would be useful for application in the clinical radiography setting according to the standard radiography guidelines. Keywords: portable digital radiography (PDR) system; X-ray source-detector angle (SDA); radiography setting
1. Introduction Digital radiography (DR) systems are frequently used in medical and industrial applications such as dental X-ray imaging, the aerospace industry, and in security and nondestructive testing [1]. Advancements in DR technologies have allowed more opportunities in the medical field because of a number of useful capabilities such as real-time acquisition, real-time post-processing, wireless techniques, and electronic data archiving and retrieval [2,3]. Such advantages of the DR system enhanced portability, mobility and smart environments due to the decreased size, light weight and better image quality depending on the improvement of the detector performance, as well as the implementation of multi-functional software. Among these advantages, mobility is an increasingly important topic in radiographic image acquisition [4]. Portable DR (PDR) detectors are suitable for Sensors 2017, 17, 531; doi:10.3390/s17030531
www.mdpi.com/journal/sensors
Sensors 2017, 17, 531
2 of 12
important topic in radiographic image acquisition [4]. Portable DR (PDR) detectors are suitable for Sensors 2017, 17, 531 2 of 12 intensive care units, emergency departments, outpatient clinics and so on [5]. Recent DR systems have increased the use of flat-panel detectors (FPDs) because they offer a lower radiation dose while preserving imaging as contrasted with thin-film transistor (TFT) [6–8]. Therefore, it intensive care units,quality emergency departments, outpatient clinics and sodetectors on [5]. Recent DR systems is important newer DR systems are designed to solve unmet needs of radiologists, have increasedthat the use of flat-panel detectors (FPDs) because theythe offer a lower radiation dose while technologists, administrators and private conjunction the [6–8]. advantages in DR preserving imaging quality as contrasted withpractices thin-filmin transistor (TFT) with detectors Therefore, it is technologies. important that newer DR systems are designed to solve the unmet needs of radiologists, technologists, In clinical radiography, the X-ray distance (SDD)inand source-detector administrators and private practices in source–to-detector conjunction with the advantages DR X-ray technologies. angleIn(SDA) are important factors for acquiring appropriate image quality. According to standard clinical radiography, the X-ray source–to-detector distance (SDD) and X-ray source-detector clinical radiography guidelines [9,10], is recommended for such imaging positions as follows: SDD angle (SDA) are important factors foritacquiring appropriate image quality. According to standard of 180 cm, and SDA of 0° (around 5°, vertical incidence) for chest AP (anterior to posterior) and clinical radiography guidelines [9,10], it is recommended for such imaging positions as follows: SDDPA of (posterior to SDA anterior) cm SDDincidence) and 15°–20° side direction incidence) 180 cm, and of 0◦imaging; (around100 5◦ , vertical forSDA chest(head AP (anterior to posterior) and for PA cervical spine AP imaging; and 100 SDD and 5°–7° side direction incidence) for lateral (posterior to anterior) imaging; 100cm cm SDD and 15◦SDA –20◦ (head SDA (head side direction incidence) for ◦ ◦ knee imaging. Currently, most commercial PDR systems are well equipped as a tape measure for cervical spine AP imaging; and 100 cm SDD and 5 –7 SDA (head side direction incidence) for lateral SDD, whereas Currently, they are relatively less equipped a specific device for SDA to the knee imaging. most commercial PDR systems are well equipped as acompared tape measure for equipment for they SDDare [11]. In clinical conditions or compared extreme medical situations SDD, whereas relatively lesssituations, equippedunexpected a specific device for SDA to the equipment such as imaging for immobilized bedridden patients, post-stroke for SDD [11]. Inexperiments clinical situations, unexpected conditions or extreme medicalhemiplegic situations patients such as and severely injured for patients are often encountered [12]. From these viewpoints, a radiographic imaging experiments immobilized bedridden patients, post-stroke hemiplegic patients and severely imaging systemare automatically determining thethese SDDviewpoints, and SDAa radiographic can be helpful for clinical injured patients often encountered [12]. From imaging system radiographic However, are few studies focusing on radiographic SDA measurement clinical automatically settings. determining the SDDthere and SDA can be helpful for clinical settings.inHowever, radiography. there are few studies focusing on SDA measurement in clinical radiography. Therefore, thepurpose purposeofof study to develop PDR system including for Therefore, the thisthis study waswas to develop a PDRasystem including a device afordevice obtaining obtaining the to SDA andthe to assess theperformance imaging performance for the diagnosis of chest imaging. the SDA and assess imaging for the diagnosis of chest imaging. 2. Materials Materials and and Methods Methods 2. 2.1. Portable Portable Digital Digital Radiography Radiography System System 2.1. The PDR PDR main main body body consisted consisted of of an an X-ray X-ray source, source, aa FPD, FPD, and and an an integral integral PC PC (Figure (Figure 1A) 1A) and and its its The detailed structure of PDR is shown in Figure 1B. The X-ray source generated 40–110 kVp and 20–100 mA detailed structure of PDR is shown in Figure 1B. The X-ray source generated 40–110 kVp and 20–100 withwith a focal spot size 1.8 of mm of 20.4 of kg20.4 (PXP-100CA, Poskom,Poskom, Goyang,Goyang, Korea). We used mA a focal spotofsize 1.8and mmweight and weight kg (PXP-100CA, Korea). a high resolution amorphousamorphous silicon (a-Si)/CsI with PIN diode sensor-based We used a high resolution siliconthin-film (a-Si)/CsItransistor thin-film(TFT) transistor (TFT) with PIN diode FPD with 3072 × 3072 pixels (pixel pitch 140 × 140 µm), image size of 43 cm × 43 cm, and sensor-based FPD with 3072 × 3072 pixels (pixel pitch 140 × 140 μm), image size of 43 cm × weight 43 cm, of 11weight kg (FXRD-1717SA, Vieworks, Seoul, Korea). Image acquisition time of thetime FPDofwas 1 s. and of 11 kg (FXRD-1717SA, Vieworks, Seoul, Korea). Image acquisition the below FPD was The integrated PC for internal processing weighed 7.95 kg and used OS, an Intel Core below 1 s. The integrated PC for internal processing weighed 7.95 kg Windows and used 7Windows 7 OS, an i5 4300U with a 1.9 GHz CPU (POC-W242, Advantech, Taiwan). In order to increase flexibility and Intel Core i5 4300U with a 1.9 GHz CPU (POC-W242, Advantech, Taiwan). In order to increase mobility ofand PDR system,of X-ray implemented twoimplemented gas springs totwo smoothly move under a zero flexibility mobility PDRsource system, X-ray source gas springs to smoothly gravity state and PDR main body installed two main wheels and an auxiliary wheel. The main body of move under a zero gravity state and PDR main body installed two main wheels and an auxiliary the PDR system attached on a slot for safekeeping an independent detector. wheel. The main body of the PDR system attached on a slot for safekeeping an independent detector.
Figure 1. A simple diagram of portable digital radiography (PDR) (A), the detailed drawing of PDR Figure 1. A simple diagram of portable digital radiography (PDR) (A), the detailed drawing of PDR structure (B) and developed PDR system (C). structure (B) and developed PDR system (C).
Sensors 2017, 17, 531
3 of 12
2.2. Internal Processing of Portable Digital Radiography Sensors 2017, 17, 531PC is detachable type and received image data from the FPD. In order to 3 of 12 The integrated process received image data, the PDR system used a two-step correction algorithm as follows. In the first step, 2.2. Internal Processing of Portable Digital Radiography bad pixels on image data were detected with a 3 × 3 differentiation operator. When the differentiation The integrated PC is detachable type and received image data from theasFPD. In pixel. order to at a given pixel was greater than a threshold value, the pixel was classified a bad Inprocess the second received image data,ofthe usedreplaced a two-step as follows. In the first step, the intensity value thePDR badsystem pixel was by correction a bi-linearalgorithm interpolating function of adjacent step, bad pixels on image data were detected with a 3 × 3 differentiation operator. When the pixel values [13]. Then, the final image data were generated to Digital Imaging and Communications differentiation at a given pixel was greater than a threshold value, the pixel was classified as a bad in Medicine (DICOM) file format. The DICOM files were stored in a specific directory on the PC hard pixel. In the second step, the intensity value of the bad pixel was replaced by a bi-linear interpolating disk storage. function of adjacent pixel values [13]. Then, the final image data were generated to Digital Imaging
and Communications in Medicine (DICOM) file format. The DICOM files were stored in a specific directory on the PC hard disk storage.
2.3. Development and Performance Test of X-ray SDA Device
To determine the flip angle degree between an X-ray source and a detector, we developed 2.3. Development andthat Performance Testofof aX-ray SDA Device an X-ray SDA device consisted main board (Arduino Uno R3 board; ARDUINO, Ivrea, determine the flip angle degree between an X-ray source a detector, developed an Italy), a 6To degrees of freedom (DOF) inertial measurement unitand (IMU) shieldwe (embedded in the X-rayaccelerometer SDA device thatand consisted of a main board Uno R3 board; ARDUINO, Ivrea, Italy),and a ADXL345 the ITG-3200 gyro; (Arduino Sparkfun Electronics, Boulder, CO, USA), a 6 degrees of freedom (DOF) inertial measurement unit (IMU) shield (embedded in the ADXL345 Bluetooth module (HC-05 master, HC-06 slave; Shenzhen Rainbowsemi Electronics Co., Guangdong, and the Electronics, CO, USA), and a2),Bluetooth China)accelerometer [14]. This device wasITG-3200 attachedgyro; to theSparkfun X-ray source (deviceBoulder, 1) and detector (device respectively. module (HC-05 master, HC-06 slave; Shenzhen Rainbowsemi Electronics Co., Guangdong, China) The Arduino main board was the core processing component coded by the Sketch software tool. The 6 [14]. This device was attached to the X-ray source (device 1) and detector (device 2), respectively. DOF IMU shield included an accelerometer and gyro sensor, and transmitted the angle data to the The Arduino main board was the core processing component coded by the Sketch software tool. The Arduino through I2C communication after sensing the angle data the X-ray andtodetector. 6 DOF IMU shield included an accelerometer and gyro sensor, andfrom transmitted the source angle data the The 6 Arduino DOF is achieved by using a microelectromechanical to sense movement through I2C communication after sensing thesystem angle data fromtranslational the X-ray source and in three detector. perpendicular axes is(surge, heave, sway) and rotational movement about threetranslational perpendicular The 6 DOF achieved by using a microelectromechanical system to sense movement threeThe perpendicular (surge,toheave, sway) and rotational about three axes (roll, pitch, in yaw). 6 DOF IMUaxes designed provide motion, position,movement and navigational sensing perpendicular axes (roll, pitch, yaw). The 6 DOF IMU designed to provide motion, position, and from a durable single device over 6 DOF (six-dimensional motion variants) [14]. The Bluetooth module navigational sensing from core a durable single device over 6 DOF (six-dimensional motion variants) was connected to the Arduino processor. [14]. The Bluetooth module was connected to the Arduino core processor. Figure 2 shows the system block for processing procedures by the SDA device. The SDA value Figure 2 shows the system block for processing procedures by the SDA device. The SDA value was measured as following steps. The Arduinos of devices 1 and 2 received angle data through I2C was measured as following steps. The Arduinos of devices 1 and 2 received angle data through I2C communication from the 6 DOF IMU Shield. Then, the Arduinos calculated the angle values for device communication from the 6 DOF IMU Shield. Then, the Arduinos calculated the angle values for 1 (XS and YS 1) and device 2 (XD and Y applying an algorithm to the angle [15]. Bluetooth device (XS and YS) and device 2 D(X) Dby and YD) by applying an algorithm to thedata angle dataThe [15]. The module (HC-06) on device 2 received the angle values (X and Y ) via serial communication from the D D Bluetooth module (HC-06) on device 2 received the angle values (XD and YD) via serial Arduino, and transmitted theArduino, angle values (XD and Ythe thevalues Bluetooth module of device communication from the and transmitted angle (XD and YD) to(HC-05) the Bluetooth D ) to module (HC-05) of devicemodule 1 in real-time. TheofBluetooth (HC-05) device 1 transmitted 1 in real-time. The Bluetooth (HC-05) device 1 module transmitted theofangle values (XD andthe YD ) to angle values (XD and YD) to the Arduino via serial communication. Finally, the Arduino of device 1 the Arduino via serial communication. Finally, the Arduino of device 1 calculated the SDA difference difference YSDA )Yas follows: XSDA =. XS − XD, YSDA = YS − YD. (XSDAcalculated and YSDAthe ) asSDA follows: XSDA(X=SDA XSand −X , = Y − Y D SDA S D
Figure 2. System block of the device X-raysource-detector source-detector angle (SDA). Figure 2. System block of the devicefor forobtaining obtaining the the X-ray angle (SDA).
Sensors 2017, 17, 531 Sensors 2017, 17, 531
4 of 12 4 of 12
In order to test the performance of SDA device, we used a digital angle meter (DL-155V, STS, ◦ /80 ◦ –90 orderwith to test the performance of SDA device, used a ◦digital meter (DL-155V, Tokyo, In Japan) accuracies of 0.1%/degree (0◦ –10we ), andangle 0.2%/degree (10◦ –80◦STS, ) as a Tokyo, Japan) with accuracies of 0.1%/degree (0°–10°/80°–90°), and 0.2%/degree (10°–80°) as a standard reference device. Figure 3 showed the developed SDA device attaching on the PDR system standard reference device. Figure 3 showed the developed SDA device attaching on the PDR system and the initial setting for achieving the angle degrees using both the SDA device and digital angle and the initial setting for achieving the angle degrees using both the SDA device and digital angle meter. First, the ‘0.0◦ ’ value on the SDA device was accordant with a ‘0.0◦ ’ reference value on the meter. First, the ‘0.0°’ value on the SDA device was accordant with a ‘0.0°’ reference value on the digital angle meter (Figure 3C). Next, the angle value of the digital angle meter changed from 0 to digital angle meter (Figure 3C). Next, the angle value of the digital angle meter changed from 0 to 60 60 degrees in units of five degrees, and then, the angle value on the SDA device measured three times. degrees in units of five degrees, and then, the angle value on the SDA device measured three times. The measured angle degree using PDR with SDA device displayed on a LCD monitor (Figure 3D). The measured angle degree using PDR with SDA device displayed on a LCD monitor (Figure 3D). Finally, thethe angle difference and SDA SDAdevice deviceinineach eachangle angle degree was Finally, angle differencebetween betweendigital digital angle angle meter meter and degree was calculated as mean difference of both angle values. calculated as mean difference of both angle values.
Figure 3. Details of the device for obtaining the SDA, which was attached on an X-ray source (A) and Figure 3. Details of the device for obtaining the SDA, which was attached on an X-ray source (A) and a detector (B). Photography showed the initial setting for obtaining the angle degree using the SDA a detector (B). Photography showed the initial setting for obtaining the angle degree using the SDA device and digital angle meter (C). The ‘0.0°’ value on the SDA device was accordant with a ‘0.0°’ device and digital angle meter (C). The ‘0.0◦ ’ value on the SDA device was accordant with a ‘0.0◦ ’ reference value on the digital angle meter. The measured angle degree using PDR with SDA device reference value on the digital angle meter. The measured angle degree using PDR with SDA device displayed on a LCD monitor (D). displayed on a LCD monitor (D).
2.4. Measurements of Image Quality and Radiation Dose 2.4. Measurements of Image Quality and Radiation Dose The image quality of the PDR system was evaluated as signal-to-noise ratio (SNR), The image quality of the was evaluated as signal-to-noise ratio contrast-to-noise ratio (CNR), and PDR spatialsystem resolution using a bar phantom (X-ray test pattern type(SNR), 18; FUNK, Berlin, Germany). The SNR and CNR were calculated the ratio (X-ray of the lead bar (0.05type mm18; contrast-to-noise ratio (CNR), and spatial resolution using a barasphantom test pattern thick)Berlin, value to noise and The the ratio theCNR lead bar-air contrast toas noise, respectively [16].bar Mean SNR FUNK, Germany). SNRofand were calculated the ratio of the lead (0.05 mm and CNR values were obtained from six image sets using the developed PDR system. The thick) value to noise and the ratio of the lead bar-air contrast to noise, respectively [16]. Mean SNR
Sensors 2017, 17, 531 Sensors 2017, 17, 531
5 of 12 5 of 12
modulation transfer function (MTF) has been used to evaluate spatial resolution of imaging systems and CNR values were obtained from six image sets using the developed PDR system. The modulation [17]. This study was used a bar phantom to generate the MTF curve and was measured image transfer function (MTF) has been used to evaluate spatial resolution of imaging systems [17]. This study resolution at 10% on MTF curve. was used a bar phantom to generate the MTF curve and was measured image resolution at 10% on The radiation dose was calculated using the method described by the International Commission MTF curve. on Radiological Protection [18]. Entrance surface dose (ESD) is the absorbed dose including the The radiation dose was calculated using the method described by the International Commission contribution from backscatter [1]. The ESD measurement was performed using a dosimeter (Piranha, on Radiological Protection [18]. Entrance surface dose (ESD) is the absorbed dose including RTI Electronics, Mölndal, Sweden). This study was measured the ESD under the conditions of 80 the contribution from backscatter [1]. The ESD measurement was performed using a dosimeter kVp tube voltage, 4 mAs current, 100 ms and 1 m SDD. (Piranha, RTI Electronics, Mölndal, Sweden). This study was measured the ESD under the conditions of 80 kVp tube voltage, 4 mAs current, 100 ms and 1 m SDD. 2.5. Chest Radiography According to Different Angle Degrees 2.5. Chest According to Different ChestRadiography radiographs were performed sixAngle timesDegrees under the conditions of different angles using the developed system with device.six Totimes assessunder the image quality according to the angulation of Chest PDR radiographs were SDA performed the conditions of different angles using the the detector in the mediolateral plane of the patient, the angle value of the SDA changed from 0° developed PDR system with SDA device. To assess the image quality according to the angulation (medial line) to in 30°the in units of 10 degrees 4). For the of chest images, and CNR of the detector mediolateral plane (Figure of the patient, theanalysis angle value of the SDA SNR changed from according to different angle degrees were measured from 15 different points using the Alderson ◦ ◦ 0 (medial line) to 30 in units of 10 degrees (Figure 4). For the analysis of chest images, SNR and Radiation Therapy (ART) angle male degrees phantom (ART-200A). The 15 image distortion according to the CNR according to different were measured from different points using the Alderson changes angulation the deformity percent changes in the to sizes cardiac RadiationofTherapy (ART)was malemeasured phantom (ART-200A). Theasimage distortion according the of changes of transverse diameter (CTD), thoracic transverse diameter (TTD), and thoracic longitudinal diameter angulation was measured the deformity as percent changes in the sizes of cardiac transverse diameter (TLD). Also, two expert radiologists (moreand thanthoracic 10 years of experience) blindly evaluated eachexpert chest (CTD), thoracic transverse diameter (TTD), longitudinal diameter (TLD). Also, two image, and (more reached a 10 consensus regarding the anatomic landmarks [19].image, The five anatomica radiologists than years of experience) blindly evaluated each chest and reached landmarks were the border of the heart, clavicle, first rib, fifth rib and scapula. Each anatomic consensus regarding the anatomic landmarks [19]. The five anatomic landmarks were the border of the landmark on chest image data was analyzed according to the radiological diagnosis on a five-point heart, clavicle, first rib, fifth rib and scapula. Each anatomic landmark on chest image data was analyzed scale: 1, definitely seen; 2, probably seen; probably not seen; and2,5,probably definitelyseen; not according to the radiological diagnosis on3,a equivocal; five-point 4, scale: 1, definitely seen; seen. 3, equivocal; 4, probably not seen; and 5, definitely not seen.
Figure 4. 4. The The angulation angulation plane plane (A) (A) and and reference reference line line setting setting (B) (B) for for the of distortion Figure the measurement measurement of distortion on chest phantom: (a) cardiac transverse diameter (CTD); (b) thoracic transverse diameter (TTD); on chest phantom: (a) cardiac transverse diameter (CTD); (b) thoracic transverse diameter (TTD); and (c) (c) thoracic thoracic longitudinal longitudinal diameter diameter (TLD). and (TLD). L L and and R R indicated indicated left left side side and and right right side. side.
Statistical Analysis 2.6. Statistical analyses were performed performed using the Statistical Package for the Social Sciences All statistical analyses Coefficient of of variance variance (CV) (CV) of SNR and CNR was (SPSS version 20.0, Chicago, IL, USA) software. Coefficient calculated for the variability of measurements in the PDR system [20]. [20]. The angle difference between and digital angle meter) was analyzed with the independent two sample both angle anglemeters meters(SDA (SDAdevice device and digital angle meter) was analyzed with the independent two t-test. According to the different degrees, variation image qualities, distortion anddistortion visibility sample t-test. According to the angle different anglethe degrees, theinvariation in image qualities, scores were analyzed with the repeated-measures analysis of variance (rmANOVA) and(rmANOVA) Tukey’s post and visibility scores were analyzed with the repeated-measures analysis of variance hoc tests. Two-sided less than 0.05 wereless considered indicate statistical and Tukey’s post hocp-values tests. Two-sided p-values than 0.05towere considered to significance. indicate statistical significance.
Sensors 2017, 17, 531 Sensors 2017, 17, 531
6 of 12 6 of 12
3. 3. Results Results 3.1. 3.1. Development Development and and Performances Performances of of PDR PDR The the PDR PDR system system were were 723 723 mm mm (width) (width) × × 650 mm (depth) (depth) × × 1376 The size size and and weight weight of of the 650 mm 1376 mm mm (height) and approximately 60 kg, respectively (Figure 1). The PDR has a smooth flexibility (height) and approximately 60 kg, respectively (Figure 1). The PDR has a smooth flexibility of of the the X-ray anan adult. TheThe mean SNR andand CNR werewere 182.47 ± 6.75 X-ray source sourceand andcan canbe bemoved movedeasily easilybyby adult. mean SNR CNR 182.47 ±(CV: 6.75 3.7%) and 141.43 ± 6.08 (4.3%), The overall values the system less were than 5%. (CV: 3.7%) and 141.43 ± 6.08 respectively. (4.3%), respectively. TheCV overall CVinvalues in thewere system less The resolution at 10% MTF (157lp/mm μm) (Figure 5) and the ESD wasthe 0.266 mGy. thanspatial 5%. The spatial resolution at was 10%3.17 MTFlp/mm was 3.17 (157 µm) (Figure 5) and ESD was Therefore, PDR system is easily applicable for imaging inpatients such as 0.266 mGy.our Therefore, our PDR system is easily applicable forbedside imagingpatients bedsideor patients or inpatients immobilized bedridden patients and post-stroke hemiplegic patients. such as immobilized bedridden patients and post-stroke hemiplegic patients.
Figure 5. Bar phantom image obtained from the PDR to calculate the signal-to-noise ratio (SNR), Figure 5. Bar phantom image obtained from the PDR to calculate the signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR) and spatial resolution (A). Line-squares on the phantom images contrast-to-noise ratio (CNR) and spatial resolution (A). Line-squares on the phantom images indicated indicated theofregions forCNR SNRmeasurements. and CNR measurements. A modulation transfer function the regions interestof forinterest SNR and A modulation transfer function (MTF, 10%) (MTF, 10%) curve was used for resolution evaluation (B). The 10% MTF value is 3.17 lp/mm (157μm) curve was used for resolution evaluation (B). The 10% MTF value is 3.17 lp/mm (157µm) on the image on the image obtained fromobtained the PDR.from the PDR.
3.2. Performance of the SDA Device 3.2. Performance of the SDA Device Figure 3 shows the features of the SDA device on the PDR system. The device was attached to Figure 3 shows the features of the SDA device on the PDR system. The device was attached the X-ray source and detector, respectively, and its size was 55 mm (width) × 80 mm (depth) × 35 mm to the X-ray source and detector, respectively, and its size was 55 mm (width) × 80 mm (depth) × (height). The angle values obtained from the SDA device were displayed in real time on the LCD 35 mm (height). The angle values obtained from the SDA device were displayed in real time on the monitor (1 times/s). The averaged angle values of the SDA device and digital angle meter are LCD monitor (1 times/s). The averaged angle values of the SDA device and digital angle meter are summarized in Table 1, and the mean difference in both angle values was 0.71° ± 0.15°. There was no summarized in Table 1, and the mean difference in both angle values was 0.71◦ ± 0.15◦ . There was no significant angle difference between the digital angle meter and our SDA device (p > 0.05). significant angle difference between the digital angle meter and our SDA device (p > 0.05). 3.3. 3.3. Chest Chest Radiographic Radiographic Study Study According According to to Different Different Angles Angles for for Clinical Clinical Application Application Figure Figure 66 shows shows the the representative representative chest chest AP AP images images obtained obtained from from the the developed developed PDR PDR according according to different angles. Imaging quality, distortion and visibility scores are summarized in Table 2. SNR2. to different angles. Imaging quality, distortion and visibility scores are summarized in Table and CNR values were not significantly different from different the angle degrees (rmANOVA, p> SNR and CNR values were not significantly different from different the angle degrees (rmANOVA, 0.05). TheThe objective image qualities were notnot significantly different ininthis p > 0.05). objective image qualities were significantly different thisstudy; study;thus, thus,ititcould could be be considered to be applicable for appropriate angle degrees in clinical experiments. However, the sizes considered to be applicable for appropriate angle degrees in clinical experiments. However, the sizes and and deformity deformity (%) (%) in in the the CTD, CTD, TTD TTD and and TLD TLD increased increased according according to to the the increment increment of of angulation angulation (rmANOVA, (rmANOVA, pp
