European Journal of Radiology 92 (2017) 37–44
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
Research paper
Difference in the craniocaudal gradient of the maximum pixel value change rate between chronic obstructive pulmonary disease patients and normal subjects using sub-mGy dynamic chest radiography with a flat panel detector system
MARK
⁎
Yoshitake Yamadaa,b, , Masako Ueyamac,1, Takehiko Abed,2, Tetsuro Arakia,3, Takayuki Abee,4, ⁎⁎ Mizuki Nishinoa,5, Masahiro Jinzakib,6, Hiroto Hatabua, , Shoji Kudohf,7 a
Department of Radiology, Center for Pulmonary Functional Imaging, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA Department of Diagnostic Radiology, Keio University School of Medicine, Tokyo, Japan Department of Health Care, Fukujuji Hospital, Japan Anti-Tuberculosis Association, Kiyose, Japan d Department of Radiology, Fukujuji Hospital, Japan Anti-Tuberculosis Association, Kiyose, Japan e Department of Preventive Medicine and Public Health, Biostatistics Unit at Clinical and Translational Research Center, Keio University School of Medicine, Tokyo, Japan f Department of Respiratory Medicine, Fukujuji Hospital, Japan Anti-Tuberculosis Association, Kiyose, Japan b c
A R T I C L E I N F O
A B S T R A C T
Keywords: X-ray Respiration Ventilation Lung Thorax
Objectives: To compare the craniocaudal gradients of the maximum pixel value change rate (MPCR) during tidal breathing between chronic obstructive pulmonary disease (COPD) patients and normal subjects using dynamic chest radiography. Materials and methods: This prospective study was approved by the institutional review board and all participants provided written informed consent. Forty-three COPD patients (mean age, 71.6 ± 8.7 years) and 47 normal subjects (non-smoker healthy volunteers) (mean age, 54.8 ± 9.8 years) underwent sequential chest radiographs during tidal breathing in a standing position using dynamic chest radiography with a flat panel detector system. We evaluated the craniocaudal gradient of MPCR. The results were analyzed using an unpaired t-test and the Tukey–Kramer method. Results: The craniocaudal gradients of MPCR in COPD patients were significantly lower than those in normal subjects (right inspiratory phase, 75.5 ± 48.1 vs. 108.9 ± 42.0 s−1 cm−1, P < 0.001; right expiratory phase, 66.4 ± 40.6 vs. 89.8 ± 31.6 s−1 cm−1, P = 0.003; left inspiratory phase, 75.5 ± 48.2 vs. 108.2 ± 47.2 s−1 cm−1, P = 0.002; left expiratory phase, 60.9 ± 38.2 vs. 84.3 ± 29.5 s−1 cm−1, P = 0.002). No significant differences in height, weight, or BMI were observed between COPD and normal groups. In the subanalysis, the gradients in severe COPD patients (global initiative for chronic obstructive lung disease [GOLD] 3 or 4,
Abbreviations: BMI, body mass index; COPD, chronic obstructive pulmonary disease; CT, computed tomography; FEV, forced expiratory volume; FIR, finite impulse response; FPD, flat panel detector; GOLD, global initiative for chronic obstructive pulmonary disease; MPCR, maximum pixel value change rate; MRI, magnetic resonance imaging; SD, standard deviation; VC, vital capacity ⁎ Corresponding author at: Department of Radiology, Center for Pulmonary Functional Imaging, Brigham and Women's Hospital, Harvard Medical School; Department of Diagnostic Radiology, Keio University School of Medicine, 75 Francis St., Boston, MA 02215, USA. ⁎⁎ Corresponding author at: Department of Radiology, Center for Pulmonary Functional Imaging, Brigham and Women's Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02215, USA. E-mail addresses:
[email protected] (Y. Yamada),
[email protected] (M. Ueyama),
[email protected] (T. Abe),
[email protected] (T. Araki),
[email protected] (T. Abe),
[email protected] (M. Nishino),
[email protected] (M. Jinzaki),
[email protected] (H. Hatabu),
[email protected] (S. Kudoh). 1 Department of Health Care, Fukujuji Hospital, Japan Anti-Tuberculosis Association, 3-1-24 Matsuyama, Kiyose, Tokyo 204-8522, Japan. 2 Department of Radiology, Fukujuji Hospital, Japan Anti-Tuberculosis Association, 3-1-24 Matsuyama, Kiyose, Tokyo 204-8522, Japan. 3 Department of Radiology, Center for Pulmonary Functional Imaging, Brigham and Women's Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02215, USA. 4 Department of Preventive Medicine and Public Health, Biostatistics Unit at Clinical and Translational Research Center, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. 5 Department of Radiology, Center for Pulmonary Functional Imaging, Brigham and Women's Hospital, Harvard Medical School, 75 Francis St., Boston, MA 02215, USA. 6 Department of Diagnostic Radiology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. 7 Department of Respiratory Medicine, Fukujuji Hospital, Japan Anti-Tuberculosis Association, 3-1-24 Matsuyama, Kiyose, Tokyo 204-8522, Japan. http://dx.doi.org/10.1016/j.ejrad.2017.04.016 Received 22 February 2017; Received in revised form 17 April 2017; Accepted 23 April 2017 0720-048X/ © 2017 The Author(s). Published by Elsevier Ireland Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
European Journal of Radiology 92 (2017) 37–44
Y. Yamada et al.
n = 26) were significantly lower than those in mild COPD patients (GOLD 1 or 2, n = 17) for both right and left inspiratory/expiratory phases (all P ≤ 0.005). Conclusions: A decrease of the craniocaudal gradient of MPCR was observed in COPD patients. The craniocaudal gradient was lower in severe COPD patients than in mild COPD patients.
(caption on next page)
38
European Journal of Radiology 92 (2017) 37–44
Y. Yamada et al.
Fig. 1. Image analysis. (a) Determining lung area on a dynamic chest radiograph and the lung area division into small blocks. The lung areas from the dynamic chest radiographs were determined manually, without touching the chest wall, diaphragm, or mediastinum, by a board-certified radiologist with 14 years of experience in interpreting chest radiography (left). Although the size of the lung field changed during breathing, subsequent analyses were conducted only within the smallest lung field at the resting end-expiratory position. The determined lung areas were then divided into small blocks of 5 × 5 pixels (1.94 mm × 1.94 mm) (right). (b) A typical example of the averaged pixel value change in one small block (upper panel), the averaged pixel value in one small block after low-pass filtering (middle panel), the pixel value change rate of each small block (lower panel). The average pixel value PAVE(x, y, t) of each small block was calculated (upper panel). Although a variety of factors influence the changes in X-ray translucency of the lungs during dynamic chest radiography, two of the most influential factors are breathing and blood flow. The 31-order finite impulse response (FIR) low-pass filter (Parks-McClellan algorithm, fc = 0.85 Hz) was applied to extract the low frequency components representing the changes of the pixel values related to ventilation (middle panel). The pixel value change rate PCR(x, y, t) of each small block was calculated by means of the inter-frame subtraction (lower panel) and the maximum pixel value change rate (MPCR) was detected during inspiratory and expiratory phases independently (lower panel, blue and red double-arrows). (c) Maximum pixel value change rate (MPCR) in inspiratory phase. (d) Calculation of the craniocaudal gradient of MPCR. The craniocaudal gradient of MPCR (c) was calculated by estimating the slope of the distance from the lung apex in the craniocaudal direction (X axis) versus the maximum pixel value change rate (Y axis) (d).
Table 1 Demographic characteristics of the study population. Demographic variables
Normal subjects (n = 47) Mean ± SD (range) or n (%)
COPD patients (n = 43) Mean ± SD (range) or n (%)
Comparison between normal subjects and COPD patients P valueb
Age (years) Female/male (%) Height (cm) Weight (kg) BMI (kg/m2) Pulmonary function test Tidal volume (L) VC (L) %VC (%) FEV1 (L) FEV1% (%) %FEV1 (%)
54.8 ± 9.8 (36–72) 27 (57.4%)/20 (42.6%) 162.2 ± 9.3 (146.1–183.7) 58.9 ± 10.4 (37.0–78.0) 22.3 ± 3.0 (15.1–31.1)
71.6 ± 8.7 (48–85) 5 (11.6%)/38 (88.4%) 163.0 ± 6.3 (149.0–176.0) 57.1 ± 10.5 (42.0–94.0) 21.5 ± 3.5 (16.3–34.5)
< 0.001a < 0.001a 0.654 0.434 0.251
0.75 ± 0.35 (0.33–1.76) 3.34 ± 0.87 (2.11–5.70) 110.4 ± 14.5 (92.1–159.6) 2.71 ± 0.73 (1.58–4.72) 82.4 ± 6.2 (70.5–97.0) 107.1 ± 14.8 (80.6–163.9)
0.98 2.71 87.7 1.25 47.6 49.1
0.003a < 0.001a < 0.001a < 0.001a < 0.001a < 0.001a
± ± ± ± ± ±
0.35 0.72 22.3 0.49 10.4 19.8
(0.39–2.30) (1.49–4.34) (44.6–140.0) (0.32–2.41) (20.9–68.9) (11.8–97.9)
a
Indicates P < 0.05. P values were calculated using Student's t test or chi-square test. SD, standard deviation; BMI, body mass index; VC, vital capacity; FEV, forced expiratory volume. b
1. Introduction
gradients of the maximum pixel value change rate (MPCR) during tidal breathing between COPD patients and normal subjects using dynamic chest radiography.
Chronic obstructive pulmonary disease (COPD) is a condition defined as incompletely reversible expiratory airflow obstruction due to the exposure of noxious inhaled particulates [1], and is one of the leading causes of morbidity and mortality worldwide [2]. The diagnosis of COPD is based on the results of pulmonary function tests; however, it is increasingly clear that spirometric measures of lung function alone are inadequate for a complete understanding of the impact of disease and are insufficient for the categorization of disease severity [3]. Recently, dynamic chest radiography using a flat panel detector (FPD), which is performed as an additional examination in conventional chest radiography and has a large field of view, was introduced for clinical use. This technique enables one to obtain sequential chest radiographs with high temporal resolution during respiration [4]. The radiation dose of dynamic chest radiography is much lower than that of computed tomography (CT), and its cost is lower than that of CT or magnetic resonance imaging (MRI) [5]. Also, while CT and MRI are performed in a supine or prone position, dynamic chest radiography can be performed in a standing or sitting position, which reflects physiologically relevant daily activity. Dynamic chest radiographs contain a wealth of functional information, such as diaphragm movement, cardiac motion, pulmonary ventilation, and circulation [6]. Most importantly, dynamic chest radiography can detect changes in X-ray translucency (radiographic pixel value) that are related to air ventilation [7]. In dynamic chest radiographs, the pixel value decreases according to an increase in air volume in the lung during the inspiratory phase; in contrast, the pixel value increases according to a decrease in air volume in the lung during the expiratory phase [6]. To the best of our knowledge, no detailed clinical study has compared the ventilation information obtained by dynamic chest radiography between COPD patients and normal subjects. We hypothesized that the distribution of the pixel value change rates during respiration in COPD patients is different from that in normal subjects. The purpose of this study was to compare the craniocaudal
2. Materials and methods 2.1. Study population This study was approved by our institutional review board and all the participants provided written informed consent. From June 2009 to August 2011, 43 consecutive COPD patients who met the following inclusion criteria for the study were recruited: (1) clinical diagnosis of pure COPD based on clinical course, clinical symptoms, chest CT scans, and laboratory data, including airflow limitation assessed by pulmonary function tests with post-bronchodilator inhalation, without acute respiratory infection or other respiratory diseases such as bronchiectasis or any type of interstitial lung disease; (2) current or ex-smokers; (3) over 20 years old; (4) scheduled for conventional chest radiography; (5) ability to follow instructions for tidal breathing. No patients were excluded from these eligible 43 COPD patients. Thus, a total of 43 COPD patients (38 men, 5 women; mean age, 71.6 ± 8.7 years; age range, 48–85 years) were finally included in the analysis. A normal (control) group of 49 consecutive volunteers who visited the health screening center of our hospital from May 2013 to February 2014 and met the following inclusion criteria was also recruited for the study: (1) over 20 years old; (2) scheduled for conventional chest radiography; (3) normal pulmonary function test results (i.e., percent vital capacity (% VC) > 80% and forced expiratory volume/forced VC (FEV1%) > 70%); (4) ability to follow tidal breathing instructions; (4) no smoking history; (5) no past medical history of respiratory diseases. Volunteers with any of the following criteria were excluded: (1) pregnant or potentially pregnant or lactating (n = 0); (2) incomplete dynamic chest radiography data sets (n = 1); (3) diaphragmatic motion could not be analyzed by the software described below (n = 0); (4) suspected malnourishment (body weight < 30 kg) (n = 1). Thus, a 39
European Journal of Radiology 92 (2017) 37–44
< 0.001 < 0.001 48.7 ± 34.9 (−31.0 to 119.9) 43.5 ± 27.4 (1.4 to 93.0)
total of 47 normal subjects (20 men, 27 women; age, 54.8 ± 9.8 years; age range, 36–72 years) were included in the analysis as a control group. For sub-analysis, 43 COPD patients were classified into a mild COPD group (percent forced expiratory volume [%FEV1] ≥ 50% predicted; global initiative for chronic obstructive lung disease [GOLD] 1 or 2, n = 17) or a severe COPD group (%FEV1 < 50% predicted; GOLD 3 or 4, n = 26). Images for the 39 COPD patients and 47 normal subjects of this study population were analyzed for a different purpose in a previous study that evaluated diaphragmatic motion but did not evaluate pixel value change in the lung [8]. The height and weight of the participants were measured, and the body mass index (BMI, weight in kilograms divided by height in meters squared) was calculated. 2.2. Imaging protocol of dynamic chest radiology Posteroanterior dynamic chest radiography (“dynamic X-ray phrenicography”) was performed using a prototype system (Konica Minolta, Inc., Tokyo, Japan) composed of an FPD (PaxScan 4030CB, Varian Medical Systems, Inc., UT, USA) and a pulsed X-ray generator (DHF155HII with cineradiography option, Hitachi Medical Corporation, Tokyo, Japan) [5,8]. All the subjects were scanned in the standing position and instructed to breathe normally in a relaxed way without deep inspiration/expiration (tidal breathing). The exposure conditions were as follows: tube voltage, 100 kV; tube current, 50 mA; duration of pulsed X-ray, 1.6–3.2 ms; source-to-image distance, 2 m; additional filter, 0.5 mm Al + 0.1 mm Cu. The additional filter was used to reduce the low-energy component (soft X-rays). The exposure time was approximately 10–15 s. The pixel size was 388 μm × 388 μm, the matrix size was 1024 × 768, and the overall image area was 40 cm × 30 cm. The dynamic image data captured at the frame interval Δt = 1/7.5 s were synchronized with the pulsed X-ray. (The pulsed Xray prevented excessive radiation exposure to the subjects.) The total entrance surface dose during the exposure time was approximately 0.3–1.0 mGy, which was calculated as the product of the entrance surface dose for a single frame and the number of frames during the exposure time. In this calculation, source-to-skin surface distance was assumed to be 174.5 cm and the entrance surface dose for a single frame at that point was estimated from the air kerma measured at 180 cm from the X-ray source and a back scatter factor. The dosimeter (RaySafe Xi Platinum Plus, Unfors RaySafe AB, Billdal, Sweden) was used for measuring the air kerma. The range of the signal intensity (incident exposure) in FPD was 16384 (14 bits), corresponding to approximately 29 mGy. In general, the pixel value of the gray-scale chest radiography was inversely proportional to the logarithm of the signal intensity. Therefore, in this study, the pixel value P(x, y, t) was represented as the logarithm of the signal intensity I(x, y, t) and normalized with a maximum value of I(x, y, t) as follows:
P values were calculated using Student's t test. P values were calculated using Tukey–Kramer method.
⎡ ln I (x, y, t ) ⎤ P (x, y, t ) = 10 3⋅⎢1 − ⎥ ⎣ ln 214 ⎦ where ln I(x,y,t) = 0 for all I(x,y,t) = 0. x, y: The coordinates of blocks in the horizontal and craniocaudal directions. t = n · Δt: The time at frame number n with frame interval Δt. 2.3. Determination of lung areas and division into small blocks The dynamic image data on sequential chest radiographs during tidal breathing were analyzed using prototype software (Konica Minolta, Inc., Tokyo, Japan) installed in an independent workstation (Operating system: Windows 7 Pro SP1, Microsoft, Redmond WA; CPU: Intel® Core™ i5-4590, 3.30 GHz; memory 16 GB). The lung areas from the dynamic chest radiographs were determined manually, without touching the chest wall, diaphragm, or mediastinum, by a boardcertified radiologist with 14 years of experience in interpreting chest radiography. Although the size of the lung field changed during
b
a
66.4 ± 40.6 (−31.0 to 148.0) 60.9 ± 38.2 (1.4 to 163.3) 89.8 ± 31.6 (29.3 to 163.9) 84.3 ± 29.5 (31.6 to 158.5) Right Left
0.003 0.002
93.5 ± 33.6 (25.2 to 148.0) 87.5 ± 37.4 (26.0 to 163.3)
0.005 0.001 58.5 ± 47.0 (−22.8 to 210.0) 55.2 ± 34.4 (−1.0 to 133.2) 101.5 ± 38.0 (26.1 to 181.0) 106.5 ± 50.6 (30.5 to 198.8) 75.5 ± 48.1 (−22.8 to 210.0) 75.5 ± 48.2 (−1.0 to 198.8) 108.9 ± 42.0 (45.6 to 235.6) 108.2 ± 47.2 (35.0 to 251.6) Right Left
Craniocaudal gradient of MPCR (inspiratory phase) (s−1 cm−1) Craniocaudal gradient of MPCR (expiratory phase) (s−1 cm−1)
Means ± SD (range) Means ± SD (range)
COPD patients all (n = 43)
< 0.001 0.002
Means ± SD (range)
Means ± SD (range)
Comparison between COPD mild and COPD severe P valueb COPD severe (GOLD 3 or 4) (n = 26) COPD mild (GOLD 1 or 2) (n = 17) Normal subjects (n = 47)
Comparison between normal subjects and COPD patients P valuea
Sub-analysis Main analysis Parameters
Table 2 Craniocaudal gradient of the maximum pixel value change rate (MPCR) in normal subjects and COPD patients with the sub-analysis between mild COPD and severe COPD patients.
Y. Yamada et al.
40
European Journal of Radiology 92 (2017) 37–44
Y. Yamada et al.
Fig. 2. Comparison of the craniocaudal gradients of MPCR between normal subjects and COPD patients (main analysis) and among normal subjects, mild COPD, and severe COPD patients (sub-analysis) expressed as medians: (left upper) right inspiratory phase; (left lower) right expiratory phase; (right upper) left inspiratory phase; (right lower) left expiratory phase. Boxes: upper to lower quartile. Thin lines: maximum and minimum (excluding outliers and extreme values). Student's t test was used to compare the craniocaudal gradients of MPCR between normal subjects and COPD patients (main analysis). Tukey–Kramer method was used to compare the craniocaudal gradients of MPCR among normal subjects, mild COPD, and severe COPD patients (sub-analysis). NS indicates not significant. ** indicates P < 0.01.
the most influential factors are breathing and blood flow. Generally, the frequency of breathing in adults is 12–20 cycles per minute, while the heart rate is between 60–90 beats per minute. In this study, the 31order finite impulse response (FIR) low-pass filter (Parks-McClellan algorithm, fc = 0.85 Hz) was applied in order to extract the low frequency components representing the changes of the pixel values related to ventilation, and to eliminate the high frequency components representing the changes of pixel values related to blood flow. The middle panel of Fig. 1b shows the averaged pixel value in one small block after low-pass filtering, FPAVE(x, y, t).
breathing, subsequent analyses were conducted only within the smallest lung field at the resting end-expiratory position. The determined lung area was then divided into small blocks of 5 × 5 pixels (1.94 mm × 1.94 mm) and the average pixel value PAVE(x, y, t) of each small block was calculated in order to reduce the noise associated with the determination. The left panel of Fig. 1a shows a typical example of the determined lung area and the right panel of Fig. 1a shows the lung area division into small blocks. 2.4. Plotting each small block's pixel value and the removal of heartbeat signals
2.5. Calculation of the craniocaudal gradient of MPCR
The upper panel of Fig. 1b shows a typical example of the averaged pixel value change in one small block. In the inspiratory phase, X-ray translucency increased and then the pixel value decreased. On the other hand, in the expiratory phase, X-ray translucency decreased and then the pixel value increased. Because the positive and negative peak of the pixel value change was sometimes unclear, the inspiratory and expiratory phases were determined by assessing diaphragm motion. The distance between the lung apex and the diaphragm was measured; increasing distance corresponded to the inspiratory phase and decreasing distance corresponded to the expiratory phase. Although a variety of factors influences the changes in X-ray translucency of the lungs during dynamic chest radiography, two of
The pixel value change rate PCR(x, y, t) of each small block was calculated by means of the inter-frame subtraction (lower panel of Fig. 1b) and the maximum of pixel value change rate MPCR(x, y, t) was detected automatically. The MPCR was detected during inspiratory and expiratory phases independently (Fig. 1e, blue and red double-arrows).
PCR (x, y, t ) =
FPAVE (x, y, n ⋅ Δt ) − FPAVE (x, y, (n − 1) ⋅ Δt ) Δt
MPCR (x, y, t ) = max PCR (x, y, t ) The craniocaudal gradient of MPCR was calculated as follows: (1) Calculate the averaged maximum pixel value change rate at each 41
European Journal of Radiology 92 (2017) 37–44
Y. Yamada et al.
Fig. 3. Representative craniocaudal gradients of MPCR in 3 cases (expiratory phases): (left) normal subject, 60 year old male; (center) mild COPD (GOLD 2) patient, 78 year old male; (right) severe COPD (GOLD 4) patient, 73 year old male.
Fig. 4. Estimated regression line of the craniocaudal gradients of MPCR in expiratory phase on percent forced expiratory volume (%FEV1). (a) Association between %FEV1 and right craniocaudal gradients of MPCR in expiratory phase. (b) Association between %FEV1 and left craniocaudal gradients of MPCR in expiratory phase. Lines show estimated regression (a, b). All scatterplots show correlations (P < 0.0001).
inspiratory and expiratory phases independently; in this study, the αM was defined as the craniocaudal gradient of MPCR (Fig. 1d).
craniocaudal direction y (Fig. 1c): N
AveMPCR (y ) =
∑ MPCR (xi , y)/N i =1
2.6. Statistical analysis
N: the number of xi ∈ defined lung area at depth y.
• Calculate
Descriptive statistics are expressed as mean ± standard deviation (SD) for continuous variables and as frequency and percentages for nominal variables. The participants’ demographic characteristics and the craniocaudal gradients of MPCR between normal subjects and COPD patients were compared using the Student's t-test for continuous variables and the chi-square test for nominal/categorical variables. The craniocaudal gradients of MPCR among the normal subjects, mild COPD, and severe COPD patients were compared using the Tukey–Kramer method because of multiple comparisons. The association between %FEV1 and the craniocaudal gradient of MPCR in the
the distance from the lung apex in the craniocaudal
direction.
D (y ) = 0.194⋅y (cm)
• Plot
D(y) versus AveMPCR (y )⋅10 3 and estimate the slope of a regression line αM. This slope of a regression line αM was estimated during the 42
European Journal of Radiology 92 (2017) 37–44
Y. Yamada et al.
inspiratory/expiratory flow rate during respiration to some extent. Our study showed a substantial craniocaudal gradient of MPCR in the normal subjects. Possible reasons for this are that the lower lungs undergo greater volume change during breathing compared with the upper lungs because of the effects of gravity on lung recoil, and that the diaphragm contributes to respiration more than the intercostal muscles during tidal breathing [13,14]; thus, the MPCR in the lower lung field would be higher than in the upper lung field because of greater motion of the lower lung compared with the upper lung. Although we cannot explain the exact reasons for the decrease of the craniocaudal gradients in COPD patients and the association between %FEV1 and the craniocaudal gradient, we speculate that airflow obstruction in COPD (especially in severe COPD) may more strongly affect the airways of the lower lung field compared with those of the upper lung field. The sub-analysis of this study showed that there were no significant differences in the craniocaudal gradients of MPCR between normal subjects and mild COPD patients because of the overlap of the values. However, there was a moderate linear correlation between %FEV1 (representing the severity of COPD) and the craniocaudal gradient (r = 0.50); therefore, the small sample size in the mild COPD group in this study (i.e., only 17 patients) may have affected the results of the sub-analysis between normal subjects and mild COPD patients. Conventionally, ventilation lung scintigraphy is used to evaluate pulmonary ventilation. However, ventilation lung scintigraphy requires nuclear isotopes and special equipment, has low spatial-resolution [15], and takes a much longer time than dynamic chest radiography. On the other hand, dynamic chest radiography is performed in a way that is almost identical to conventional chest radiography; has high-spatial resolution with flat panel detector; and does not require nuclear isotopes, special equipment for a radioisotope, or a contrast medium. While CT and MRI now allow both dynamic and static evaluation of regional ventilation and simultaneous acquisition of anatomical and ventilation images [15–17], they pose significant problems: CT exposes patients to ionizing radiation, particularly with serial scanning [18], and MRI is of limited utility due to high cost, low-accessibility, and lowthroughput. On the contrary, dynamic chest radiography is performed in seconds with a radiation dose comparable to conventional posteroanterior plus lateral chest radiography [19], and does not require significant preparation, although it is 2-dimentional imaging. These features could be suited for longitudinal follow-up of COPD during treatment in addition to the initial diagnosis. Our study has several limitations. First, we included only 47 normal subjects and 43 COPD patients at a single institution, and further studies in larger patient populations are required for confirming these preliminary findings, especially those regarding the sub-analysis between normal patients and mild COPD patients. Second, we did not assess inter-observer or intra-observer variability. Third, in this study, the soft tissues surrounding the lungs and rib movement over the lung fields may have affected the attenuation of the X-rays. However, the craniocaudal gradient of MPCR in this study was calculated as the estimated slope of a regression line of MPCR, thus being averaged in the whole lung; therefore, the noise by the tissues surrounding the lungs and rib movement in the lung fields were reduced in this study. In conclusion, a decrease of the craniocaudal gradient of MPCR during tidal breathing was observed in COPD patients. The craniocaudal gradient was lower in severe COPD patients than in mild COPD patients.
expiratory phase was evaluated by means of the Pearson's correlation coefficient. The significance level for all tests was 5% (two-sided). All data were analyzed using a commercially available software program (JMP; version 12, SAS, Cary, North Carolina, USA). 3. Results 3.1. Participant characteristics Table 1 shows the demographic characteristics of the normal subjects and the COPD patients. No significant differences in height, weight, or BMI were observed between the normal subjects and COPD patients. 3.2. Comparison of the craniocaudal gradients of MPCR between COPD patients and normal subjects The craniocaudal gradients of MPCR in COPD patients were significantly lower than those in normal subjects (right inspiratory phase, 75.5 ± 48.1 vs. 108.9 ± 42.0 s−1 cm−1, P < 0.001; right expiratory phase, 66.4 ± 40.6 vs. 89.8 ± 31.6 s−1 cm−1, P = 0.003; left inspiratory phase, 75.5 ± 48.2 vs. 108.2 ± 47.2 s−1 cm−1, P = 0.002; left expiratory phase, 60.9 ± 38.2 vs. 84.3 ± 29.5 s−1 cm−1, P = 0.002) (Table 2). In the sub-analysis, the gradients in severe COPD patients (GOLD 3 or 4, n = 26) were significantly lower than those in mild COPD patients (GOLD 1 or 2, n = 17) for both right and left inspiratory/expiratory phases (all P ≤ 0.005) (Table 2), while there were no significant differences in the gradients between normal subjects and mild COPD patients (Fig. 2). Fig. 3 illustrates the representative craniocaudal gradients of MPCRs of 3 cases: a normal subject, mild COPD (GOLD 2), and severe COPD (GOLD 4) patients. 3.3. Association between %FEV1 and the craniocaudal gradient of MPCR in expiratory phase Higher %FEV1 was associated with an increased craniocaudal gradient of MPCR in the expiratory phase (right, r = 0.49, P < 0.0001; left, r = 0.49, P < 0.0001) (Fig. 4a and b). 4. Discussion Our study demonstrated that the craniocaudal gradients of MPCR obtained by dynamic chest radiography during tidal breathing were significantly lower in COPD patients than in normal subjects, and that the gradients in severe COPD patients were significantly lower than those in mild COPD patients. These findings are noteworthy because the sub-mGy dynamic chest radiography was able to detect kinetic and physiological differences between normal subjects and COPD patients and could detect the severity of COPD during tidal breathing, without the need for forced breathing. Currently, pulmonary function tests are performed during forced breathing and it has been reported that forced expiration could be difficult for the elderly and for subjects with cognitive impairment [9,10]. Elderly patients are reported to be at risk for misdiagnosis and inappropriate treatment of respiratory disease [11,12], which may be compounded by pulmonary function test under utilization and the inappropriate acceptance of suboptimal test quality, due to low performance expectations [10]. Thus, dynamic chest radiography is a potentially useful method for the quantitative evaluation of ventilation, especially in elderly patients and subjects with cognitive impairment, because it can be performed during tidal breathing without a subject's active cooperation. Since respiration influences the pixel values in a radiograph and the pixel value change rate in this study was calculated as the first derivation (differential) of the change in the pixel value, the pixel value change rate (i.e., pixel value change speed) could reflect
Source of funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflicts of interest HH received a research grant from Konica Minolta, Inc. The 43
European Journal of Radiology 92 (2017) 37–44
Y. Yamada et al.
[6] R. Tanaka, Dynamic chest radiography: flat-panel detector (FPD) based functional X-ray imaging, Radiol. Phys. Technol. 9 (2) (2016) 139–153. [7] R. Tanaka, S. Sanada, M. Suzuki, T. Kobayashi, T. Matsui, H. Inoue, N. Yoshihisa, Breathing chest radiography using a dynamic flat-panel detector combined with computer analysis, Med. Phys. 31 (8) (2004) 2254–2262. [8] Y. Yamada, M. Ueyama, T. Abe, T. Araki, T. Abe, M. Nishino, M. Jinzaki, H. Hatabu, S. Kudoh, Difference in diaphragmatic motion during tidal breathing in a standing position between COPD patients and normal subjects: time-resolved quantitative evaluation using dynamic chest radiography with flat panel detector system (“dynamic X-ray phrenicography”), Eur. J. Radiol. 87 (2017) 76–82. [9] N. Carvalhaes-Neto, H. Lorino, C. Gallinari, S. Escolano, A. Mallet, F. Zerah, A. Harf, I. Macquin-Mavier, Cognitive function and assessment of lung function in the elderly, Am. J. Respir. Crit. Care Med. 152 (5 Pt 1) (1995) 1611–1615. [10] J.M. Haynes, Pulmonary function test quality in the elderly: a comparison with younger adults, Respir. Care 59 (1) (2014) 16–21. [11] L. Dow, L. Fowler, L. Phelps, K. Waters, D. Coggon, A.L. Kinmonth, S.T. Holgate, Prevalence of untreated asthma in a population sample of 6000 older adults in Bristol, UK, Thorax 56 (6) (2001) 472–476. [12] M.J. Connolly, Obstructive airways disease: a hidden disability in the aged, Age Ageing 25 (4) (1996) 265–267. [13] A. Ratnovsky, D. Elad, P. Halpern, Mechanics of respiratory muscles, Respir. Physiol. Neurobiol. 163 (1–3) (2008) 82–89. [14] R. Tanaka, S. Sanada, N. Okazaki, T. Kobayashi, M. Suzuki, T. Matsui, O. Matsui, Detectability of regional lung ventilation with flat-panel detector-based dynamic radiography, J. Digit. Imaging 21 (1) (2008) 109–120. [15] E.J. Chae, J.B. Seo, H.W. Goo, N. Kim, K.S. Song, S.D. Lee, S.J. Hong, B. Krauss, Xenon ventilation CT with a dual-energy technique of dual-source CT: initial experience, Radiology 248 (2) (2008) 615–624. [16] C. Salito, A. Aliverti, D.S. Gierada, G. Deslee, R.A. Pierce, P.T. Macklem, J.C. Woods, Quantification of trapped gas with CT and 3 He MR imaging in a porcine model of isolated airway obstruction, Radiology 253 (2) (2009) 380–389. [17] Y. Ohno, H. Koyama, M. Nogami, D. Takenaka, S. Matsumoto, M. Obara, K. Sugimura, Dynamic oxygen-enhanced MRI versus quantitative CT: pulmonary functional loss assessment and clinical stage classification of smoking-related COPD, Am. J. Roentgenol. 190 (2) (2008) W93–W99. [18] A. Sodickson, P.F. Baeyens, K.P. Andriole, L.M. Prevedello, R.D. Nawfel, R. Hanson, R. Khorasani, Recurrent CT, cumulative radiation exposure, and associated radiation-induced cancer risks from CT of adults, Radiology 251 (1) (2009) 175–184. [19] IAEA, International basic safety standards for protection against ionizing radiation and for the safety of radiation sources, International Atomic Energy Agency Vienna, Safety Report Series No. 115, 1996.
remaining authors (YY, MU, TA, TA, TA, MN, MJ, and SK) have no potential conflicts of interest related with this article. Acknowledgements The authors acknowledge the valuable assistance of Hideo Ogata MD PhD, Norihisa Motohashi MD PhD, Misako Aoki MD, Yuka Sasaki MD PhD, and Hajime Goto MD PhD from the Department of Respiratory Medicine; Yuji Shiraishi MD PhD from the Department of Respiratory Surgery; and Masamitsu Ito MD PhD, Atsuko Kurosaki MD, Yoichi Akiyama RT, Kenta Amamiya RT, and Kozo Hanai RT PhD from the Department of Radiology, Fukujuji Hospital for their important suggestions. The authors also acknowledge the valuable assistance of Alba Cid MS for editorial work on the manuscript. Yoshitake Yamada MD PhD is the recipient of a research fellowship from the Uehara Memorial Foundation. References [1] K.F. Rabe, S. Hurd, A. Anzueto, P.J. Barnes, S.A. Buist, P. Calverley, Y. Fukuchi, C. Jenkins, R. Rodriguez-Roisin, C. van Weel, J. Zielinski, Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary, Am. J. Respir. Crit. Care Med. 176 (6) (2007) 532–555. [2] N.A. Hanania, Chronic obstructive pulmonary disease (COPD) is a major cause of morbidity and mortality worldwide, Respir. Med. 106 (Suppl. 2) (2012) S1–S2. [3] G.R. Washko, The role and potential of imaging in COPD, Med. Clin. N. Am. 96 (4) (2012) 729–743. [4] R. Tanaka, S. Sanada, T. Kobayashi, M. Suzuki, T. Matsui, O. Matsui, Computerized methods for determining respiratory phase on dynamic chest radiographs obtained by a dynamic flat-panel detector (FPD) system, J. Digit. Imaging 19 (1) (2006) 41–51. [5] Y. Yamada, M. Ueyama, T. Abe, T. Araki, T. Abe, M. Nishino, M. Jinzaki, H. Hatabu, S. Kudoh, Time-resolved quantitative analysis of the diaphragms during tidal breathing in a standing position using dynamic chest radiography with a flat panel detector system (“dynamic X-ray phrenicography”): initial experience in 172 volunteers, Acad. Radiol. 24 (4) (2017) 393–400.
44
