- Email: [email protected]
CT Angiography in Patients with Peripheral Arterial Disease
ARTICLE IN PRESS
Original Investigation
CT Angiography in Patients with Peripheral Arterial Disease: Effect of Small Focal Spot Imaging and Iterative Model Reconstruction on the Image Quality Seitaro Oda, MD, PhD, Akira Yoshimura, MD, Keiichi Honda, RT, Yuji Iyama, MD, Kazuhiro Katahira, MD, PhD, Takeshi Nakaura, MD, PhD, Daisuke Utsunomiya, MD, PhD, Yoshinori Funama, PhD, Hideaki Yuki, MD, PhD, Masafumi Kidoh, MD, PhD, Kenichiro Hirata, MD, Narumi Taguchi, MD, Shinichi Tokuyasu, AD, Yasuyuki Yamashita, MD, PhD Rationale and Objectives: We investigated the effects of small focal spot (SFS) imaging and iterative model reconstruction (IMR) on the image quality of computed tomography angiographs (CTA) in patients with peripheral arterial disease. Materials and Methods: We divided 60 consecutive patients with suspected or confirmed peripheral artery disease into two equal groups. One group underwent large focal spot scanning under our standard CTA protocol with hybrid iterative reconstruction (iDose4) (protocol 1), and the other underwent scanning with the SFS protocol and IMR (protocol 2). Quantitative image quality parameters, ie, arterial computed tomography attenuation, image noise, and the contrast-to-noise ratio, were compared and the visual image quality (depiction of each vessel) was scored on a 5-point scale. Results: There was no significant difference in the arterial attenuation among all evaluated slice levels. The mean image noise was significantly lower under protocol 2 and the contrast-to-noise ratio was significantly higher at all slice levels. The visual scores assigned to the two protocols for the depiction of large vessels, such as the abdominal aorta and iliac artery, were comparable. However, the mean visual scores for small vessels in the lower extremities were significantly higher under protocol 2. Conclusion: CTA with SFS and IMR yielded significantly better qualitative and quantitative image quality especially for small vessels. Key Words: CT angiography; peripheral arterial disease; iterative reconstruction; small focal spot; image quality. © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION Peripheral artery disease (PAD) is a common, chronic, progressive health problem (1). It affects up to 8.5 million (7.2%) Americans in their 40s and is associated with significant morbidity and mortality (2). The 5-, 10-, and 15-year morbidity and mortality rates from all causes in patients with PAD are approximately 30%, 50%, and 70%, respectively. Coronary artery disease is the most common cause of death among Acad Radiol 2016; ■:■■–■■ From the Department of Diagnostic Radiology, Faculty of Life Sciences, Kumamoto University, 1-1-1 Honjyo, Chuou-ku, Kumamoto 860-8556, Japan (S.O., T.N., D.U., H.Y., M.K., K.H., N.T., Y.Y.); Department of Radiology, Kumamoto Chuo Hospital (A.Y., K.H., Y.I., K.K.); Department of Medical Physics, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan (Y.F.); CT Clinical Science, Philips Electronics Japan, Tokyo, Japan (S.T.). Received March 14, 2016; revised May 25, 2016; accepted May 30, 2016. Address correspondence to: S.O. e-mail: [email protected] © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.acra.2016.05.011
patients with PAD (40%–60%); cerebral artery disease accounts for 10%–20% of deaths (3). Early diagnosis and appropriate medical intervention can mitigate limb-specific symptoms, improve the quality of life, and decrease systemic cardiovascular risks (4). Digital subtraction angiography (DSA) is considered the reference standard for diagnosing PAD. However, it is invasive and carries limitations and risks (5). Computed tomography angiography (CTA), a less invasive and safer examination, is an alternative to DSA and has gained widespread clinical acceptance for diagnosing PAD (5). Although CTA of lower extremities is more sensitive, specific, and accurate for assessing the location and extent of peripheral artery stenosis than DSA (6), its spatial resolution is inferior to DSA, and the visualization of small vessels, such as the peripheral small artery and collateral vessels, is suboptimal. The focal spot size in the x-ray tube defines the spatial resolution of a CT system (7). Many CT tubes feature small and large focal spot (SFS and LFS) sizes. The SFS size facilitates 1
Academic Radiology, Vol ■, No ■, ■■ 2016
ODA ET AL
more detailed imaging but at the cost of x-ray intensity. Most conventional CT protocols employ the LFS because SFS imaging restricts the x-ray tube power to the lower output level to prevent overheating of the tube. Advances in CT technology have surmounted this limitation because the newer detector systems make more efficient use of the available tube power. In addition, the latest x-ray generators with advanced x-ray tube cooling systems can deliver a higher tube current (mA) to the SFS; this yields an x-ray intensity similar to LFS imaging. Furthermore, the iterative reconstruction (IR) techniques allow the use of protocols with a lower tube current (8). SFS imaging benefits from this advance. Iterative model reconstruction (IMR) is the latest advance in the field of reconstruction techniques. The IMR uses a knowledge-based approach to accurately determine the data and image statistics and the system models, which depict the geometry and physical characteristics of the CT scanner, and yields improved image quality (9,10). Under the hypothesis that the combination of SFS and IMR improves small vessel visualization on CTA in patients with PAD, we investigated the effects of SFS and IMR on the image quality of CTA scans. MATERIALS AND METHODS We obtained institutional review board (IRB) approval and prior written informed consent from all patients participating in this prospective study (Hospital IRB Number: #29-03). Study Population
Between December 2014 and January 2016, we enrolled 60 consecutive patients (41 men and 19 women; mean age: 73.4 years) with suspected or confirmed PAD. All underwent CTA. The inclusion criteria were no lower limb amputation, no renal failure (estimated glomerular filtration rate <30 mL/min/1.73 m2), no hemodialysis, and no history of allergic reactions to iodinated contrast material. The enrolled patients were randomized and scanned under one of two CTA protocols based on a random-number table. One group (n = 30) underwent scanning with our standard CTA protocol with LFS (protocol 1) and the other group (n = 30) underwent scanning under the SFS protocol (protocol 2). Patient characteristics are summarized in Table 1. CT Scanning and Contrast Infusion Protocols
All CT examinations were performed on a 256-slice CT system (Brilliance iCT; Philips Healthcare, Cleveland, OH). The parameters were detector configuration, 128 × 0.625 mm; slice thickness, 1.0 mm; section interval, 0.5 mm; gantry rotation time, 0.75 seconds; beam pitch, 0.59; tube voltage, 100 kVp; and reference tube current time product, 231 mAs (effective mAs) with auto-modulation (Dose Right; Philips Healthcare). CTA data were acquired in the craniocaudal direction 2
TABLE 1. Patient Demographics LFS Protocol SFS Protocol (n = 30) (n = 30) P (Protocol 1) (Protocol 2) Value Sex (male/female) 21/9 Age (y) 72.8 ± 8.6 Body height (cm) 157.6 ± 7.9 Body weight (kg) 61.2 ± 13.0 24.5 ± 4.3 Body mass index (kg/m2) 33.1 ± 30.4 eGFR (mL/min/1.73 m2)
20/10 74.1 ± 9.3 158.9 ± 8.3 59.7 ± 15.9 23.4 ± 4.6 32.5 ± 30.8
0.72 0.55 0.53 0.70 0.34 0.94
eGFR, estimated glomerular filtration rate; LFS, large focal spot; SFS, small focal spot. Note: Data are mean ± standard deviation.
TABLE 2. Imaging and Contrast Material Parameters of the LFS and the SFS Protocols LFS Protocol (Protocol 1) CT scanner Collimation Tube voltage Effective tube current Rotation time Helical pitch Total amount of contrast medium Injection duration Bolus tracking trigger Scan delay Image reconstruction Section thickness/interval
SFS Protocol (Protocol 2)
256-slice CT (Brilliance iCT, Philips Healthcare) 128 × 0.625 mm 100 kVp 231 eff. mAs (reference) with auto-modulation 0.75 s/rot 0.585 500 mgI/mL 25 s 150 HU (abdominal aorta) 15 s IMR iDose4 1.0/0.5 mm
CT, computed tomography; IMR, iterative model reconstruction; LFS, large focal spot; SFS, small focal spot.
from the suprarenal aorta to the ankles. Using a doublehead power injector (Auto Enhance A-250; Nemoto Kyorindo, Tokyo, Japan), iopamidol (iodine concentration 300 or 370 mgI/mL [Iopamiron 300 or 370; Bayer HealthCare, Osaka, Japan]) was injected via a 20-gauge catheter inserted into an antecubital vein. The amount of contrast material was adjusted to the body weight of each patient (500 mgI/kg) and injected at a fixed injection duration of 25 seconds. Contrast administration was followed by the injection of 40 mL of a saline solution delivered at the same injection rate as the contrast medium. An automatic bolus-tracking program was used to time the start of scanning for each phase after contrast material injection. Monitoring was at the L1 vertebral body level; a region of interest (ROI) cursor (1.0–2.0 cm2) was placed on the abdominal aorta. Data acquisition started 15 seconds after triggering. The acquisition parameters for protocols 1 and 2 are summarized in Table 2.
Academic Radiology, Vol ■, No ■, ■■ 2016
CT Image Reconstruction
We reconstructed the raw data over a field of view of 35.0– 40.0 cm; the pixel matrix was 512 × 512, and the section thickness and intervals were 1.0 and 0.5 mm, respectively. Images acquired under protocol 1 were reconstructed with hybrid IR (iDose4, level 4). Protocol 2 images were reconstructed with IMR (body routine, level 2). Original 1.0mm axial images were processed on a commercially available image-processing workstation (Virtual Place Advance Plus; AZE, Tokyo, Japan) for three-dimensional (3D) reconstruction by a radiology technologist with 15 years of experience (K.H.). The 3D reconstruction of a single case on the imageprocessing workstation took 15–20 minutes. Quantitative Image Quality Analysis
A board-certified radiologist (A.Y., with 9 years of experience reading general CT scans), blinded to the protocols used, performed measurements on axial source images. Arterial CT attenuation was measured at six levels from the abdominal aorta to the artery of the lower leg (Fig 1). For portions of levels
CT ANGIOGRAPHY IN PAD PATIENTS
2–6, the CT number of the bilateral arteries was recorded to calculate the mean CT number. When the artery on one side was occluded, only the non-occluded contralateral vessel was measured. Attempts were made to select ROIs of approximately 100, 40, 30, 20, 15, and 5 mm2 for the abdominal aorta, the common iliac, external iliac, superficial femoral, popliteal, and anterior or posterior tibial artery, respectively. Their sizes were chosen to obtain ROIs large enough not to be affected by pixel variability and small enough to avoid including the arterial wall and its calcification. The image noise, ie, the average of the standard deviation of muscle attenuation at the same slice level, was recorded. The contrast-to-noise ratio (CNR) of each artery was calculated using the equation
CNR = (HU A − HU M ) image noise , where HUA and HUM are the CT attenuation in the artery and muscle, respectively, and HUM is the CT attenuation in the muscle. These parameters were compared in protocols 1 and 2. Qualitative Image Quality Analysis
All images were reviewed and interpreted on PACS workstations (EV Insite, PSP Corp., Tokyo, Japan). Two boardcertified radiologists with 8 and 11 years of experience interpreting CTA scans were blinded to the protocols for their consensual visual evaluation of the maximum intensity projection 3D CTA images. The images were divided into the aortoiliac, femoropopliteal, and lower leg section. For qualitative analysis, we used 11 vessels of various sizes in the various locations (Fig 2). The visualization of each vessel, ie, the diagnostic quality of the images obtained with the two protocols, was graded on a 5-point scale, where 5 = excellent (excellent visualization of the entire portion of the artery), 4 = good (sufficient visualization of almost the entire portion of the artery), 3 = adequate (partially ambiguous but yielding sufficient diagnostic information), 2 = fair (insufficient visualization of some portion of the artery with partially limited diagnostic information), and 1 = uninterpretable (insufficient visualization of the entire portion of the artery). CT Radiation Dose
Based on each patient’s dose information page, we recorded the volume CT dose index (CTDIvol) in milligray (mGy) and the dose-length product (DLP) in mGy·cm for the CTA phase. Statistical Analysis Figure 1. Measurement sites for quantitative image quality analysis. Level 1: Abdominal aorta at the third lumbar vertebra level. Level 2: Common iliac artery at the iliac crest level. Level 3: External iliac artery at the hip joint level. Level 4: Superficial femoral artery at the center of the femoral bone. Level 5: Popliteal artery at the level of the patella. Level 6: Center of the anterior or posterior tibial artery.
Numerical data were expressed as mean ± standard deviation. All qualitative and quantitative image parameters of the two protocols were compared. CTDIvol and DLP were also assessed. Differences in the mean values between the two protocols with normally and non-normally distributed data were determined with the two-tailed independent t test and the 3
ODA ET AL
Academic Radiology, Vol ■, No ■, ■■ 2016
Figure 2. Evaluation sites for qualitative image quality analysis. The assessed vessels were the abdominal aorta (a), renal (b), common-external iliac (c), internal iliac (d), superior-inferior gluteal (e), superficial femoral-popliteal (f), deep femoral (g), descending branch of the lateral femoral circumflex (h), descending genicular (i), and the center of the anterior tibial (j) arteries. Measurements were also obtained at the lateral malleolus of the dorsalis pedis artery (k).
Figure 3. There was no significant difference between the two protocols with respect to arterial computed tomography attenuation at all evaluated slice levels (a). The mean image noise was significantly lower under the small focal spot (SFS) protocol with iterative model reconstruction (IMR) than the standard large focal spot (LFS) protocol at all slice levels (b). The contrast-to-noise ratio was significantly higher under the SFS protocol with IMR at all slice levels (c).
Mann-Whitney U test, respectively. A P value of less than 0.05 was considered to indicate a statistically significant difference. We used software for statistical analyses (JMP 9.0.2; SAS Institute, Cary, NC). RESULTS Quantitative Image Quality Analysis
There was no significant difference between the protocols with respect to arterial CT attenuation at all evaluated slice levels. However, under protocol 2, the mean image noise was significantly lower at all slice levels and CNR was significantly higher (Figs 3 and 4). 4
Qualitative Image Quality Analysis
The visual scores assigned to large vessels, such as the abdominal aorta and the renal and iliac arteries, were comparable for the two protocols. On the other hand, under protocol 2, the mean visual scores for small vessels in the lower extremities were significantly higher (Table 3). Representative case is shown in Figure 5. CT Radiation Dose
With protocol 1, the calculated mean CTDI vol and DLP were 9.4 ± 1.5 mGy (range: 4.8–12.6 mGy) and 1259.4 ± 253.7 mGy·cm (range: 646.7–2001.1 mGy·cm),
Academic Radiology, Vol ■, No ■, ■■ 2016
CT ANGIOGRAPHY IN PAD PATIENTS
Figure 4. A 66-year-old man with a body mass index (BMI) of 25.9 kg/m2 (a) and a 61-year-old woman with a BMI of 25.7 kg/m 2 (b) with peripheral artery disease. Transverse computed tomography images obtained at six slice levels for quantitative image quality analysis under standard large focal spot (LFS) protocol (a) and the small focal spot (SFS) protocol with iterative model reconstruction (IMR) (b) in a optimized window setting (width, 850 HU; level, 250 HU). On computed tomography angiographic images acquired with the SFS protocol with IMR, image noise was reduced compared to images obtained with the standard LFS protocol (11.4 HU and 25.1 HU, respectively, at the abdominal aorta level [level 1]).
TABLE 3. Qualitative Assessment of Image Quality
Abdominal aorta Renal artery Common-external iliac artery Internal iliac artery Superior-inferior gluteal artery Superficial femoral artery-popliteal artery Deep femoral artery Descending branch of the lateral femoral circumflex artery Descending genicular artery Tibial artery Dorsalis pedis artery
LFS Protocol (Protocol 1)
SFS Protocol (Protocol 2)
P Value
4.9 ± 0.3 4.4 ± 0.8 4.7 ± 0.5 4.2 ± 0.6 3.4 ± 0.6 4.3 ± 0.5 4.2 ± 0.6 3.7 ± 0.4 3.6 ± 0.6 3.8 ± 0.7 3.5 ± 0.7
5.0 ± 0.0 4.7 ± 0.6 4.9 ± 0.2 4.7 ± 0.5 4.2 ± 0.7 4.8 ± 0.4 4.8 ± 0.4 4.4 ± 0.7 4.5 ± 0.7 4.3 ± 0.8 4.2 ± 0.7
0.51 0.08 0.08 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01
LFS, large focal spot; SFS, small focal spot. Data are mean ± standard deviation.
5
ODA ET AL
Academic Radiology, Vol ■, No ■, ■■ 2016
Figure 5. A 75-year-old man with peripheral artery disease. He was the only patient who underwent computed tomography under both protocols during the study period. Maximum intensity images of the lower extremities under the large focal spot (LFS) (a) and the small focal spot (SFS) (b) protocols. Visualization of the small vessels, including the collateral arteries, was better under the SFS protocol.
respectively. Under protocol 2, these values were 9.0 ± 1.3 mGy (range: 5.1–12.2 mGy) and 1188.3 ± 225.2 mGy·cm (range: 680.5–1819.0 mGy·cm). There was no significant difference between the two protocols. DISCUSSION Our results demonstrate that the SFS protocol with IMR (protocol 2) yielded a significantly higher CNR of the abdominal arteries and the arteries in the lower extremities. Consequently, under protocol 2, the visual scores assigned on 3D CTA images of small vessels in the lower extremities were significantly higher. The radiation dose delivered with both protocols was equivalent. The spatial resolution of a CT system is defined as the ability to differentiate between close objects of different densities against a uniform background. In conventional CT systems, the x-ray tube has dual filaments; they provide two focal spot sizes ranging from 0.5 to 2 mm. The focal spot size is one of the main geometric factors that affects the spatial resolution of a CT system. X-ray photons emitted from a measurable focal spot are responsible for producing anatomic details with blurred edges. Thus, a smaller focal spot size produces CT images with higher spatial resolution or less blurriness. However, the SFS size is associated with the buildup of heat in the x-ray tube that may damage the tube. Consequently, the x-ray intensity and the duration of x-ray production must be lowered to obviate this problem and the SFS size is inappropriate for the routine scanning of larger human body volumes with a long z-axis. The development of multidetector CT scanners that feature better tube cooling systems resulted in shorter imaging times and superior temporal resolution, and the x-ray tube of the newer systems can handle a higher mA for SFS scanning (11). Therefore, the use of SFS for routine body CT imaging, including CTA, is now possible. 6
Oh et al. (11), who evaluated the efficacy of SFS imaging for abdominal CTA, reported that it produced images with better vessel wall clarity and fewer beam-hardening artifacts from vessel calcification. Although their findings are concordant with ours, they did not assess the small vessels in the lower extremities. We found that SFS is more effective than LFS imaging for depicting small vessels in the lower extremities. This is of practical importance because the diagnostic assessment of small vessels in the lower extremities, including collateral arteries, is important for the development of an effective therapeutic strategy. IR helps to reduce the quantum noise associated with standard convolution filtered back projection (FBP) reconstruction. It has been increasingly integrated into clinical practice (12). Earlier studies indicated that compared to FBP, the use of hybrid-type IR, such as iDose4, which is now widely applied in standard CT protocols, improves image quality, allows for a reduction in radiation exposure, and increases the percentage of diagnostically useful studies (13,14). Hybrid-type IR involves two denoising components: a sinogram restoration phase that reduces correlated noise and bias artifacts in the projection space, and an iterative denoising process in the image space that reduces the uncorrelated quantum-mottle noise. However, a certain amount of image noise persists and artifacts may be introduced because of the non-global model of noise reduction. IMR is the latest development in the field of reconstruction techniques. It is not only more complex mathematically than prior generation- and hybrid-type IR, but also more accurate, and it yields significantly better image quality (10,15). Uchida (16), who evaluated the image quality of CTA images of the abdominal viscera obtained with the SFS and the IMR techniques, reported that the former yielded higher quality CTA images. However, the study was not comparative, it included only seven subjects, and it did not evaluate the small vessels in the lower extremities. To our
Academic Radiology, Vol ■, No ■, ■■ 2016
knowledge, ours is the first comparative assessment of the image quality of CTA scans acquired with SFS and IMR in patients with PAD. Our study has some limitations. First, as it included a relatively small number of patients seen at a single center, our findings must be rigorously evaluated in large-scale prospective studies. Second, the body size of our patients was smaller than of most North American and European individuals in whom higher noise levels are to be expected. Although our observations remain to be confirmed in larger subjects, we suspect that the IMR algorithm is particularly effective in larger patients because unlike the FBP and the hybrid IR algorithms, it features nonlinear image characteristics. Third, we did not evaluate the diagnostic performance of our protocols for detecting arterial stenoses. We did not correlate ours with DSA images because for most patients we had no reference standard. Rather, we focused on comparing the quality of CTA images obtained with SFS plus IMR and standard LFS. In conclusion, qualitatively and quantitatively, CTA using the SFS and IMR techniques yielded significantly better image quality especially with respect to the small vessels in the lower extremities of PAD patients. REFERENCES 1. Dhaliwal G, Mukherjee D. Peripheral arterial disease: epidemiology, natural history, diagnosis and treatment. Int J Angiol 2007; 16:36–44. 2. Allison MA, Ho E, Denenberg JO, et al. Ethnic-specific prevalence of peripheral arterial disease in the United States. Am J Prev Med 2007; 32:328– 333.
CT ANGIOGRAPHY IN PAD PATIENTS
3. Norgren L, Hiatt WR, Dormandy JA, et al. Inter-society consensus for the management of peripheral arterial disease (TASC II). J Vasc Surg 2007; 45(suppl S):S5–S67. 4. Hirsch AT, Criqui MH, Treat-Jacobson D, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001; 286:1317–1324. 5. Pollak AW, Norton PT, Kramer CM. Multimodality imaging of lower extremity peripheral arterial disease: current role and future directions. Circ Cardiovasc Imaging 2012; 5:797–807. 6. Heijenbrok-Kal MH, Kock MC, Hunink MG. Lower extremity arterial disease: multidetector CT angiography meta-analysis. Radiology 2007; 245:433– 439. 7. Haidekker MA. Medical imaging technology. 1st ed. New York: Springer, 2013. 8. Oda S, Utsunomiya D, Funama Y, et al. A knowledge-based iterative model reconstruction algorithm: can super-low-dose cardiac CT be applicable in clinical settings? Acad Radiol 2014; 21:104–110. 9. Brown KM, Zabic S, Koehler T. Acceleration of ML iterative algorithms for CT by the use of fast start images. Proc SPIE 2012; 8313:831339. 10. Oda S, Weissman G, Vembar M, et al. Iterative model reconstruction: improved image quality of low-tube-voltage prospective ECG-gated coronary CT angiography images at 256-slice CT. Eur J Radiol 2014; 83:1408– 1415. 11. Oh LC, Lau KK, Devapalasundaram A, et al. Efficacy of “fine” focal spot imaging in CT abdominal angiography. Eur Radiol 2014; 24:3010–3016. 12. Geyer LL, Schoepf UJ, Meinel FG, et al. State of the art: iterative CT reconstruction techniques. Radiology 2015; 276:339–357. 13. Oda S, Utsunomiya D, Funama Y, et al. A hybrid iterative reconstruction algorithm that improves the image quality of low-tube-voltage coronary CT angiography. AJR Am J Roentgenol 2012; 198:1126–1131. 14. Itatani R, Oda S, Utsunomiya D, et al. Reduction in radiation and contrast medium dose via optimization of low-kilovoltage CT protocols using a hybrid iterative reconstruction algorithm at 256-slice body CT: phantom study and clinical correlation. Clin Radiol 2013; 68:e128–e135. 15. Oda S, Weissman G, Vembar M, et al. Cardiac CT for planning redo cardiac surgery: effect of knowledge-based iterative model reconstruction on image quality. Eur Radiol 2015; 25:58–64. 16. Uchida M. High-quality three-dimensional computed tomography angiography of abdominal viscera with small focal spot, low tube voltage, and iterative model reconstruction technique. O J Rad 2015; 5:8–12.
7
Original Investigation
CT Angiography in Patients with Peripheral Arterial Disease: Effect of Small Focal Spot Imaging and Iterative Model Reconstruction on the Image Quality Seitaro Oda, MD, PhD, Akira Yoshimura, MD, Keiichi Honda, RT, Yuji Iyama, MD, Kazuhiro Katahira, MD, PhD, Takeshi Nakaura, MD, PhD, Daisuke Utsunomiya, MD, PhD, Yoshinori Funama, PhD, Hideaki Yuki, MD, PhD, Masafumi Kidoh, MD, PhD, Kenichiro Hirata, MD, Narumi Taguchi, MD, Shinichi Tokuyasu, AD, Yasuyuki Yamashita, MD, PhD Rationale and Objectives: We investigated the effects of small focal spot (SFS) imaging and iterative model reconstruction (IMR) on the image quality of computed tomography angiographs (CTA) in patients with peripheral arterial disease. Materials and Methods: We divided 60 consecutive patients with suspected or confirmed peripheral artery disease into two equal groups. One group underwent large focal spot scanning under our standard CTA protocol with hybrid iterative reconstruction (iDose4) (protocol 1), and the other underwent scanning with the SFS protocol and IMR (protocol 2). Quantitative image quality parameters, ie, arterial computed tomography attenuation, image noise, and the contrast-to-noise ratio, were compared and the visual image quality (depiction of each vessel) was scored on a 5-point scale. Results: There was no significant difference in the arterial attenuation among all evaluated slice levels. The mean image noise was significantly lower under protocol 2 and the contrast-to-noise ratio was significantly higher at all slice levels. The visual scores assigned to the two protocols for the depiction of large vessels, such as the abdominal aorta and iliac artery, were comparable. However, the mean visual scores for small vessels in the lower extremities were significantly higher under protocol 2. Conclusion: CTA with SFS and IMR yielded significantly better qualitative and quantitative image quality especially for small vessels. Key Words: CT angiography; peripheral arterial disease; iterative reconstruction; small focal spot; image quality. © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.
INTRODUCTION Peripheral artery disease (PAD) is a common, chronic, progressive health problem (1). It affects up to 8.5 million (7.2%) Americans in their 40s and is associated with significant morbidity and mortality (2). The 5-, 10-, and 15-year morbidity and mortality rates from all causes in patients with PAD are approximately 30%, 50%, and 70%, respectively. Coronary artery disease is the most common cause of death among Acad Radiol 2016; ■:■■–■■ From the Department of Diagnostic Radiology, Faculty of Life Sciences, Kumamoto University, 1-1-1 Honjyo, Chuou-ku, Kumamoto 860-8556, Japan (S.O., T.N., D.U., H.Y., M.K., K.H., N.T., Y.Y.); Department of Radiology, Kumamoto Chuo Hospital (A.Y., K.H., Y.I., K.K.); Department of Medical Physics, Faculty of Life Sciences, Kumamoto University, Kumamoto, Japan (Y.F.); CT Clinical Science, Philips Electronics Japan, Tokyo, Japan (S.T.). Received March 14, 2016; revised May 25, 2016; accepted May 30, 2016. Address correspondence to: S.O. e-mail: [email protected] © 2016 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.acra.2016.05.011
patients with PAD (40%–60%); cerebral artery disease accounts for 10%–20% of deaths (3). Early diagnosis and appropriate medical intervention can mitigate limb-specific symptoms, improve the quality of life, and decrease systemic cardiovascular risks (4). Digital subtraction angiography (DSA) is considered the reference standard for diagnosing PAD. However, it is invasive and carries limitations and risks (5). Computed tomography angiography (CTA), a less invasive and safer examination, is an alternative to DSA and has gained widespread clinical acceptance for diagnosing PAD (5). Although CTA of lower extremities is more sensitive, specific, and accurate for assessing the location and extent of peripheral artery stenosis than DSA (6), its spatial resolution is inferior to DSA, and the visualization of small vessels, such as the peripheral small artery and collateral vessels, is suboptimal. The focal spot size in the x-ray tube defines the spatial resolution of a CT system (7). Many CT tubes feature small and large focal spot (SFS and LFS) sizes. The SFS size facilitates 1
Academic Radiology, Vol ■, No ■, ■■ 2016
ODA ET AL
more detailed imaging but at the cost of x-ray intensity. Most conventional CT protocols employ the LFS because SFS imaging restricts the x-ray tube power to the lower output level to prevent overheating of the tube. Advances in CT technology have surmounted this limitation because the newer detector systems make more efficient use of the available tube power. In addition, the latest x-ray generators with advanced x-ray tube cooling systems can deliver a higher tube current (mA) to the SFS; this yields an x-ray intensity similar to LFS imaging. Furthermore, the iterative reconstruction (IR) techniques allow the use of protocols with a lower tube current (8). SFS imaging benefits from this advance. Iterative model reconstruction (IMR) is the latest advance in the field of reconstruction techniques. The IMR uses a knowledge-based approach to accurately determine the data and image statistics and the system models, which depict the geometry and physical characteristics of the CT scanner, and yields improved image quality (9,10). Under the hypothesis that the combination of SFS and IMR improves small vessel visualization on CTA in patients with PAD, we investigated the effects of SFS and IMR on the image quality of CTA scans. MATERIALS AND METHODS We obtained institutional review board (IRB) approval and prior written informed consent from all patients participating in this prospective study (Hospital IRB Number: #29-03). Study Population
Between December 2014 and January 2016, we enrolled 60 consecutive patients (41 men and 19 women; mean age: 73.4 years) with suspected or confirmed PAD. All underwent CTA. The inclusion criteria were no lower limb amputation, no renal failure (estimated glomerular filtration rate <30 mL/min/1.73 m2), no hemodialysis, and no history of allergic reactions to iodinated contrast material. The enrolled patients were randomized and scanned under one of two CTA protocols based on a random-number table. One group (n = 30) underwent scanning with our standard CTA protocol with LFS (protocol 1) and the other group (n = 30) underwent scanning under the SFS protocol (protocol 2). Patient characteristics are summarized in Table 1. CT Scanning and Contrast Infusion Protocols
All CT examinations were performed on a 256-slice CT system (Brilliance iCT; Philips Healthcare, Cleveland, OH). The parameters were detector configuration, 128 × 0.625 mm; slice thickness, 1.0 mm; section interval, 0.5 mm; gantry rotation time, 0.75 seconds; beam pitch, 0.59; tube voltage, 100 kVp; and reference tube current time product, 231 mAs (effective mAs) with auto-modulation (Dose Right; Philips Healthcare). CTA data were acquired in the craniocaudal direction 2
TABLE 1. Patient Demographics LFS Protocol SFS Protocol (n = 30) (n = 30) P (Protocol 1) (Protocol 2) Value Sex (male/female) 21/9 Age (y) 72.8 ± 8.6 Body height (cm) 157.6 ± 7.9 Body weight (kg) 61.2 ± 13.0 24.5 ± 4.3 Body mass index (kg/m2) 33.1 ± 30.4 eGFR (mL/min/1.73 m2)
20/10 74.1 ± 9.3 158.9 ± 8.3 59.7 ± 15.9 23.4 ± 4.6 32.5 ± 30.8
0.72 0.55 0.53 0.70 0.34 0.94
eGFR, estimated glomerular filtration rate; LFS, large focal spot; SFS, small focal spot. Note: Data are mean ± standard deviation.
TABLE 2. Imaging and Contrast Material Parameters of the LFS and the SFS Protocols LFS Protocol (Protocol 1) CT scanner Collimation Tube voltage Effective tube current Rotation time Helical pitch Total amount of contrast medium Injection duration Bolus tracking trigger Scan delay Image reconstruction Section thickness/interval
SFS Protocol (Protocol 2)
256-slice CT (Brilliance iCT, Philips Healthcare) 128 × 0.625 mm 100 kVp 231 eff. mAs (reference) with auto-modulation 0.75 s/rot 0.585 500 mgI/mL 25 s 150 HU (abdominal aorta) 15 s IMR iDose4 1.0/0.5 mm
CT, computed tomography; IMR, iterative model reconstruction; LFS, large focal spot; SFS, small focal spot.
from the suprarenal aorta to the ankles. Using a doublehead power injector (Auto Enhance A-250; Nemoto Kyorindo, Tokyo, Japan), iopamidol (iodine concentration 300 or 370 mgI/mL [Iopamiron 300 or 370; Bayer HealthCare, Osaka, Japan]) was injected via a 20-gauge catheter inserted into an antecubital vein. The amount of contrast material was adjusted to the body weight of each patient (500 mgI/kg) and injected at a fixed injection duration of 25 seconds. Contrast administration was followed by the injection of 40 mL of a saline solution delivered at the same injection rate as the contrast medium. An automatic bolus-tracking program was used to time the start of scanning for each phase after contrast material injection. Monitoring was at the L1 vertebral body level; a region of interest (ROI) cursor (1.0–2.0 cm2) was placed on the abdominal aorta. Data acquisition started 15 seconds after triggering. The acquisition parameters for protocols 1 and 2 are summarized in Table 2.
Academic Radiology, Vol ■, No ■, ■■ 2016
CT Image Reconstruction
We reconstructed the raw data over a field of view of 35.0– 40.0 cm; the pixel matrix was 512 × 512, and the section thickness and intervals were 1.0 and 0.5 mm, respectively. Images acquired under protocol 1 were reconstructed with hybrid IR (iDose4, level 4). Protocol 2 images were reconstructed with IMR (body routine, level 2). Original 1.0mm axial images were processed on a commercially available image-processing workstation (Virtual Place Advance Plus; AZE, Tokyo, Japan) for three-dimensional (3D) reconstruction by a radiology technologist with 15 years of experience (K.H.). The 3D reconstruction of a single case on the imageprocessing workstation took 15–20 minutes. Quantitative Image Quality Analysis
A board-certified radiologist (A.Y., with 9 years of experience reading general CT scans), blinded to the protocols used, performed measurements on axial source images. Arterial CT attenuation was measured at six levels from the abdominal aorta to the artery of the lower leg (Fig 1). For portions of levels
CT ANGIOGRAPHY IN PAD PATIENTS
2–6, the CT number of the bilateral arteries was recorded to calculate the mean CT number. When the artery on one side was occluded, only the non-occluded contralateral vessel was measured. Attempts were made to select ROIs of approximately 100, 40, 30, 20, 15, and 5 mm2 for the abdominal aorta, the common iliac, external iliac, superficial femoral, popliteal, and anterior or posterior tibial artery, respectively. Their sizes were chosen to obtain ROIs large enough not to be affected by pixel variability and small enough to avoid including the arterial wall and its calcification. The image noise, ie, the average of the standard deviation of muscle attenuation at the same slice level, was recorded. The contrast-to-noise ratio (CNR) of each artery was calculated using the equation
CNR = (HU A − HU M ) image noise , where HUA and HUM are the CT attenuation in the artery and muscle, respectively, and HUM is the CT attenuation in the muscle. These parameters were compared in protocols 1 and 2. Qualitative Image Quality Analysis
All images were reviewed and interpreted on PACS workstations (EV Insite, PSP Corp., Tokyo, Japan). Two boardcertified radiologists with 8 and 11 years of experience interpreting CTA scans were blinded to the protocols for their consensual visual evaluation of the maximum intensity projection 3D CTA images. The images were divided into the aortoiliac, femoropopliteal, and lower leg section. For qualitative analysis, we used 11 vessels of various sizes in the various locations (Fig 2). The visualization of each vessel, ie, the diagnostic quality of the images obtained with the two protocols, was graded on a 5-point scale, where 5 = excellent (excellent visualization of the entire portion of the artery), 4 = good (sufficient visualization of almost the entire portion of the artery), 3 = adequate (partially ambiguous but yielding sufficient diagnostic information), 2 = fair (insufficient visualization of some portion of the artery with partially limited diagnostic information), and 1 = uninterpretable (insufficient visualization of the entire portion of the artery). CT Radiation Dose
Based on each patient’s dose information page, we recorded the volume CT dose index (CTDIvol) in milligray (mGy) and the dose-length product (DLP) in mGy·cm for the CTA phase. Statistical Analysis Figure 1. Measurement sites for quantitative image quality analysis. Level 1: Abdominal aorta at the third lumbar vertebra level. Level 2: Common iliac artery at the iliac crest level. Level 3: External iliac artery at the hip joint level. Level 4: Superficial femoral artery at the center of the femoral bone. Level 5: Popliteal artery at the level of the patella. Level 6: Center of the anterior or posterior tibial artery.
Numerical data were expressed as mean ± standard deviation. All qualitative and quantitative image parameters of the two protocols were compared. CTDIvol and DLP were also assessed. Differences in the mean values between the two protocols with normally and non-normally distributed data were determined with the two-tailed independent t test and the 3
ODA ET AL
Academic Radiology, Vol ■, No ■, ■■ 2016
Figure 2. Evaluation sites for qualitative image quality analysis. The assessed vessels were the abdominal aorta (a), renal (b), common-external iliac (c), internal iliac (d), superior-inferior gluteal (e), superficial femoral-popliteal (f), deep femoral (g), descending branch of the lateral femoral circumflex (h), descending genicular (i), and the center of the anterior tibial (j) arteries. Measurements were also obtained at the lateral malleolus of the dorsalis pedis artery (k).
Figure 3. There was no significant difference between the two protocols with respect to arterial computed tomography attenuation at all evaluated slice levels (a). The mean image noise was significantly lower under the small focal spot (SFS) protocol with iterative model reconstruction (IMR) than the standard large focal spot (LFS) protocol at all slice levels (b). The contrast-to-noise ratio was significantly higher under the SFS protocol with IMR at all slice levels (c).
Mann-Whitney U test, respectively. A P value of less than 0.05 was considered to indicate a statistically significant difference. We used software for statistical analyses (JMP 9.0.2; SAS Institute, Cary, NC). RESULTS Quantitative Image Quality Analysis
There was no significant difference between the protocols with respect to arterial CT attenuation at all evaluated slice levels. However, under protocol 2, the mean image noise was significantly lower at all slice levels and CNR was significantly higher (Figs 3 and 4). 4
Qualitative Image Quality Analysis
The visual scores assigned to large vessels, such as the abdominal aorta and the renal and iliac arteries, were comparable for the two protocols. On the other hand, under protocol 2, the mean visual scores for small vessels in the lower extremities were significantly higher (Table 3). Representative case is shown in Figure 5. CT Radiation Dose
With protocol 1, the calculated mean CTDI vol and DLP were 9.4 ± 1.5 mGy (range: 4.8–12.6 mGy) and 1259.4 ± 253.7 mGy·cm (range: 646.7–2001.1 mGy·cm),
Academic Radiology, Vol ■, No ■, ■■ 2016
CT ANGIOGRAPHY IN PAD PATIENTS
Figure 4. A 66-year-old man with a body mass index (BMI) of 25.9 kg/m2 (a) and a 61-year-old woman with a BMI of 25.7 kg/m 2 (b) with peripheral artery disease. Transverse computed tomography images obtained at six slice levels for quantitative image quality analysis under standard large focal spot (LFS) protocol (a) and the small focal spot (SFS) protocol with iterative model reconstruction (IMR) (b) in a optimized window setting (width, 850 HU; level, 250 HU). On computed tomography angiographic images acquired with the SFS protocol with IMR, image noise was reduced compared to images obtained with the standard LFS protocol (11.4 HU and 25.1 HU, respectively, at the abdominal aorta level [level 1]).
TABLE 3. Qualitative Assessment of Image Quality
Abdominal aorta Renal artery Common-external iliac artery Internal iliac artery Superior-inferior gluteal artery Superficial femoral artery-popliteal artery Deep femoral artery Descending branch of the lateral femoral circumflex artery Descending genicular artery Tibial artery Dorsalis pedis artery
LFS Protocol (Protocol 1)
SFS Protocol (Protocol 2)
P Value
4.9 ± 0.3 4.4 ± 0.8 4.7 ± 0.5 4.2 ± 0.6 3.4 ± 0.6 4.3 ± 0.5 4.2 ± 0.6 3.7 ± 0.4 3.6 ± 0.6 3.8 ± 0.7 3.5 ± 0.7
5.0 ± 0.0 4.7 ± 0.6 4.9 ± 0.2 4.7 ± 0.5 4.2 ± 0.7 4.8 ± 0.4 4.8 ± 0.4 4.4 ± 0.7 4.5 ± 0.7 4.3 ± 0.8 4.2 ± 0.7
0.51 0.08 0.08 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01
LFS, large focal spot; SFS, small focal spot. Data are mean ± standard deviation.
5
ODA ET AL
Academic Radiology, Vol ■, No ■, ■■ 2016
Figure 5. A 75-year-old man with peripheral artery disease. He was the only patient who underwent computed tomography under both protocols during the study period. Maximum intensity images of the lower extremities under the large focal spot (LFS) (a) and the small focal spot (SFS) (b) protocols. Visualization of the small vessels, including the collateral arteries, was better under the SFS protocol.
respectively. Under protocol 2, these values were 9.0 ± 1.3 mGy (range: 5.1–12.2 mGy) and 1188.3 ± 225.2 mGy·cm (range: 680.5–1819.0 mGy·cm). There was no significant difference between the two protocols. DISCUSSION Our results demonstrate that the SFS protocol with IMR (protocol 2) yielded a significantly higher CNR of the abdominal arteries and the arteries in the lower extremities. Consequently, under protocol 2, the visual scores assigned on 3D CTA images of small vessels in the lower extremities were significantly higher. The radiation dose delivered with both protocols was equivalent. The spatial resolution of a CT system is defined as the ability to differentiate between close objects of different densities against a uniform background. In conventional CT systems, the x-ray tube has dual filaments; they provide two focal spot sizes ranging from 0.5 to 2 mm. The focal spot size is one of the main geometric factors that affects the spatial resolution of a CT system. X-ray photons emitted from a measurable focal spot are responsible for producing anatomic details with blurred edges. Thus, a smaller focal spot size produces CT images with higher spatial resolution or less blurriness. However, the SFS size is associated with the buildup of heat in the x-ray tube that may damage the tube. Consequently, the x-ray intensity and the duration of x-ray production must be lowered to obviate this problem and the SFS size is inappropriate for the routine scanning of larger human body volumes with a long z-axis. The development of multidetector CT scanners that feature better tube cooling systems resulted in shorter imaging times and superior temporal resolution, and the x-ray tube of the newer systems can handle a higher mA for SFS scanning (11). Therefore, the use of SFS for routine body CT imaging, including CTA, is now possible. 6
Oh et al. (11), who evaluated the efficacy of SFS imaging for abdominal CTA, reported that it produced images with better vessel wall clarity and fewer beam-hardening artifacts from vessel calcification. Although their findings are concordant with ours, they did not assess the small vessels in the lower extremities. We found that SFS is more effective than LFS imaging for depicting small vessels in the lower extremities. This is of practical importance because the diagnostic assessment of small vessels in the lower extremities, including collateral arteries, is important for the development of an effective therapeutic strategy. IR helps to reduce the quantum noise associated with standard convolution filtered back projection (FBP) reconstruction. It has been increasingly integrated into clinical practice (12). Earlier studies indicated that compared to FBP, the use of hybrid-type IR, such as iDose4, which is now widely applied in standard CT protocols, improves image quality, allows for a reduction in radiation exposure, and increases the percentage of diagnostically useful studies (13,14). Hybrid-type IR involves two denoising components: a sinogram restoration phase that reduces correlated noise and bias artifacts in the projection space, and an iterative denoising process in the image space that reduces the uncorrelated quantum-mottle noise. However, a certain amount of image noise persists and artifacts may be introduced because of the non-global model of noise reduction. IMR is the latest development in the field of reconstruction techniques. It is not only more complex mathematically than prior generation- and hybrid-type IR, but also more accurate, and it yields significantly better image quality (10,15). Uchida (16), who evaluated the image quality of CTA images of the abdominal viscera obtained with the SFS and the IMR techniques, reported that the former yielded higher quality CTA images. However, the study was not comparative, it included only seven subjects, and it did not evaluate the small vessels in the lower extremities. To our
Academic Radiology, Vol ■, No ■, ■■ 2016
knowledge, ours is the first comparative assessment of the image quality of CTA scans acquired with SFS and IMR in patients with PAD. Our study has some limitations. First, as it included a relatively small number of patients seen at a single center, our findings must be rigorously evaluated in large-scale prospective studies. Second, the body size of our patients was smaller than of most North American and European individuals in whom higher noise levels are to be expected. Although our observations remain to be confirmed in larger subjects, we suspect that the IMR algorithm is particularly effective in larger patients because unlike the FBP and the hybrid IR algorithms, it features nonlinear image characteristics. Third, we did not evaluate the diagnostic performance of our protocols for detecting arterial stenoses. We did not correlate ours with DSA images because for most patients we had no reference standard. Rather, we focused on comparing the quality of CTA images obtained with SFS plus IMR and standard LFS. In conclusion, qualitatively and quantitatively, CTA using the SFS and IMR techniques yielded significantly better image quality especially with respect to the small vessels in the lower extremities of PAD patients. REFERENCES 1. Dhaliwal G, Mukherjee D. Peripheral arterial disease: epidemiology, natural history, diagnosis and treatment. Int J Angiol 2007; 16:36–44. 2. Allison MA, Ho E, Denenberg JO, et al. Ethnic-specific prevalence of peripheral arterial disease in the United States. Am J Prev Med 2007; 32:328– 333.
CT ANGIOGRAPHY IN PAD PATIENTS
3. Norgren L, Hiatt WR, Dormandy JA, et al. Inter-society consensus for the management of peripheral arterial disease (TASC II). J Vasc Surg 2007; 45(suppl S):S5–S67. 4. Hirsch AT, Criqui MH, Treat-Jacobson D, et al. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001; 286:1317–1324. 5. Pollak AW, Norton PT, Kramer CM. Multimodality imaging of lower extremity peripheral arterial disease: current role and future directions. Circ Cardiovasc Imaging 2012; 5:797–807. 6. Heijenbrok-Kal MH, Kock MC, Hunink MG. Lower extremity arterial disease: multidetector CT angiography meta-analysis. Radiology 2007; 245:433– 439. 7. Haidekker MA. Medical imaging technology. 1st ed. New York: Springer, 2013. 8. Oda S, Utsunomiya D, Funama Y, et al. A knowledge-based iterative model reconstruction algorithm: can super-low-dose cardiac CT be applicable in clinical settings? Acad Radiol 2014; 21:104–110. 9. Brown KM, Zabic S, Koehler T. Acceleration of ML iterative algorithms for CT by the use of fast start images. Proc SPIE 2012; 8313:831339. 10. Oda S, Weissman G, Vembar M, et al. Iterative model reconstruction: improved image quality of low-tube-voltage prospective ECG-gated coronary CT angiography images at 256-slice CT. Eur J Radiol 2014; 83:1408– 1415. 11. Oh LC, Lau KK, Devapalasundaram A, et al. Efficacy of “fine” focal spot imaging in CT abdominal angiography. Eur Radiol 2014; 24:3010–3016. 12. Geyer LL, Schoepf UJ, Meinel FG, et al. State of the art: iterative CT reconstruction techniques. Radiology 2015; 276:339–357. 13. Oda S, Utsunomiya D, Funama Y, et al. A hybrid iterative reconstruction algorithm that improves the image quality of low-tube-voltage coronary CT angiography. AJR Am J Roentgenol 2012; 198:1126–1131. 14. Itatani R, Oda S, Utsunomiya D, et al. Reduction in radiation and contrast medium dose via optimization of low-kilovoltage CT protocols using a hybrid iterative reconstruction algorithm at 256-slice body CT: phantom study and clinical correlation. Clin Radiol 2013; 68:e128–e135. 15. Oda S, Weissman G, Vembar M, et al. Cardiac CT for planning redo cardiac surgery: effect of knowledge-based iterative model reconstruction on image quality. Eur Radiol 2015; 25:58–64. 16. Uchida M. High-quality three-dimensional computed tomography angiography of abdominal viscera with small focal spot, low tube voltage, and iterative model reconstruction technique. O J Rad 2015; 5:8–12.
7