Abstract
BACKGROUND:
Computed tomography (CT) pulmonary angiography (CTPA) examination has been frequently applied in detecting suspected pulmonary embolism (PE). How to reduce radiation dose to patients is also of concern.
OBJECTIVE:
To assess the value of using 640-slice CT wide-detector volume scan with adaptive statistical iterative reconstruction (ASIR) algorithm in low-dose CTPA.
METHODS:
Fifty-eight patients who performed with CTPA were divided into two groups randomly. In the first experimental group (n = 30), ASIR combined with volume scan were performed on the patients, while in the second conventional group (n = 28), patients received ASIR combined with conventional spiral scan. General data including age and body mass index, image quality, pulmonary arterial phase, and radiation dose were analyzed by t test in the two groups.
RESULTS:
In both groups, all images revealed the 5-order or higher pulmonary arterial branches and fully met the needs for clinical diagnosis. There was no statistical difference in general data between the two groups. In terms of pulmonary phase accuracy, compared with the conventional group, images at pulmonary arterial phase could be captured more accurately in the experimental group. CTDI in the experimental group decreased by 30% compared with that in the conventional group. The actual radiation dose in the experimental group was 1.5 mSv, which is reduced by 53% compared to that in the conventional group.
CONCLUSIONS:
Compared with the conventional spiral scan, using 640-slice CT volume scan with ASIR in CTPA is more accurate in scanning phase and has lower radiation dose. There is no significant difference in image quality between the two groups.
Introduction
Since the introduction of multi-slice computed tomography (CT) with high spatial and temporal resolution, CT pulmonary angiography (CTPA) examination has become the method of choice for imaging the pulmonary vasculature when pulmonary embolism (PE) is suspected currently [1–5]. However, because of the well-documented high radiation exposure associated with CT, there are serious concerns regarding the potential hazards of the resulting radiation dose [4–7]. How to reduce the radiation dose in patients is also of concern. Studies of reducing CTPA doses are rising, such as the automatic tube current or low voltage technique [2, 7], and the minimum radiation dose that provides diagnostic-quality studies is recommended. In this light, the use of all available equipment-specific dose reduction techniques is strongly endorsed [5].
At present, spiral CT with adaptive statistical iterative reconstruction (ASIR) technique has been mostly researched and applied in daily practice [6, 8]. The application of ASIR can reduce the CT radiation dose while ensuring the CTPA image quality. On the other hand, with the update of CT equipment, the current multi-slice CT also allows a fast tube rotation. If combined with a wide detector, i.e., 16 cm, the protocol can cover the main pulmonary arteries once time without table feed, thus greatly shortens the scanning time. The reduction of the scanning time also means that the radiation dose is further reduced under the same scanning parameters. Given the ALARA (“as low as reasonably achievable”) principle in CT examination [1, 8], wide-detector volume scan with ASIR technique may be beneficial to children, young people, or women in suspected PE clinically [4]. In this study, we aimed at investigating the difference in image quality of CTPA and radiation dose between Toshiba 640-slice CT wide-detector volume scan combined with ASIR and conventional spiral scan.
Materials and methods
Patients
From October 1, 2017 to January 31, 2018, we performed CTPA examinations in 58 patients (24 males and 34 females) suspected of having PE clinically. The age range was from 37 to 89 years. By a double-blinding, they were divided into two groups, experimental group (n = 28; ASIR combined with wide-detector volume scan) and conventional group (n = 30; conventional spiral scan). The study was approved by the Ethics Committee of our Hospital. All patients who enrolled in the study signed informed consent.
Inspection equipment
Toshiba 640-slice spiral CT (Aquilion ONE) in 16 cm detector width was used. Specifically, two scanning models could be selected: one was spiral scanning model with the adjustable collimation width. The other was volume scanning model with the collimation width of 16 cm in once scan or w-volume model, but the later had the maximal coverage of a single volume scan was 10.2 cm. Thus, we did not adopt it in this study.
Inspection method
In this study, 35 mL contrast agent iopromide (370 mgI/mL) was injected via the elbow median vein at a flow rate of 4 mL/s using a high pressure double syringe in both the conventional and experimental groups. During scanning, put patient’s foot first, and lift hands over the head. The bolus tracking technique was used for the enhanced scan. A region of interest (ROI) was placed on the main pulmonary arteries or the left pulmonary artery to avoid the influence of the contrast agent on the superior vena cava and the range was slightly smaller than that of the blood vessel section. For this model, the enhancing peak values could be monitored in real time and the pulmonary arterial phase was automatically triggered with a certain delay once the enhancing peak values reached a preset Hounsfield units (HU) threshold.
In the conventional group, spiral scan was used with scanning triggered after inhalation and breath hold. The trigger threshold was 80 HU with 4 s delay for scanning. The scanning length was fixed to 24 cm, and scanning parameters included 120 kV tube voltage, automatic tube current, and collimation width 64×0.5 mm. The rotation time was 0.5 s, the pitch factor was 0.887, and the field of vision (FOV) was 28 cm. The image reconstruction was performed using an ASIR with a slice thickness of 0.5 mm and an intersection of 0.5 mm. In the experimental group, a wide-detector volume scan with the fixed scanning length of 16 cm was used. The scan was triggered after the patient’s breathhold at the end of expiration to achieve the goal of including as main pulmonary arteries as possible. The trigger threshold of the experimental group was 160 HU with 1.8 s delay for scanning. Parameters included 120 kV, automatic tube current, collimation width of 320×0.5 mm, the rotation time was 0.5 s, FOV was 28 cm, and image reconstruction was also performed using the ASIR with a slice thickness of 0.5 mm and intersection of 0.5 mm.
Image processing and analysis
The scanned raw data was sent to the picture archiving and communication system (PACS, Synapse, Fuji, Japan) workstation. Two radiologists (F.Q.L., and Q.K. with 18 and 10 years of experience in body CT, respectively) carried out post-processing and analysis of these images respectively by a double-blind method. Image interpretations used the axial, multiplanar reconstruction (MPR), including curved planar reconstruction (CPR) images as well as maximal intensity projection (MIP) and virtual rendering (VR) images. Accurate identification of the pulmonary artery trunk, right and left main pulmonary arteries, interlobar arteries, segmental, and subsegmental arteries were in meticulous analysis of both mediastinal and lung window settings to identify the arteries according to their relationship to the bronchi [9]. The presence of vascular signs of PE based on the reported literature [10]. They scored and recorded data required by objective and subjective evaluations, and the mean value of two sets of data was adopted.
Subjective ratings
Pulmonary artery image quality evaluation
The pulmonary artery image quality is assessed using a 5-grade scale, and both pulmonary arteries are analyzed. To be specific, poor image quality and inability to observe blood vessels owing to poor contrast enhancement or patient motion artifacts (grade 1); poor image quality affecting image outcomes (grade 2); image quality not bad, but not affecting image outcomes (grade 3); good image quality but few artifacts (grade 4); good image quality with clear blood vessels (grade 5).
Pulmonary phase accuracy evaluation
A 5-grade scale was also used in pulmonary phase accuracy evaluation: no pulmonary artery was filled with contrast medium (grade 1); pulmonary artery attenuation lower than that of pulmonary vein or aorta (grade 2); similar attenuation of pulmonary artery and vein, and pulmonary arteries needed to be identified carefully (grade 3); higher attenuation of the pulmonary artery than that of the pulmonary vein, and the VR image reconstruction being slightly affected (grade 4); high attenuation of the pulmonary artery, no opacification or slight opacification of the pulmonary veins, and the pulmonary artery VR image reconstruction being not affected at all (grade 5).
Objective evaluation
Measurement of CT attenuated values and calculations based on the reported literatures [3, 11]: the main pulmonary artery and the lobar pulmonary artery of the right and left lower lobe were measured. The ROIs was equal to the diameter of the vessel. When there is a thrombus, place the ROI away from it. The mean value of the measured CT values was calculated as the mean pulmonary artery value (ROImpa). At the pulmonary vein level, a ROI with a diameter of approximately 1 cm was taken in the presternal air area, in which the background noise was measured on the assessment of standard deviation (SD) of attenuation under the lung window settings to avoid clothing interference. The CT values of the bilateral spine muscles were measured separately and the mean value was selected as ROImuscle. Two reviewers independently measured and obtained the mean values. The signal to noise ratio (SNR) and contrast to noise ratio (CNR) of the pulmonary artery were calculated, where: SNR = ROImpa/SD, CNR = (ROImpa - ROImuscle) /SD.
Radiation dose parameters and calculations
CT dose index volume (CTDIvol) and dose length product (DLP) were calculated. DLP (mGy · cm) = CTDIvol (mGy)×scan length (L, cm). The actual radiation dose (E) = 0.017×DLP [3, 5], that is E = 0.017×CTDI×L. In the case of a fixed scan length, E is positively correlated with CTDI. In this article, the length of the experimental group was 16 cm, while in the conventional group, the length was 24 cm. In order to compare the DLP in both groups under the same scanning length, we virtually converted the scanning length of 24 cm into 16 cm in the conventional group, that is, 24 cm×2/3.
Data analysis
A statistical data analysis software (SPSS 19.0) was used and measurement data was expressed as
Results
All the patients in both groups completed the examination successfully. In the experimental group, the scanning acquisition time was 0.5 s, and all images showed the 5-order pulmonary arterial branch or higher that could meet the need for clinical diagnosis of PE (Figs. 1–3). In the conventional group, the scanning acquisition time was 4.2 s. Among them, 28 patients had full chest coverage from the lung apex to the diaphragm, and only two patients had incomplete coverage away from the apex and the diaphragmatic dome of 1–2 cm. Although incomplete coverage may cut out the peripheral smaller arteries, more than 5-order arterial branches could be depicted, and it did not affect to assess the subsegmental arteries in those patients. There was no statistical difference in general data such as age or body mass index between the two groups (Table 1).
A 71-year-old woman with deep venous thrombosis shows a longer lung field due to emphysema. Volume scan with 16-cm detector shows high attenuation of the pulmonary arteries from the right and left main pulmonary arteries to 5-order branches clearly, but the “roof effect” also can be seen (white arrows).
A 54-year-old man with pulmonary hypertension presents a shorter lung field, in which almost all of pulmonary arterial branches are covered when volume scan with 16-cm detector.
A 51-year-old woman who had a acute chest pain and deep venous thrombosis as well. CTPA reveals filling defects in the 3-order pulmonary arterial branches on both the axial image (a) and coronary reformatted image (b) (white arrow). On the curved planar reconstruction (CPR) image (c), the left main pulmonary artery down to the 6-order branches (white arrow) along with filling defects in the 3-order branches (black arrow) are depicted clearly using volume scan with 16-cm detector.
Comparison of patient’s general data and CTDIvol
Note: # indicates DLP virtual value at the length of 16 cm in the CG, that is, DLP = 24 cm×2/3×CTDIvol (mGy), which resulting in a virtual DLP and E were 126.73±36.86 (mGy·cm), 2.15±0.63 (mSv), respectively. EG: experimental group; CG: conventional group.
With regard to the subjective score, there was no significant difference in images between the two groups (P > 0.05) (Table 2), indicating both of them could meet the diagnostic requirements. In terms of pulmonary phase accuracy, compared with the conventional group, images at pulmonary arterial phase could be captured more accurately in the experimental group (t = 4.965; P < 0.05). As for objective score, there was no statistical difference in two sets of data between the two groups (Table 3).
Comparison of subjective scores of images
Comparison of subjective scores of images
EG: experimental group; CG: conventional group.
Comparison of objective scores of images
EG: experimental group; CG: conventional group; mpa: mean pulmonary artery.
The radiation dose was showed statistical significance after comparison between the experimental group and the conventional group (P < 0.05) (Table 1). Compared with the conventional group, the CTDIvol of the experimental group decreased by 30% and DLP would decreased by 30% if the same scan length was taken in the conventional group. The actual mean radiation dose was 1.5 mSv by the calculation of 0.017×CTDIvol×L in the experimental group, which decreased by 53% compared with the actual radiation dose being 3.2 mSv (24 cm) in the conventional group.
Discussion
With the increase of patients suffering from PE in clinic, due to application of D-dimer in high-risk populations and recognition of the contribution of CT in the diagnosis of PE [4, 12], CTPA is increasingly being used clinically for the diagnosis of PE [1–7]. Then how to accomplish an effective CTPA examination at a lower dose are our concerns. Previous studies in CTPA most focused on ASIR combined with automatic tube current to decrease the radiation dose with estimated radiation savings of 25–60% [6, 13]. The potential of ASIR technique allows reduction in radiation dose while image quality is maintained [7, 14]. Indeed, dose reduction methods still include tube current modulation, tube voltage selection, or shorter scanning acquisition time (i.e., high-pitch scanning models or scanning range adjustment) [7]. However, simply lowering tube current or voltage may increase image noise, particularly in those obese patients [7, 14]. Therefore, combination of using available dose-reducing techniques is strongly endorsed [4].
Nevertheless, in the previous study of low-dose CTPA, the actual radiation dose was more than 2 mSv [15, 16], which was similar to our conventional spiral scan protocol. When volume scan model with 16 cm wide-detector CT used, the actual radiation dose in this study was only 1.5 mSv reduced by more than 25% compared with other studies [15, 16]. The difference in this study was that Toshiba 640-slice CT finished the scanning length within 0.5 s for 16 cm pulmonary artery in the experimental group. While, the scanning acquisition time was 4.2 s for 24 cm scanning length in the conventional group. In fact, the acquisition time also need 3.9 s when covering the 16-cm scanning length with this spiral scan model, which was much longer than the volume scan one. Considering the dose reduction strategies, both appropriately reducing scanning range and effectively shortening scanning time may be regarded as a feasible way.
As for the pulmonary image quality, the trigger threshold of wide-detector volume scan was increased to close to the ideal pulmonary artery CT value without table feed to better catch the timing of the pulmonary artery and obtain a more satisfactory image in contrast to the spiral scan model. In this study, the proportion of pulmonary artery in the experimental group in grade 5 was 80%, and the rest ones were grade 4, while in the conventional group, most of them were in grade 4.
The most common reasons for nondiagnostic CT images are poor contrast enhancement of pulmonary vessels, patient motion, and increased image noise due to excessive patient obesity [17]. Thus, CTPA using wide-detector volume scan combined with ASIR technique may effectively shorten scanning time, almost eliminate motion artifacts, as well as decrease radiation dose. These advantages may favor those acute chest pain patients, children, and young women in suspected PE in current clinical practice [4].
The artifacts of images generated by wide-detector volume scan are “all or nothing”, that is, if there are artifacts, all images will have, while in the spiral scan, artifacts are generally existent in some images. In this study, for a patient with a short breath hold but in active cooperation, a wide-detector volume scan can finish the examination more easily. While for this type of patient, artifacts are present in the inferior lung by spiral scan. Therefore, it is also advised that patients with short breath hold take the wide-detector volume scan for a satisfactory image. For those who are unable to cooperate, fewer artifacts are presented in the wide-detector volume scan for its short scanning time compared to the conventional scan.
Honestly, the wide-detector volume scan also has its own shortcomings, such as “roof effect” of the image which is imperfect, limited scanning length, and unable to fully cover the chest, but these drawbacks may be offset partially considering its lower radiation dose and more accurately to catch the pulmonary arterial phase than that of conventional spiral scan at the present study. There is also a w-volume scanning mode in wide-detector volume scan that automatically removes the roof effect, which is equivalent to multiple volume scans, but the maximum scan length for a single volume scan in this mode is 10.2 cm, and 23.2 cm is the maximum scanning length for volume scan twice. Inhomogeneous opacification of contrast agent will also appear in pulmonary veins and aorta, which is caused by time difference between two volume scans. In addition, the pulmonary vessel joining is poor due to malposition from inconsistent two breath-holds. The most important issue is that radiation dose is not substantial decrease compared with the conventional spiral scan. Thus, this scanning mode was not adopted in the study.
In summary, this study demonstrated that using Toshiba 640-slice CT wide-detector volume imaging scan can reduce the radiation dose of CTPA and obtain satisfactory images compared to the conventional spiral scan.