CT chest and gantry rotation time: does the rotation time influence image quality?

Introduction

Computed tomography (CT) imaging of the chest is one of the most frequent radiologic examinations in the world, as it is a non-invasive way in which to evaluate a variety of clinical issues, such as pulmonary infiltration or potential malignancies. Even though multiple techniques are designed to reduce radiation exposure, chest CT imaging still exposes patients to a higher amount of radiation than other imaging modalities (e.g. X-ray). However, radiation exposure in chest CT imaging has decreased remarkably in recent years using low kV imaging, iterative reconstruction algorithms, and campaigns by radiological societies around the world, such as the “Image Wisely” campaign from the Radiological Society of North America and American College of Radiology (1–3).

State-of-the-art CT scanners implement rapid imaging strategies to reduce radiation exposure, including wide detectors, fast gantry rotation times, two detectors, or a combination of these strategies (3–5). In nearly every CT machine, fast tube detector rotation times of ≤0.5 s / rotation (rot) are possible and can be used in different clinical situations. Fast gantry rotation speeds combined with an adapted pitch-factor shorten image acquisition times, which is advantageous when imaging the thoracic region because it reduces the number of motion artifacts. Unfortunately, accelerating the gantry rotation time can also introduce additional artifacts (e.g. streaking) that are not visible using slower speeds. Therefore, the aim of this study was to investigate the influence of gantry rotation time on image quality, noise, artifacts, and radiation exposure during chest CT examinations.

Material and Methods

Patients and CT protocols

This single-center, observer-blinded, retrospective study was approved by the local Ethics Committee and written informed consent was obtained by all patients. Data of consecutive patients that underwent clinically indicated CT of the chest between November 2011 and April 2013 were analyzed. All CT examinations included in this study were diagnostic and did not require repeating because of unsatisfactory image quality or artifacts.

Patients were divided into four groups of 40 individuals each based on gantry rotation time (Table 1). All groups underwent chest CT scanning on a dual-source CT operated in single-source mode functioning as a 128 slice CT (Somatom Definition Flash, Siemens Healthcare, Forchheim, Germany) with a 1.2 pitch, 128 × 0.6 mm collimation, 120 kV tube potential, and 110 reference mAs. Gantry rotation times for each group were as follows: group 1, 1 s/rot; group 2, 0.5 s/rot; group 3, 0.33 s/rot; and group 4, 0.28 s/rot. Automatic exposure control was used for all scans (CARE Dose 4D, Siemens Healthcare, Forchheim, Germany). Data were acquired in the caudocranial direction during deep inspiratory breath-hold.

OPEN IN VIEWER

Table 1.

Examination parameters.

	Group 1	Group 2	Group 3	Group 4
Slice × collimation	128 × 0.6	128 × 0.6	128 × 0.6	128 × 0.6
Pitch	1.2	1.2	1.2	1.2
kV / ref. mAs	120 / 110	120 / 110	120 / 110	120 / 110
Rotation time (seconds/rotation)	1.0	0.5	0.33	0.28
Patients	40	40	40	40
Male	24	23	20	27
Female	16	17	20	13
Age (years)	67 (25–85)	63.5 (26–87)	70 (36–90)	63 (42–77)
BMI (kg/m2)	26.3 ± 3.4 (18.7–31.3)	27.3 ± 3.0 (19.5–32.5)	28.1 ± 2.9 (20.2–31.1)	27.2 ± 3.1 (19.1–30.8)

Transverse images were reconstructed with a 5 mm slice thickness in 5 mm increments using a hard convolution kernel in filtered back-projection (B60f), a matrix of 512 × 512, and a lung window (center, –500 HU; width, 2000 HU) for the overview. For further analysis, transverse 1.0 mm slices in 0.7 mm increments were reconstructed. For 3D evaluation, coronal and sagittal reformations with a 2 mm slice thickness in 2 mm increments were reconstructed. Additional patient parameters such as body mass index (BMI) and scan range were also evaluated.

Results

The median exam duration in group 4 (2.4 s; range, 0.6 s) was significantly shorter compared to groups 1 (7.7 s; range, 1.6 s; P < 0.05) and 2 (4.2 s; range, 1.5 s; P < 0.05; Table 2). In addition, the median exam duration was significantly shorter in group 1 relative to group 3 (2.6 s; range, 0.6 s; P < 0.05; Table 2). Furthermore, there was no statistical difference in BMI (P > 0.5; Table 1), DLP, CTDIvol, or mAs (P > 0.05; Table 2) between groups.

OPEN IN VIEWER

Table 2.

Examination parameters.

	Group 1	Group 2	Group 3	Group 4	P value: Group 1 vs. Group 2/3/4
Scan range (cm)	35.5 (31.9–39.3)	38.5 (30.8–44.5)	36.5 (32.0–40.6)	38.8 (31.7–45.3)	0.06
Scan duration (s)	7.7 (6.9–8.5)	4.2 (3.3–4.8)	2.6 (2.3–3.1)	2.4 (2.2–2.9)	1 vs. 2: 0.3; 1 vs. 3: 0.003; 1 vs. 4: 0.0005 2 vs. 3: 0.1; 2 vs. 4: 0.05; 3 vs. 4: 0.6
Dose-length product (mGy × cm)	271 (171–443)	266 (143–323)	253 (166–483)	236 (176–311)	0.1
CTDI vol.	7.4 (4.6–17.8)	7.2 (4.2–13.5)	7.5 (4.7–13.9)	6.8 (3.8–10.4)	0.6
mAs	110 (68–265)	110 (69–205)	105 (62–200)	99 (55–130)	0.4
Image noise (HU)	24.8 (16.0–30.8)	27.0 (23.1–35.8)	34.3 (33.1–35.9)	35.8 (32.6–47.8)	1 vs. 2: 0.2; 1 vs. 3: <0.01; 1 vs. 4: <0.01 2 vs. 3: <0.01; 2 vs. 4: <0.01; 3 vs. 4: 0.3
SNR	2.0 (1.4–3.1)	1.8 (1.2–2.8)	1.4 (1.3–1.6)	1.5 (0.6–1.6)	1 vs. 2: 0.05; 1 vs. 3: <0.01; 1 vs. 4: <0.01 2 vs. 3: <0.01; 2 vs. 4: 0.02; 3 vs. 4: 0.5

Evaluation of image noise as an objective image quality parameter showed that slower gantry rotation time protocols in groups 1 and 2 resulted in significantly less image noise compared with faster rotation times in groups 3 and 4 (P < 0.01; Table 2; Fig. 1). The SNR-values were influenced additionally to the image noise (Table 2). The best image quality was obtained with 0.5 and 1 s/rot rotation times (Table 3), although image quality differences were not significant overall (Fig. 2). Interestingly, only Radiologist/Observer 2 found significant differences in subjective image analysis ratings between groups 1 and 3 (P = 0.007) and groups 1 and 4 (P = 0.04; Table 3), the rest did not reach statistical significance. However, there was an obvious increase of image noise and streak artifacts in higher gantry rotation time groups (Tables 2 and 3). As image noise increased, artifacts became more visible, which influenced subjective image scoring in some cases. OPEN IN VIEWER

OPEN IN VIEWER

Fig. 2.

Sufficient image quality in all cases: Four patients: (a) 1.0 s / rot. rotation time – nearly no noise artifacts in the picture and surroundings; (b) 0.5 s / rot. rotation time – less noise artifacts in the picture and surroundings; (c) 0.33 s / rot. rotation time – increasing noise artifacts in the picture and surroundings; (d) 0.28 s / rot. rotation time – most artifacts in the picture and surroundings.

OPEN IN VIEWER

Table 3.

Image quality rating between the different groups.

		Group 1	Group 2	Group 3	Group 4	Overall Kappa
Streak artefacts	Observer 1	1 ± 0.5 (1–2)	1 ± 1 (1–3)	2 ± 0.5 (1–3)	2 ± 0.9 (1–3)	0.8
	Observer 2	1 ± 0.5 (1–3)	2 ± 0.5 (1–3)	2 ± 0.8 (1–3)	3 ± 1 (1–4)
Kappa		0.93	0.82	0.78	0.65
Overall quality	Observer 1	2 ± 0.6 (1–3)	1 ± 0.7 (1–3)	2 ± 0.6 (1–3)	2 ± 0.8 (1–3)	0.6
	Observer 2	1 ± 0.5 (1–2)	2 ± 0.6 (1–3)	2 ± 0.7 (1–3)	2 ± 0.7 (1–3)
Kappa		0.40	0.64	0.75	0.8

Discussion

Ongoing research has shown that increasing detector width and rotation speed not only make CT imaging faster, but also improves workflow, making this imaging modality more robust against motion artifacts (4,6). Motion is a very common and significant problem in CT imaging because patients are not always able to hold their breath for a defined period of time or may simply be non-compliant (4). Unfortunately, accelerating rotation speed can introduce additional artifacts, noise, or other image quality issues that are not visible at slower speeds. Therefore, we investigated the influence of gantry rotation time on various imaging parameters and radiation exposure during chest CT examinations.

Our study shows for the first time that the slowest tube rotation time setting (1 s/rot) results in a breath-hold period of 6–9 s (Table 2). All patients examined in this mode were able to and did hold their breath for this period of time. Although most previously published reports focus on “non-breath-hold” imaging of the chest (7,8), data on both “breath-hold” and “non-breath-hold” situations is important for adaptation of CT protocols in different clinical situations. The SNR values showed some differences, so group 1 and 2 offered the highest values compared to group 3 and 4 (Table 2).

A previous abdominal imaging report from 2008 demonstrated that using a faster tube-detector rotation time saturates the X-ray tube more often (at its limit of current), leading to an increase in image noise and deterioration of image quality (9). Conversely, a 2012 report concluded that rotation time and pitch factors have only a small influence on image quality (10). Present results agree with both studies in that although faster tube-detector rotation time minimally influenced overall image quality, image noise significantly increased. Although image quality was highest at longer gantry rotation times (0.5 and 1 s/rot; Table 3), differences between rotation times were not significant (Fig. 2). On the other hand, image noise was significantly higher using faster rotation (Tables 2 and 3). Moreover, with increasing image noise, artifacts sometimes become more visible, which influences subjective image scoring. Furthermore, image noise might influence image quality differently according to the body region examined.

Although a shorter CT scan time would technically reduce the radiation exposure time, it is unclear whether faster gantry rotation directly affects the absorbed dose. Interestingly, a 2006 report posited that using faster gantry rotation times at a high tube load might lead to a larger beam focus, which might increase overall radiation exposure (11). However, our data demonstrated that there was no statistical difference in DLP or CTDIvol between rotational times investigated in this study (P > 0.05; Table 2).

In conclusion, faster CT gantry rotation speed reduces scan time and motion artifacts. However, current results, together with previous studies, demonstrate that accelerating rotation time does not always provide the best image quality, noise, or artifact presentation. Furthermore, in CT machines with wide detectors, a fast gantry rotation is not always necessary to acquire images in a short period of time. Therefore, a slower CT gantry rotation speed is still recommended for higher image quality in some clinical cases.