Wide-detector CT combined iterative reconstruction in pediatric low-dose scan of the paranasal sinus

Abstract

BACKGROUND:

Computed tomography (CT) is considered a standard modality for imaging the paranasal sinus (PS), but increasingly radiation dose is of concern, especially in children.

Objective:

This study aims to investigate the feasibility of using a 320-detector CT scanner with a 16 cm wide-detector combined with iterative reconstruction (IR) algorithm to further reduce radiation dose when scanning the PS.

METHODS:

A total of 90 children who underwent CT of the PS were randomly allocated into three groups namely, (1) the experimental group using low-dose wide-detector scan (n = 30, 9±4 years); (2) low-dose helical group (n = 30, 9±4 years); and (3) pediatric conventional group (n = 30, 8±4 years). Statistical software SPSS 19.0 was used for one-way ANOVA analysis of the general data (age, BMI), image quality, and radiation dose. Multiple comparisons of data without homogeneity of variance were analyzed by Bonferroni test and Tamhane’s test.

RESULTS:

All patients underwent successful CT examinations. No significant differences in the general data and image quality evaluation were detected between three groups (all P values > 0.05). CTDIvol and DLP were 2.87 mGy and 32.58 mGy·cm in the experimental group, 4.92 mGy and 70.84 mGy·cm in the low-dose helical group, and 9.95 mGy and 131.83 mGy·cm in the conventional group, respectively, which were significantly different among these three groups as indicated by multiple comparisons (all P values < 0.05). In the experimental group, the effective radiation dose was 0.07 mSv, which was reduced by 76% and 56% comparing to the conventional group and the low-dose helical group, respectively.

CONCLUSIONS:

The 320-detector CT scanner equipped with the wide-detector combined with IR can further reduce radiation dose, while maintaining good image quality comparing to the low-dose helical or pediatric modes.

Keywords

320-detector computed tomography (CT)wide-detector iterative reconstruction pediatric paranasal sinus low dose

1 Introduction

In children, the paranasal sinuses (PS) are difficult to evaluate due to variability in shape, size and patterns of development. A wide range of conditions, such as traumatic, inflammatory and neoplastic diseases, and congenital abnormalities, usually affects these sinuses [1 –4]. Currently, the CT is the most commonly used and represents the gold standard for imaging the PS because of the high spatial resolution, accuracy and rapidity [2, 3]. Nevertheless, CT carries a high degree of radiation that usually affects X-ray sensitive organs, such as the eye lens and thyroid gland because they are nearby the PS [2 , 6]. Thus, it is extremely important to optimize the CT examination by reducing the radiation dose, thus protecting the patients [3 , 8].

At present, the most popular methods used for reducing the radiation dose include iterative reconstruction (IR), reducing the tube current or voltage, automatic tube voltage or current, and shorten scanning acquisition time [2 , 8]. The combination of these methods, especially those involving IR [4 , 8], have shown an optimal effect. A recent development of 16 cm wide-detector, which uses a single axial rotation, has attracted a lot of attention. Using this tool, researchers have been largely focusing on investigating cardiac or pulmonary imaging [9 , 11]; while a low-dose CT has been increasingly applied in pediatric imaging [6 , 13]. Nonetheless, low-dose CT has been rarely studied in relation to wide-detector combined IR to reduce radiation dose [9]. The aim of this study was to investigate the feasibility of 320-detector CT scanner with 16 cm wide-detector combined with IR, to further reduce the radiation dose when imaging the PS.

2 Materials and methods

2.1 Study subjects

A total of 90 children less than 14 years old who underwent CT of the PS from December 2017 to August 2018 were recruited into this study. Patients were randomly divided into three groups (random number table): 1, the experimental group with low-dose wide-detector scan; 2, the low-dose helical group, in which a commonly used helical CT scan mode (64-detector CT) was applied but with the same tube voltage, current, and rotation time as the experimental group; 3, the conventional group, namely pediatric scan mode. Patients who failed to comply with examinations due to motion or other reasons were excluded from the study.

2.2 Inspection instrument

All CT examinations were performed with Toshiba 320-detector CT scanner (Toshiba Aquilion ONE), The major difference between this scanner with other commonly used CT scanner (i.e., 64-detector CT) is that it allows the maximally adaptive 16 cm wide detectors in contrast to 3.2 to 4.0 cm-wide detectors on 64-detector CT. Additionally, it can provide both axial volumetric acquisitions (i.e., volume mode with a maximal z-axis coverage of 16 cm) and conventional helical acquisitions with adaptive collimation, such as acquisition configuration of 160×0.5 mm, 64×0.5 mm. In particular, a pediatric acquisition mode (“sinuses child”, named by the manufacturer) was also provided by the scanner with lower kV (i.e., the default of 100 kV compared to 120 kV in adults when imaging the PS).

2.3 Imaging method

All metal objects affecting the quality of image were removed from the examination area, and protective measures (physical shielding with lead clothes) were deployed on other areas. Patients were placed in supine position with the head in front and instructed meticulously not to move during examination. The range in three sets of scan was up to the superior wall of the frontal sinus, and down to the bottom wall of the maxillary sinus; the scan length (L) was fixed to 10 cm. The field of vision (FOV) covered the entire PS, lateral to the bilateral mastoids, at size of 20 cm×20 cm.

Parameters in the experimental group were: wide-detector mode, collimation width 320×0.5 mm with an adjusted z-axis coverage from 16 cm to 10 cm, tube voltage 100 kV, tube current 50 mA, rotation time 0.5 s, scan duration 0.5 s, slice thickness 0.5 mm; parameters in the low-dose helical scan group were: helical mode, collimation width 64×0.5 mm, tube voltage 100 kV, tube current 50 mA, rotation time 0.5 s, scan duration 3.2 s, slice thickness 0.5 mm, pitch factor 0.641; and the parameters in the conventional group were: helical mode, collimation width 64×0.5 mm, default pediatric conditions 100 kV and 100 mA, rotation time 0.5 s, scan duration 3.2 s, slice thickness 0.5 mm, pitch factor 0.641. Image reconstruction was performed with iterative standards in the experimental and low-dose helical CT groups, and with filtered back projection (FBP) in the conventional group.

The acquired raw data were reconstructed into two sets in three groups: 1, both axial slice thickness and intersection of 3 mm using soft tissue window, and both axial slice thickness and intersection of 4 mm using bone window mainly for image quality comparisons in three groups [3] and daily practice; 2, submillimeter reconstruction with both 0.5 mm slice thickness and intersection using both bone and soft window was also performed for multiplanar reformation evaluation because of the high spatial resolution and contrast of bones and soft tissue. Other specifications included the age and body mass index (BMI).

This study was approved by the Ethical Committee of Wuxi Second People’s Hospital. In addition, all participants signed the informed consent by their parents.

2.4 Image and data analysis

2.4.1 Subjective evaluation

A five-point Likert scale was adopted to evaluate the image quality in depicting bony and soft tissue anatomical landmarks, such as the ostiomeatal complex (OMC), septa of the ethmoid cells, and eye muscles on the axial images in three groups with the same window width and center [2, 3]: 5, diagnostic and excellent image quality (without artifact and optimal contrast between bony and soft tissue structures); 4, diagnostic and good image quality with mild noise or artifacts; 3,diagnostic image quality but with moderate image noise or artifacts; 2, limited diagnostic image quality with substantial image noise or artifacts; and 1,non-diagnostic, referred to the poor image quality due to severe image artifacts or noise. Two blinded readers (Q.K. and P.Q. with 10 and 4 years experience in CT, respectively) assessed the image quality in consensus based on the criteria above [5].

2.4.2 Objective evaluation

Comparisons were made in the image noise with mean standard deviation (SD) and contrast-to-noise ratios (CNR). The CT attenuated values with SD, which represented the noise of relevant tissue, were measured in Hounsfield units (HU) on the same axial image using three circle ROIs with an area of 8– 15 mm² at three deferent positions: inferior turbinate mucosa, air (outside of the patient’s head) [2], and the fat of infratemporal fossa. The measurement was repeated twice by two readers, and the mean result of ROIs was recorded as ROI_mucosa, ROI_air, and ROI_fat, respectively. The SD value of air was served as the image background noise, and it was recorded as SD_back [2]. CNR was quantified using the following formula: CNR = (ROI_mucosa – ROI_fat)/SD_back.

2.4.3 Radiation dose

To compare the radiation dose in three groups, CT dose index volume (CTDI_vol) and dose length product (DLP) were recorded after examination. According to the European guidelines, DLP and effective radiation dose (ED) were calculated using the following equation: DLP = CTDI_vol×L; ED = k×DLP, and the conversion factor (k) was of 0.0023 mSv/mGy·cm [14].

2.4.4 Statistical analysis

Data were analyzed using SPSS 19.0. Measurement data were expressed by x±s. One-way ANOVA was adopted for 3-group comparisons in the general data, measured results of CT values, calculated results of radiation dose, and evaluation score of image quality. Multiple comparisons of data without homogeneity of variance were analyzed with Bonferroni test and Tamhane’s test. P < 0.05 was considered statistically significant.

3 Results

CT examination was successfully preformed in all children and no need for sedation. In addition, scan length of 10 cm could cover the entire PS. In three groups, the mean age and BMI were 8.27, 8.63, and 8.37 years, respectively, and 19.18, 17.28, and 18.39 kg/m², respectively. No significant difference in the general data was observed between groups in view of age and BMI (F = 0.099, 1.972, respectively, all P values >0.05) (Table 1).

Table 1
Statistical comparisons in the general data and CTDIvol

Groups Number Age (years) BMI (kg/m²) CTDI_vol (mGy)^* DLP (mGy·cm)^* ED (mSv)

EG 30 8.27±3.24 19.18±4.24 2.87±0.20 32.58±5.97 0.07±0.01

HG 30 8.63±3.20 17.28±3.65 4.92±0.19 70.84±6.13 0.16±0.14

CG 30 8.37±3.46 18.39±3.23 9.95±0.58 131.83±7.36 0.30±0.01

F 0.099 1.972 2955 1770.95

P 0.906 0.145 < 0.001 < 0.001

Groups	Number	Age (years)	BMI (kg/m²)	CTDI_vol (mGy)^*	DLP (mGy·cm)^*	ED (mSv)
EG	30	8.27±3.24	19.18±4.24	2.87±0.20	32.58±5.97	0.07±0.01
HG	30	8.63±3.20	17.28±3.65	4.92±0.19	70.84±6.13	0.16±0.14
CG	30	8.37±3.46	18.39±3.23	9.95±0.58	131.83±7.36	0.30±0.01
F		0.099	1.972	2955	1770.95
P		0.906	0.145	< 0.001	< 0.001

^*Note: multiple comparisons of three groups by Bonferroni test and Tamhane’s test showed P < 0.001 in CTDI_vol and DLP. EG: experimental group; HG: Low-dose helical group; CG: pediatric conventional group.

3.1 Dose comparisons

Multiple comparisons in the experimental group, conventional group, and low-dose helical group are shown in Table 1. Both the CTDI_vol and DLP in three groups showed significant difference (F = 2955, 1770.95, respectively, all P values < 0.001) with the lowest ED dose of 0.07 mSv in the experimental group, which was reduced by 76% and 56% compared with the conventional group (0.30 mSv) and the low-dose helical group (0.16 mSv), respectively.

3.2 Image quality

The subjective evaluation is shown in Table 2. In the experimental group, three patients presented mild image noise and one had mild motion artifacts. In contrast, three patients showed mild motion artifacts in the low-dose helical group and the other two in the conventional group, respectively. Although the image quality in the experimental group demonstrated slightly lower mean scores (4.87) than that of the low-dose helical group (4.90) and conventional group (4.93), multiple comparisons in three groups showed insignificant difference (F = 0.289, P = 0.750), and the diagnostic criteria were met in all groups in this study (Figs. 1 and 2).

Fig.1

Axial reformatted images with 3 mm slice thickness at the inferior turbinate level using three CT scan modes. IR image using low-dose wide-detector mode in a 6-year-old child (a) demonstrates bilateral maxillary sinusitis (black arrows) and adenoid hypertrophy (white asterisk) as well. Although a slightly increased image noise may be perceived compared to the low-dose helical mode in a 14-year-old child (b) and pediatric conventional mode using FBP reconstruction in a 5-year-old child (c), the standard deviation (SD) at three different positions on three groups is similar, and all images can meet the diagnostic criteria. Also, slightly thickening of ethmoid and maxillary mucosa both in (b) (white arrows) and (c) (white arrows) can be depicted.

Fig.2

Coronary reformatted images with 0.5 mm slice thickness at the ostiomeatal complex (OMC) level using IR with low-dose wide-detector mode in a 6-year-old child (a), low-dose helical mode in a 14-year-old child (b), and three CT scan modes, and pediatric conventional mode using FBP reconstruction in a 12-year-old child (c), respectively. Either (a) or (b), or (c) clearly shows bony and soft tissue anatomical landmarks, such as the septa of the ethmoid cells (white arrows), and eye muscles (white arrowheads), as well as the thickening of maxillary and/or ethmoid mucosa (white asterisks).

Table 2

Comparisons of 3 groups in the noise, CNR, and image quality

Groups	SD (IT mucosa)	SD (ITF fat)	SD (air)	CNR	Subjective evaluation
EG	8.87±2.37	9.33±2.93	5.38±1.98	30.07±10.35	4.87±0.43
HG	8.25±2.05	8.77±3.22	5.05±1.50	32.96±13.20	4.90±0.31
CG	7.99±1.38	8.37±2.40	4.36±1.81	36.28±9.47	4.93±0.25
F	1.581	0.860	2.551	2.347	0.289
P	0.212	0.427	0.084	0.102	0.750

IT: inferior turbinate; ITF: infratemporal fossa.

The objective evaluation is also summarized in Table 2. Compared to the low-dose helical group and pediatric conventional group, the relatively increased noises at three deferent positions and decreased CNR in the experimental group were observed, but the differences were also not significant (CNR: F = 2.347, P = 0.102).

4 Discussions

The use of pediatric CT has been rapidly increasing and estimated to represent about 11% of all CT procedures [15]. Yet, due to its increased radiation exposure, pediatric CT has become a public health concern. Some epidemiological studies have suggested an association between CT radiation dose and cancer risk, especially when applied in teenagers [16 –20]. Thus, it is of extreme importance to reduce the radiation dose when ensuring the correct diagnosis. At present, the ALARA (“as low as reasonably achievable”) principle has been advocated worldwide [2 , 21– 23], this means that maintaining diagnostic quality while reducing radiation dose is recommended. In other words, the use of all available equipment-specific dose reduction techniques has been strongly endorsed [24].

Currently, tube voltage reduction, tube current reduction, and scan duration reduction are some of the conventional methods used for radiation dose reducing. However, reducing tube voltage or current may substantially increase noises which decrease the image quality [5, 25]. Some researchers have investigated alternative imaging techniques, such as cone-beam CT (CBCT), or even magnetic resonance imaging (MRI) for the PS examinations in order to reduce or prevent radiation dose. Nevertheless, either CBCT or MRI has longer acquisition times and is inferior to CT in evaluating bony structures in the PS [3, 4]. Recently, IR technique has been introduced into the clinical practice because it can reduce radiation dose while maintaining a good image quality [4 , 25]. Then again, the PS is an organ containing air, with good contrast, which holds a natural advantage in reducing the radiation dose. At present, the proposed methods for reducing CT dose of the PS include IR combined with low tube voltage or current [4 –6] and an extremely low radiation dose (< 0.1 mSv) could be performed on CT of the PS in children [4 –6].

In the present study, we used 320-detector CT scanner with 16 cm wide-detector combined with IR to further reduce the radiation dose in the PS. Compared to a conventional low-dose helical scan and conventional pediatric conditions, we found a significantly reduced radiation dose in the experimental group with the dose savings of 56%, 76%, respectively. Some studies have presented that the ED allows for a significant reduction while maintaining a good diagnostic image quality in CT of the PS when IR combined with lower tube voltage or current [4 –6]. In this study, the mean ED was 0.07 mSv, which was lower than the report by Schaafs et al. [5] but higher than those reports by Sun et al. [4] and Pirimoglu et al. [6]. We think that the differences may result from different protocols for imaging the PS, e.g., tube voltage, current, or scan length and conversion factor adopted as well.

The other benefits of volumetric scanning mode (wide-detector mode in this study) with sub-second acquisition time is no over-ranging, which occurs in helical scan, this may reduce patient radiation dose. Also, the mode can effectively shorter scan duration and may obviate the need for sedation in those children [9, 12], though mild motion artifact was shown in one of 30 patients. Considering the image quality, slightly increased noises and decreased CNRs in the experimental group were also observed compared to other two groups. However, the mean quality score showed insignificant difference and met the needs for diagnosis of the PS.

This study has following limitations: 1, a tube voltage of 100 kV was applied in this study, which was not the lowest voltage setting compared to 80 kV reported in the cutting-edge studies [4 , 12]; 2, in order to properly compare radiation dose changes caused by different scan modes, the rotation time in wide-detector scan was set to be 0.5 s (same as in helical mode) but not 0.35 s, which is an optical setting usually used in cardiac CT. Thus, further studies should be performed in our scanner, and 3, wide-detector CT scanner is still uncommon, which makes it impossible to be used in large-scale application clinically.

5 Conclusion

Axial volumetric 320-detector CT with wide-detector mode is a feasible method for imaging the PS in children. Low-dose scan combined with IR technique on the scanner can further reduce radiation dose compared to the low-dose helical or pediatric mode while maintaining the image quality.

Conflicts of interest

This study was carried out in accordance with the following statement of contribution, accepting no sponsorship from any relevant companies. The conflicts of interest involved no parties concerned.

Authorship contribution claim

Que Kong is in charge of study design, data collection, manuscript preparation, and statistical analysis of results; Fengqi Lu instructs the manuscript preparation and revision; Yu Gao, Peng Qiao, and Min Shao are in charge of data collection.

References

Borner

, Anschuetz

, Kaiser

, et al., Blunt nasal trauma in children: A frequent diagnostic challenge, Eur ARCH OTO-RHINO-L 276 (2019), 85–91.

Schell

, Bauer

R.W.

, Lehnert

, et al., Low-dose computed tomography of the paranasal sinus and facial skull using a high-pitch dual-source system– first clinical results, Eur Radiol 21 (2011), 107–112.

Wuest

, May

, Saake

, et al., Low-Dose CT of the paranasal sinuses: Minimizing X-Ray exposure with spectral shaping, Eur Radiol 26 (2016), 4155–4161.

Sun

, Zhang

, Duan

, et al., Application of a full model-based iterative reconstruction (MBIR) in 80 kVp ultra-low-dose paranasal sinus CT imaging of pediatric patients, Radiol Med 123 (2018), 117–124.

Schaafs

L-A.

, Lenk

, Hamm

, et al., Reducing the dose of CT of the paranasal sinuses: Potential of an iterative reconstruction algorithm, Dentomaxillofac Radiol 45 (2016), 20160127.

Pirmoglu

, Sade

, Sakat

M.S.

, et al., Low-dose noncontrast examination of the paranasal sinuses using volumetric computed tomography, J Comput Assit Tomogr 42 (2018), 482–486.

Dauer

L.T

, Ainsbury

E.A.

, Dynlacht

, et al., Guidance on radiation dose limits for the lens of the eye: Overview of the recommendations in NCRP Commentary No. 26, Int J Radiat Biol 93 (2017), 1015–1023.

Gay

, Pavia

, Pierrat

, et al., Dose reduction with adaptive statistical iterative reconstruction for paediatric CT: Phantom study and clinical experience on chest and abdomen CT, Eur Radiol 24 (2014), 102–111.

Johnston

J.H.

, Podberesky

D.J.

, Yoshizumi

T.T.

, et al., Comparison of radiation dose estimates, image noise, and scan duration in pediatric body imaging for volumetric and helical modes on 320-detector CT and helical mode on 64-detector CT, Pediatr Radiol 43 (2013), 1117–1127.

10.

Shuman

W.P.

, Green

D.E.

, Busey

J.M.

, et al., Wide-detector axial CT versus 4cm detector helical CT for transcatheter aortic valve replacement: Iodine dose, radiation, and image quality, Clin Imaging 40 (2016), 1213–1218.

11.

F.Q.

, Gao

, Kong

, et al., Application of 640-slice CT wide-detector volume scan in low-dose CT pulmonary angiography, J Xray Sci Technol 27(2019), 197–205.

12.

Kroft

L.J.M.

, Roelofs

J.J.H.

, Geleijhs

, Scan time and patient dose for thoracic imaging in neonates and small children using axial volumetric 320-detector row CT compared to helical 64-, 32-, and 16- detector row CT acquisitions, Pediatr Radiol 40 (2010), 294–300.

13.

, Wang

, Li

, et al., 3D computed tomography angiography as a novel post-processing approach in diagnosis of pediatric malignant bone tumors, J Xray Sci Technol 26 (2018), 147–154.

14.

Bongarts

, Golding

S.J.

, Jurik

A.G.

, et al., European guidelines on quality criteria for computed tomography, Report EUR 16262 (1999).

15.

Mettler

F.A.

, Wiest

P.W.

, Locken

J.A.

, et al., CT scanning: Patterns of use and dose, J Radiol Prot 20 (2000), 353–359.

16.

Brenner

D.J.

, Elliston

C.D.

, Hall

E.J.

, et al., Estimated risks of radiation-induced fatal cancer from pediatric CT, AJR 176 (2001), 289–296.

17.

de González

A.B.

, Mahesh

, Kim

K.P.

, et al., Projected cancer risks from computed tomographic scans performed in the United States in 2017, Arch Intern Med 169 (2009), 2071–2077.

18.

Pearce

M.S.

, Salotti

J.A.

, Little

M.P.

, et al., Radiation exposure from CT scans in childhood and subsequent risk of leukaemia and brain tumours: A retrospective cohort study, Lancet 380 (2012), 499–505.

19.

Mathews

J.D.

, Forsythe

A.V.

, Brady

, et al., Cancer risk in 680 000 people exposed to computed tomography scans in childhood or adolescence: Data linkage study of 11 million Australians, BMJ 346 (2013), f2360.

20.

Miglioretti

D.L.

, Johnson

, Williams

, et al., The use of computed tomography in pediatrics and the associated radiation exposure and estimated cancer risk, JAMA Pediatr 167 (2013), 700–707.

21.

Mayo-Smith

W.W.

, Hara

A.K.

, Mahesh

, et al., How I do it: Managing radiation dose in CT, Radiology 273 (2014), 657–672.

22.

Sodhi

K.S.

, Krishna

, Saxena

A.K.

, et al., Clinical application of ‘justification’ and ‘optimization’ principle of ALARA in pediatric CT imaging: “How many children can be protected from unnecessary radiation?”, EJR 84 (2015), 1752–1757.

23.

Journy

N.M.Y.

, Dreuil

, Boddaert

, et al., Individual radiation exposure from computed tomography: A survey of paediatric in French university hospital, 2010-2013, Eur Radiol 28 (2018), 630–641.

24.

Remy-Jardin

, Pistolesi

, Goodman

L.R.

, et al., Management of suspected acute pulmonary embolism in the era of CT angiography: A statement from the Fleischner Society, Radiology 245 (2007), 315–329.

25.

Schulz

, Beeres

, Bodelle

, et al., Performance of iterative image reconstruction in CT of the paranasal sinuses: A phantom study, AJNR 34 (2013), 1072–1076.