Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction

Abstract

Background

The novel advanced modeled iterative reconstruction (ADMIRE) algorithm in ultra-high-resolution (UHR) computed tomography (CT) of the temporal bone has not yet been systematically evaluated.

Purpose

To assess the potential of ADMIRE in temporal bone UHR CT.

Material and Methods

Forty-four patients who underwent UHR CT of the temporal bone using z-axis UHR protocol were retrospectively selected for analysis. Images were reconstructed using filtered back projection (FBP) and ADMIRE with multiple strength levels. Regions of interest were drawn in the posterior fossa and petrous bone. The average density (in Hounsfield units [HU]) and the image noise (standard deviation of density values) were extracted. The signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were then calculated. Additionally, a subjective qualitative analysis was performed using a five-point Likert scale. The potential dose reduction was extrapolated from objective analysis and confirmed in an additional phantom study.

Results

The image noise was significantly lower, and the SNR and CNR were significantly higher in ADMIRE reconstructions levels A2–A5 than in FBP (P < 0.001, respectively). Subjective image quality was significantly higher in ADMIRE levels A2–A5 than in FBP (P < 0.001) and it was rated best in ADMIRE level A3. Confirmed by the results from the phantom study, a dose reduction of at least 40% was feasible while maintaining image quality.

Conclusions

The ADMIRE reconstruction algorithm significantly improves image quality and reduces noise on temporal bone UHR CT scans. Thus, it allows for substantial dose reduction.

Keywords

Computed tomography CT temporal bone iterative reconstruction advanced modeled iterative reconstruction ADMIRE

Introduction

Computed tomography (CT) of the temporal bone allows for detailed evaluation of the bony anatomy of the middle and inner ear and identification of potential abnormalities and anatomic variants (1,2). In preoperative management it may influence the indication for any surgical intervention as well as the approach planning (1,3 –5). Reduction of ionizing radiation exposure is of high concern, especially in the pediatric population (3,6). However, the use of low-dose protocols may drastically impair the image quality of small anatomical structures, such as the stapes (7).

Iterative reconstruction algorithms in CT are used to process the raw projection or reconstructed data repeatedly to reduce image noise while preserving image quality and display of subtle details (8,9). The advanced modeled iterative reconstruction (ADMIRE) algorithm belongs to the category of statistical iterative reconstruction algorithms, also called “model-based iterative reconstruction algorithms” (8,10,11). ADMIRE uses statistical modeling both in the raw projection data and in the image domains. Thus, different statistical weighting is applied according to the quality of the projection data. This is different from the traditional filtered back projection (FBP) imaging, in which all projections are weighted equally, independent of their quality (8). Compared with FBP, ADMIRE has shown potential in substantially reducing radiation dose while maintaining low-contrast detectability (8).

However, the use of ADMIRE for reconstruction of ultra-high resolution (UHR) temporal bone CT has not yet been sufficiently evaluated.

Therefore, this study sought to assess the potential of ADMIRE regarding noise reduction and ameliorating image quality in UHR CT of the temporal bone.

Material and Methods

Study design and ethics

This study is a retrospective cross-sectional observational study following the STROBE guidelines (12) that was approved by our university’s local institutional review board (reference no. 260/2018BO). The trial was conducted based on the principles of the International Conference on Harmonization: Good Clinical Practice guidelines and the 2013 revised version of the Declaration of Helsinki. Written informed consent was waived given the retrospective study design. All data were anonymized before image analysis.

Patient selection and stratification

The study group was selected from 58 consecutive patients suffering from profound sensorineural hearing loss between November 2017 and January 2018 for cochlear implant screening without any history of middle ear disease. The group consisted of 32 men and 26 women (mean age = 47.5 ± 21.3 years). Thereupon, 14 patients were excluded because of middle or inner ear anatomic variations (n = 5) and middle (n = 6) and inner ear pathologies (n = 3). The final study group included 44 patients (25 men [57%] and 19 women [43%]; mean age = 50.0 ± 17.9 years) with normal CT scans of the temporal bone.

Procedures and techniques

CT image acquisition and reconstruction

All scans were acquired using a third-generation single source CT (SOMATOM® Definition AS+, Siemens Healthcare, Erlangen, Germany) with a fully integrated circuit detector (Stellar® detector, Siemens Healthcare, Erlangen, Germany). The imaging parameters were as follows: gantry rotation time = 1.0 s; tube current = 230 reference mAs using an automatic tube current modulation (CARE Dose4D®, Siemens Healthcare, Erlangen, Germany) yielding an average effective tube current of 201 mAs and an average CT dose index (CTDI_vol) of 43.2; tube voltage = 120 kV; and pitch = 0.85. The effective detector collimation was 16 × 0.3 mm using a z-axis UHR and flying focal spot technique (13), resulting in an effective 0.4-mm slice thickness. The scans were reconstructed using FBP and ADMIRE (Siemens Healthcare, Erlangen, Germany) at five different strength levels (A1–A5) (14) for the corresponding bone kernel.

Potential dose reduction was extrapolated from CTDI_vol and image noise results from FBP and ADMIRE A3 reconstructions as follows:

C T D I_{e x t r a p o l a t e d} = \frac{C T D I_{a d m i n i s t e r e d} \times N o i s e_{A D M I R E A 3}}{N o i s e_{F B P}}

To confirm these approximations, an additional phantom study was performed. This phantom study was performed as twofold measurements on a head phantom (3M, St. Paul, MN, USA) using 44 different combinations of static tube voltage and tube current (see Suppl. Table 1), but otherwise identical scan and image reconstruction parameters as follows: tube voltage = 80 kV: tube current = 40, 80, 120, 160, 200, 240, 280, 320, 360, 400, and 440 mAs; tube voltage =100 kV: tube current = 40, 80, 120, 160, 200, 240, 280, 320, and 360 mAs; tube voltage = 120 kV: tube current = 40, 80, 100, 120, 140, 160, 180, 200, 220, 240, 260, 280, 300, and 320 mAs; tube voltage =140 kV: tube current = 40, 60, 80, 100, 120, 140, 160, 180, 200, and 220 mAs). The CTDI_vol was in the range of 2.5–70.7.

Table 1.

Anatomic structures of the temporal bone assessed in qualitative image analysis, adapted from Nauer et al. (6).

n	Anatomic structure	Review criteria
1	Cochlea	Contour
2	Cochlear patency	Assessment of calcifications
3	Spiral osseous lamina	Presence, integrity
4	Modiolus	Presence, integrity
5	Vestibular labyrinth	Contour, density
6	Vestibular aqueduct	Contour, density
7	Cochlear aqueduct	Contour, density
8	Cochlear nerve canal	Presence
9	Bill’s bar	Contour
10	Internal auditory canal	Contour
11	Facial nerve canal, cochlear segment	Contour, course
12	Facial nerve canal, tympanic segment	Contour, course
13	Facial nerve canal, mastoid segment	Contour, course
14	Middle ear cavity	Aeration
15	Ossicle chain, Malleus	Presence of all parts
16	Ossicle chain, Incus	Presence of all parts
17	Ossicle chain, Stapes	Presence of all parts
18	Oval window	Contour, footplate position
19	Round window	Contour, aperture
20	Round window niche	Contour, aeration
21	Internal carotid artery canal	Contour, osseous wall dehiscence
22	Jugular foramen	Contour, osseous wall dehiscence
23	Mastoid	Contour, aeration

Average density and image noise values as well as SNRs with their standard deviations from different regions of interest in temporal bone CT. SNR and CNR are dimensionless.

HU, Hounsfield units; SNR, signal-to-noise ratio; CNR, contrast-to-noise ratio; ROI, region of interest; A_n, n^th level of ADMIRE (advanced modeled iterative reconstruction); FBP, filtered back projection.

Image analysis and observer setting

Image analyses were performed independently by two board-certified physicians who were blinded to the clinical diagnoses. Both readers had seven years of experience in temporal bone cross-sectional imaging.

Image and region of interest (ROI) analyses were performed using syngo.via® (Siemens Healthcare, Erlangen, Germany). To enable comparability with previous studies (7,15 –17), the ROIs were drawn in the brain stem (ROI 1) and petrous bone of the otic capsule (ROI 2) and copied to the other patient reconstructions (Fig. 1).

Fig. 1.

Metric assessment from UHR CT of the temporal bone. Measurement of mean density (in Hounsfield units [HU]) and image noise (measured as standard deviation of density in HU) in axial reformatted reconstructions of UHR CT of the temporal bone. Regions of interest (ROI) were drawn in the posterior fossa and petrous bone.

For quantitative measurements, the following parameters were calculated from each ROI:

CT density in Hounsfield units (HU);

image noise, measured as the standard deviation (SD) of the density values in HU;

and the signal-to-noise ratio (SNR): $S N R = \frac{{average HU}_{R O I n}}{{SD}_{R O I n}}$ .

The contrast-to-noise ratio (CNR) was calculated as follows (16):

(iv) $C N R = \frac{(average {HU}_{R O I 2} - {average HU}_{R O I 1})}{\sqrt{\frac{(S D_{R O I 2}^{2} + S D_{R O I 1}^{2})}{2}}}$

Qualitative image analyses were adapted to those of Nauer et al. (7). Image quality for delineation of 23 anatomic substructures (Table 1) “that are the most susceptible to image-quality degradation with increasing image noise such as the stapes or the cochlear spiral osseous lamina” (7), and that are relevant for cochlear implant screening, were assessed using a five-point Likert scale of 1–5: 1 = no identifiable anatomic structures due to poor image quality; 2 = identifiable structures but no assembled details, resulting in insufficient image quality; 3 = anatomical structures still fully assembled in all parts with acceptable image quality; 4 = clear delineation of structures and good image quality; and 5 = very good delineation of structures and excellent image quality.

Statistical analyses

Data analyses were performed using IBM SPSS Statistics® Version 24 (IBM, Armonk, NY, USA). The intra-class correlation coefficient (ICC [3,1; absolute agreement]) was calculated for testing inter-observer agreement, and it was interpreted according to Cicchetti (18): 0.00–0.39 = poor; 0.40–0.59 = fair; 0.60–0.74 = good; and 0.75–1.00 = excellent correlation. The Shapiro–Wilk test proved normal distributions of all variables. Friedman’s two-way analysis of variance by ranks with post-hoc Bonferroni correction compared the SD, SNR, and CNR values of each reconstruction for both ROIs. It also compared the scores of each reconstruction from subjective image analysis. Tests of the two a priori hypotheses were conducted using Bonferroni adjusted alpha levels of 0.025 per test (0.05/2).

Results

Inter-rater agreement

Regarding quantitative image analysis, the inter-rater agreement was excellent (ICC = 0.923; 95% confidence interval [CI] = 0.901–0.955) for ROI delineations. For qualitative image analysis, the inter-rater agreement was good (ICC = 0.714; 95% CI = 0.601–0.796).

Quantitative image analysis

Table 2 shows the average CT density and image noise values with their standard deviations from each ROI. The table includes the calculated SNR and CNR. Image noise was significantly lower in ADMIRE levels A2–A5 than in FBP both in ROI 1 (P < 0.001) and ROI 2 (P < 0.001). Additionally, the SNR was significantly higher in ADMIRE levels A2–A5 than in FBP both in ROI 1 (P < 0.001) and ROI 2 (P < 0.001). Finally, the CNR was significantly higher in ADMIRE levels A2–A5 than in FBP (P < 0.001). All values are illustrated in Figs. 2 and 3.

Fig. 2.

Density, image noise, and SNR in FBP and ADMIRE in UHR CT of the temporal bone. Box plots illustrate density (a, d), image noise (b, e), and the SNR (c, f) of FBP and ADMIRE reconstructions in temporal bone UHR CT. The values were extracted from ROI 1 (posterior fossa; a–c) and ROI 2 (petrous bone; d–f). FBP, filtered back projection; A_n, n^th level of ADMIRE (advanced modeled iterative reconstruction); HU, Hounsfield units; ROI, region of interest; UHR, ultra-high resolution; CT, computed tomography.

Fig. 3.

CNR ratio in FBP and ADMIRE in UHR CT of the temporal bone. Box plots illustrate the calculated CNR density of FBP and ADMIRE reconstructions in temporal bone UHR CT. FBP, filtered back projection; A_n, n^th level of ADMIRE (advanced modeled iterative reconstruction); HU, Hounsfield units; ROI, region of interest; UHR, ultra-high resolution; CT, computed tomography.

Table 2.

Values from objective ROI analysis in temporal bone CT.

	FBP	A1	A2	A3	A4	A5
ROI 1 – posterior fossa
Density (HU)	42.1 ± 4.8	41.3 ± 4.7	41.0 ± 4.6	40.7 ± 4.6	40.1 ± 4.6	39.7 ± 4.4
Noise (HU)	187.4 ± 18.6	160.0 ± 16.0	133.0 ± 13.4	106.1 ± 10.9	79.4.3 ± 8.3	53.1 ± 5.8
SNR	0.23 ± 0.03	0.26 ± 0.04	0.31 ± 0.04	0.39± 0.06	0.51 ± 0.08	0.76 ± 0.11
ROI 2 – petrous bone
Density (HU)	1751.4 ± 50.4	1749.6 ± 50.0	1746.7 ± 50.0	1744.0 ± 50.1	1741.8 ± 50.4	1739.2 ± 50.8
Noise (HU)	239.3 ± 29.9	206.9 ± 25.9	175.0 ± 21.9	143.7 ± 18.0	113.6 ± 14.7	85.7 ± 12.6
SNR	7.43 ± 0.98	8.58 ± 1.13	10.13 ± 1.31	12.32 ± 1.57	15.57 ± 2.02	20.70 ± 3.03
CNR	8.05 ± 0.97	9.35 ± 1.12	11.10 ± 1.32	13.64 ± 1.60	17.57 ± 2.08	24.17 ± 3.08

Average density and image noise values as well as SNRs ratios with their standard deviations from different ROIs in temporal bone CT. SNR and CNR are dimensionless.

Qualitative image analysis

Table 3 demonstrates the results of the subjective image analysis for temporal bone structure delineation. Subjective image quality was significantly higher in ADMIRE levels A2–A5 than in FBP (P < 0.001). The overall image quality was rated highest in ADMIRE level A3 (Fig. 4).

Fig. 4.

Subjective image quality in FBP and ADMIRE in UHR CT of the temporal bone. Bar graphs illustrate the overall subjective image quality using a five-point Likert scale of FBP and ADMIRE reconstructions in UHR CT of the temporal bone, in the range of 1–5, as follows: 1 = no identifiable anatomic structures due to poor image quality; 2 = identifiable structures identifiable but no assembled details, resulting in insufficient image quality; 3 = fully assembled anatomic structures still in all parts and acceptable image quality; 4 = clear delineation of structure and good image quality; and 5 = very good delineation of structure and excellent image quality. FBP, filtered back projection; A_n, n^th level of ADMIRE (advanced modeled iterative reconstruction); HU, Hounsfield units; ROI, region of interest; UHR, ultra-high resolution; CT, computed tomography.

Table 3.

Subjective image analysis of anatomic structures of the temporal bone, adapted from Nauer et al. (6)

n	Anatomic structure	FBP	A1	A2	A3	A4	A5	Sum
1	Cochlea	3.86	4.23	4.39	4.46	4.35	4.14	4.24
2	Cochlear patency	3.54	3.76	4.18	4.58	4.46	4.11	4.11
3	Spiral osseous lamina	2.75	2.88	3.05	3.34	3.21	2.97	3.03
4	Modiolus	3.33	3.56	3.65	3.90	3.78	3.62	3.64
5	Vestibular labyrinth	3.58	3.86	3.98	4.13	4.02	3.87	3.91
6	Vestibular aqueduct	3.13	3.50	3.60	3.91	3.90	3.69	3.62
7	Cochlear aqueduct	2.93	3.04	3.12	3.41	3.41	3.20	3.18
8	Cochlear nerve canal	3.36	3.49	3.58	3.74	3.79	3.68	3.60
9	Bill’s bar	3.41	3.58	3.63	3.79	3.84	3.63	3.65
10	Internal auditory canal	4.32	4.55	4.63	4.77	4.51	4.32	4.52
11	Facial nerve canal, cochlear segment	3.37	3.70	3.82	4.06	3.95	3.70	3.77
12	Facial nerve canal, tympanic segment	3.48	3.61	3.86	4.23	4.19	3.80	3.86
13	Facial nerve canal, mastoid segment	3.48	3.76	3.93	4.17	4.07	3.88	3.88
14	Middle ear cavity	4.34	4.34	4.39	4.41	4.41	4.40	4.38
15	Ossicle chain, Malleus	3.89	3.95	3.97	4.02	4.01	3.97	3.97
16	Ossicle chain, Incus	3.95	3.95	3.97	4.02	4.01	3.97	3.98
17	Ossicle chain, Stapes	3.63	3.80	4.01	4.10	4.09	4.04	3.95
18	Oval window	3.15	3.16	3.48	3.68	3.56	3.43	3.41
19	Round window	3.35	3.57	3.67	3.94	3.92	3.87	3.72
20	Round window niche	3.65	3.78	3.92	4.19	4.08	3.92	3.92
21	Internal carotid artery canal	4.53	4.62	4.63	4.73	4.70	4.71	4.65
22	Jugular foramen	4.45	4.77	4.77	4.84	4.79	4.53	4.69
23	Mastoid	4.19	4.33	4.34	4.54	4.55	4.40	4.39
Sum	3.64	3.82	3.94	4.13	4.07	3.91

Subjective image quality scores of FBP and ADMIRE reconstructions in UHR CT of the temporal bone.

FBP, filtered back projection; A_n, n^th level of ADMIRE (advanced modeled iterative reconstruction).

Estimate of potential dose reduction

Using CTDI_vol and image noise values from Table 2, a potential dose reduction of 40% from 43.2 to 25.9 CDTI_vol was extrapolated. The confirmatory results from the supplementary phantom study are displayed in supplementary Table 1. CTDI_vol could be reduced about approximately 50% while maintaining image noise using FBP and ADMIRE A3 reconstructions.

Discussion

The aim of this study was to assess the potential of ADMIRE in temporal bone UHR CT. The ADMIRE algorithm can significantly reduce noise and increases image quality in temporal bone UHR CT as compared with FBP. In subjective analysis, an ADMIRE strength level of A3 performed best for delineating temporal bone substructures.

Our results from quantitative analyses demonstrated that the image noise was significantly lower in ADMIRE strength levels A2–A5 than in FBP. Additionally, SNR and CNR were significantly higher in ADMIRE strength levels A2–A5 than in FBP. So far, these findings confirm Meyer et al. (17), who compared the impact of acquisition mode and reconstruction algorithm between three different CT generations, and Leng et al. (19), who applied the sinogram-affirmed iterative (SAFIRE) reconstruction algorithm for comparison between different reconstruction scan modes in temporal bone CT. However, in a systematic comparison, we could demonstrate that effective and increasing noise reduction as well as improvement of image quality could be achieved by using ADMIRE strength levels of A2 or higher. Our findings so far confirm those of Solomon et al. (8), who demonstrated in phantom studies that “low-contrast detectability performance increased with […] ADMIRE strength” and ADMIRE allows for “substantial radiation dose reduction while preserving low-contrast detectability” (8). For temporal bone CT, we did not find corresponding reports in the literature.

In qualitative analyses, our data shows that an ADMIRE strength level of A3 performed best in delineation of temporal bone substructures. This reconstruction level may be an optimal balance between progressive noise reduction and small-structure identifiability (Tables 2 and 3; Fig. 5). We noticed a slight decrease in subjective image quality when applying an ADMIRE strength level A4 and above. This finding was surprising in view of the constantly ascending image quality and noise reduction that went along with increasing ADMIRE strength level in quantitative analysis. In detail, image quality decreased in ADMIRE levels A4 and A5 regarding high contrast structures, such as the internal auditory canal or the jugular foramen, whose bony margins seemed blurred or showed notches (Fig. 5e and f), as compared to FBP and ADMIRE A1–A3 (Fig. 5a–d). Additionally, small and subtle structures such as the cochlear spiral osseous lamina or the stapes (Suppl. Fig. 1) seemed washed out in ADMIRE levels A4 and A5 (Fig. 5k and l) in comparison with FBP and ADMIRE A1–A3 (Fig. 5g–k). We did not find any corresponding reports in the literature and consider our findings to be novel.

Fig. 5.

Subjective image quality in in FBP and ADMIRE in UHR CT of the temporal bone. Decrease of image quality in ADMIRE levels A4 and A5 (e, f, k, l) as compared to FBP (a, g) and ADMIRE levels A1–A3 (b–d, h–j). In high contrast structures, such as the internal auditory canal (a–f), the bony margins show notches (e, f). Small and subtle structures such as the cochlear spiral osseous lamina seem blurred (k, l). FBP, filtered back projection; ADMIRE, advanced modeled iterative reconstruction

Regarding clinical impact, we extrapolated a potential dose reduction of approximately 40% between FBP and ADMIRE A3 reconstructions, which performed best in qualitative image analysis. These findings could be confirmed in an additional phantom study, which yielded even better results. However, ADMIRE allows for substantial dose reduction while maintaining image quality.

This study is limited by its retrospective study design and lack of ADMIRE comparisons with other iterative reconstruction algorithms, such as SAFIRE or algorithms from other vendors.

In conclusion, the ADMIRE algorithm significantly reduced noise and subjectively and objectively increased image quality in temporal bone UHR CT. For effective noise reduction, an ADMIRE strength level of at least A2 is recommended. In qualitative analysis, ADMIRE strength level of A3 performed best in delineating temporal bone substructures. Using ADMIRE, substantial dose reduction is feasible for attaining the same image quality as compared to FBP.

Supplemental Material

Supplemental Material1 - Supplemental material for Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction

Supplemental material, Supplemental Material1 for Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction by Johann-Martin Hempel, Malte Niklas Bongers, Katharina Braun, Ulrike Ernemann and Georg Bier in Acta Radiologica

Supplemental Material

Supplemental Material2 - Supplemental material for Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction

Supplemental material, Supplemental Material2 for Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction by Johann-Martin Hempel, Malte Niklas Bongers, Katharina Braun, Ulrike Ernemann and Georg Bier in Acta Radiologica

Supplemental Material

Supplemental Material3 - Supplemental material for Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction

Supplemental material, Supplemental Material3 for Noise reduction and image quality in ultra-high resolution computed tomography of the temporal bone using advanced modeled iterative reconstruction by Johann-Martin Hempel, Malte Niklas Bongers, Katharina Braun, Ulrike Ernemann and Georg Bier in Acta Radiologica

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Casselman

Offeciers

Foer

B de

, et al. CT and MR imaging of congential abnormalities of the inner ear and internal auditory canal. Eur J Radiol 2001; 40:94–104.

Koch

Hamilton

Hudgins

et al. , editors. Diagnostic Imaging: Head and Neck. Philadelphia, PA: Elsevier; 2017.

Young

Ryan

Young

NM.

Preoperative imaging of sensorineural hearing loss in pediatric candidates for cochlear implantation.

Radiographics 2014; 34:E133–49.

Fishman

AJ.

Imaging and anatomy for cochlear implants.

Otolaryngol Clin North Am 2012; 45:1–24.

Rutherford

Lerer

Schoem

, et al. Evaluation of pediatric sensorineural hearing loss: A survey of pediatric otolaryngologists. Ann Otol Rhinol Laryngol 2011; 120:674–681.

NIH - National Cancer Institute. Radiation Risks and Pediatric Computed Tomography: A Guide for Health Care Providers. Available at: https://www.cancer.gov/about-cancer/causes-prevention/risk/radiation/pediatric-ct-scans (last accessed 12 Sep 2017).

Nauer

Rieke

Zubler

, et al. Low-dose temporal bone CT in infants and young children: effective dose and image quality. Am J Neuroradiol 2011; 32:1375–1380.

Solomon

Mileto

Ramirez-Giraldo

, et al. Diagnostic performance of an advanced modeled iterative reconstruction algorithm for low-contrast detectability with a third-generation dual-source multidetector CT scanner: potential for radiation dose reduction in a multireader study. Radiology 2015; 275:735–745.

Beister

Kolditz

Kalender

WA.

Iterative reconstruction methods in X-ray CT.

Phys Med 2012; 28:94–108.

10.

Thibault

J-B

Sauer

Bouman

, et al. A three-dimensional statistical approach to improved image quality for multislice helical CT. Med Phys 2007; 34:4526–4544.

11.

Pickhardt

Lubner

Kim

, et al. Abdominal CT with model-based iterative reconstruction (MBIR): initial results of a prospective trial comparing ultralow-dose with standard-dose imaging. Am J Roentgenol 2012; 199:1266–1274.

12.

Elm

E von

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. J Clin Epidemiol 2008; 61:344–349.

13.

Flohr

Stierstorfer

Suss

, et al. Novel ultrahigh resolution data acquisition and image reconstruction for multi-detector row CT. Med Phys 2007; 34:1712–1723.

14.

Siemens Healthcare. ADMIRE – Advanced Modeled Iterative Reconstruction. Available at: https://www.healthcare.siemens.de/computed-tomography/technologies-innovations/admire (last accessed 13 Sep 2017).

15.

Wang

, et al. Improving depiction of temporal bone anatomy with low-radiation dose CT by an integrated circuit detector in pediatric patients: a preliminary study. Medicine 2014; 93:e325.

16.

Niu

Olszewski

Zhang

, et al. Experimental study and optimization of scan parameters that influence radiation dose in temporal bone high-resolution multidetector row CT. Am J Neuroradiol 2011; 32:1783–1788.

17.

Meyer

Haubenreisser

Raupach

, et al. Initial results of a new generation dual source CT system using only an in-plane comb filter for ultra-high resolution temporal bone imaging. Eur Radiol 2015; 25:178–185.

18.

Cicchetti

DV.

Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology.

Psychol Assessment 1994; 6:284–290.

19.

Leng

Diehn

Lane

, et al. Temporal bone CT: Improved image quality and potential for decreased radiation dose using an ultra-high-resolution scan mode with an iterative reconstruction algorithm. Am J Neuroradiol 2015; 36:1599–1603.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.96 MB

1.47 MB

0.06 MB