Radiation Dose Reduction in Paranasal Sinus CT Using Model-Based Iterative Reconstruction

J.M. Hoxworth; D. Lal; G.P. Fletcher; A.C. Patel; M. He; R.G. Paden; A.K. Hara

doi:10.3174/ajnr.A3749

Article Figures & Data

Figures

Tables

Download figure
Open in new tab
Download powerpoint
Fig 1.
Graphic representation of mean image noise for the pons, globe, masseter, and air within the maxillary sinus for each of the tested CT series: soft-tissue algorithm, filtered back-projection, low dose; bone algorithm, filtered back-projection, low dose; Veo model-based iterative reconstruction, low dose; soft-tissue algorithm, filtered back-projection, standard dose; and bone algorithm, filtered back-projection, standard dose.
Download figure
Open in new tab
Download powerpoint
Fig 2.
Graphic representation of mean image quality for the brain, cribriform plate, globe, inferior rectus muscle, lamina papyracea, middle turbinate, nasal septum, and optic nerve for each of the tested CT series: soft-tissue algorithm, filtered back-projection, low dose; bone algorithm, filtered back-projection, low dose; Veo model-based iterative reconstruction, low dose; soft-tissue algorithm, filtered back-projection, standard dose; and bone algorithm, filtered back-projection, standard dose.
Download figure
Open in new tab
Download powerpoint
Fig 3.
Axial noncontrast sinus CT (window level, 40 HU; window width, 400 HU) through the level of the orbits for the following imaging series: soft-tissue algorithm, filtered back-projection, standard dose (A); bone algorithm, filtered back-projection, standard dose (B); soft-tissue algorithm, filtered back-projection, low dose (C); bone algorithm, filtered back-projection, low dose (D); and Veo model-based iterative reconstruction, low dose (E). The decreased image noise secondary to Veo allows improved visualization of orbital and intracranial soft-tissue structures relative to the other low-dose protocols but not to the level of the standard-dose sinus CT.
Download figure
Open in new tab
Download powerpoint
Fig 4.
Coronal noncontrast sinus CT (window level, 800 HU; window width, 3500 HU) through the level of the maxillary sinuses for the following imaging series: soft-tissue algorithm, filtered back-projection, standard dose (A); bone algorithm, filtered back-projection, standard dose (B); soft-tissue algorithm, filtered back-projection, low dose (C); bone algorithm, filtered back-projection, low dose (D); and Veo model-based iterative reconstruction, low dose (E). The noise reduction achieved with Veo results in loss of edge enhancement, making thin bones such as the nasal septum (white arrow) and left lamina papyracea (white arrowhead) difficult to see.

Tables

Figures

View popup

Table 1:

Image-noise measurement^a

	Pons (n = 20)	Globe (n = 20)	Masseter (n = 20)	Maxillary Sinus (n = 20)	Total (n = 80)
SOFT FBP LD	74.9 (10.60)^b	55.6 (8.79)^b	52.6 (8.79)^b	45.3 (13.16)^b	57.1 (15.06)^b
BONE FBP LD	278.8 (36.73)^b	173.0 (19.66)^b	175.3 (27.03)^b	105.5 (15.64^{)^b}	183.1 (67.38)^b
VEO LD	18.9 (2.98)	12.4 (1.37)	14.3 (2.16)	12.3 (2.11)	14.5 (3.49)
SOFT FBP SD	24.7 (5.14)^b	19.5 (3.06)^b	19.1 (3.35)^b	17.5 (2.48)^b	20.2 (4.48)^b
BONE FBP SD	97.4 (14.47)^b	66.9 (7.78)^b	66.3 (8.96)^b	50.1 (6.42)^b	70.2 (19.73)^b

↵a Values are expressed as mean (SD) of image noise measured in Hounsfield units.
↵b Statistically significant difference (P < .001) in paired comparison with VEO LD.

View popup

Table 2:

Image-quality assessment^a

	Brain (n = 40)	Globe (n = 40)	Inferior Rectus Muscle (n = 40)	Optic Nerve (n = 40)	Cribriform Plate (n = 40)	Lamina Papyracea (n = 40)	Middle Turbinate (n = 40)	Nasal Septum (n = 40)	Total (n = 320)
SOFT FBP LD	1.5 (0.51)^b	1.6 (0.55)^b	1.9 (0.43)^b	1.8 (0.52)^b	3.3 (0.47)	3.1 (0.38)^b	3.3 (0.44)	3.1 (0.38)	2.4 (0.90)^b
BONE FBP LD	1.0 (0.00)^b	1.0 (0.16)^b	1.0 (0.16)^b	1.1 (0.27)^b	3.7 (0.62)^b	3.7 (0.56)^b	3.4 (0.58)	3.6 (0.59)^b	2.3 (1.35)^b
VEO LD	2.8 (0.41)	2.9 (0.35)	3.1 (0.32)	3.1 (0.32)	3.2 (0.48)	2.7 (0.53)	3.2 (0.53)	3.0 (0.45)	3.0 (0.46)
SOFT FBP SD	3.4 (0.55)^b	4.0 (0.42)^b	4.0 (0.58)^b	4.1 (0.40)^b	4.0 (0.51)^b	3.9 (0.47)^b	4.0 (0.55)^b	3.9 (0.56)^b	3.9 (0.54)^b
BONE FBP SD	2.0 (0.28)^b	2.5 (0.51)^b	3.0 (0.53)	3.1 (0.45)	4.7 (0.47)^b	4.9 (0.33)^b	4.7 (0.46)^b	4.9 (0.36)^b	3.7 (1.19)^b

↵a Values are expressed as mean (SD). Image-quality scale ranges from 1 (unacceptable noise, nondiagnostic) to 5 (excellent image quality, best diagnostic value).
↵b Statistically significant difference (P < .001) in paired comparison with VEO LD.