Variability in Heart-to-Mediastinum Ratio from Planar 123I-MIBG Images of a Thorax Phantom for 6 Common γ-Camera Models

Article Figures & Data

Download figure
Open in new tab
Download powerpoint
FIGURE 1.
Single-chamber ¹²³I-MIBG thorax phantom. Photograph of unfilled phantom (A) and example planar acquisition after filling with ¹²³I (B). Polygonal heart and square mediastinum ROIs of standardized image analysis are shown.
Download figure
Open in new tab
Download powerpoint
FIGURE 2.
Box-and-whisker plot of head 1 H/M ratio. Bottom and top of box are first and third quartiles, respectively. Horizontal line in box is median. + in box represents mean. Ends of the whiskers are minimum and maximum. Cam1 = GE Healthcare Millennium MG/LEHR (n = 10); Cam2 = GE Healthcare Discovery/LEHR (n = 8); Cam3 = GE Healthcare Infinia/LEHR (n = 10); Cam4 = Philips Brightview/LEHR (n = 2); Cam5 = Philips Cardio 60/LEHR (n = 2); Cam6 = Philips Cardio MD/LEHR (n = 4); Cam7 = Philips Vertex/VXGP (vertex general purpose) or VXHR (vertex high resolution) (n = 8); Cam8 = Siemens e.cam/LEHR (n = 8); Cam9 = Siemens Symbia/LEHR (n = 8). VXGP = vertex general purpose; VXHR = vertex high resolution.
Download figure
Open in new tab
Download powerpoint
FIGURE 3.
Bland–Altman plot for reader 1. Mean H/M (i.e., values for both heads) for each camera is plotted against difference between head 1 H/M and head 2 H/M. Results from reader 2 were similar to those shown here. VXGP = vertex general purpose; VXHR = vertex high resolution.

TABLE 1

LEHR Collimator Specifications for Studying γ-Cameras

Manufacturer	Model	LEHR collimator type	Mean septal thickness (mm)
GE Healthcare	Millennium MG	Cast	0.20
GE Healthcare	Infinia and Discovery NM series	Foil	0.20
Siemens	e.cam and Symbia	Foil	0.16
ADAC/Philips	Vertex	Foil	0.152 (26)
Philips	Brightview	Foil	0.156 (27)

TABLE 2

Summary of Camera Models, Collimator Types, and Sample Size Per Model Included in Study

TABLE 3

Analysis of Intracamera Variability in H/M Ratio

Measurement	GE Healthcare, Millennium MG/LEHR (n = 10)	GE Healthcare, Discovery/LEHR (n = 8)	GE Healthcare, Infinia/LEHR (n = 10)	Philips, Brightview/LEHR (n = 2)	Philips, Cardio 60/LEHR (n = 2)	Philips, Cardio MD/LEHR (n = 4)	Philips, Vertex/VXGP or VXHR (n = 8)	Siemens, e.cam/LEHR (n = 8)	Siemens, Symbia/LEHR (n = 8)
Mean ± SD	1.648 ± 0.0776	1.624 ± 0.0849	1.607 ± 0.1193	1.665 (0.0257)	1.677 ± 0.0302	1.342 ± 0.0262	1.611 ± 0.0866	1.566 ± 0.1118	1.635 ± 0.0455
Median	1.65 (range, 1.52–1.76)	1.62 (range, 1.51–1.75)	1.59 (range, 1.39–1.78)	1.66 (range, 1.65–1.68)	1.68 (range, 1.66–1.70)	1.34 (range, 1.31–1.37)	1.63 (range, 1.49–1.77)	1.53 (range, 1.41–1.73)	1.63 (range, 1.55–1.69)
95% CI	1.592–1.703	1.553–1.695	1.521–1.692	1.434–1.895	1.406–1.949	1.300–1.384	1.539–1.683	1.472–1.659	1.597–1.673

Analysis is based on average of H/M ratios from 2 independent readers on camera head 1 for each camera model.
n = sample size for each camera/collimator combination; VXGP = vertex general purpose; VXHR = vertex high resolution.

TABLE 4

Analysis of Intercamera Variability of H/M Ratio for Thorax Phantom

Camera/collimator	Sample size	H/M LSM ± SE^*	LSM difference ± SE^*	P^*
GE Healthcare, Millennium MG/LEHR	10	1.645 ± 0.0286
GE Healthcare, Discovery/LEHR	8	1.621 ± 0.0319	0.024 ± 0.0429	0.5804
GE Healthcare, Infinia/LEHR	10	1.609 ± 0.0286	0.036 ± 0.0404	0.3824
Philips, Vertex/VXGP or VXHR	8	1.604 ± 0.0319	0.040 ± 0.0429	0.3496
Siemens, e.cam/LEHR	8	1.564 ± 0.0319	0.081 ± 0.0429	0.0655
Siemens, Symbia/LEHR	8	1.631 ± 0.0319	0.014 ± 0.0429	0.7496

↵* Obtained from mixed model with repeated measures on H/M ratios for first head tested as dependent variable and camera/collimator combination as independent variable. Each camera model is compared with reference camera Millennium MG (GE Healthcare).
LSM = least-squares mean; VXGP = vertex general purpose; VXHR = vertex high resolution.
Only camera/collimator categories with a sample size of at least 8 are included in model, as specified in statistical analysis plan.