Table 3.

The effect of the choice of the parameters 1 − α, q(ρ) and r on m, the number of repetitions required for calibration, and on the ED sample size estimates n₁, n₂, n₃, or n for a helical chest protocol on the 320-row scanner with parameter set a = 0.035, b = 1.180, c = 0.002, e = 27.7 mV, β = 1/3, and σ_β = 0.05% β.

				Three-region approach			Single-region approach
	Precision		Calibration	Region 1	Region 2	Region 3	Sample
Confidence	of ED	CV of $\widehat{\beta }$	sample size	sample size	sample size	sample size	size
1− α (%)	q (%)	r < ρ	m	n1	n2	n3	n
95	5	0.01 < 0.026	25	0.6	3.3	0.1	2.8
95	5	0.024 < 0.026	4.3	3.3	18.7	0.8	15.9
95	10	0.024 < 0.051	4.3	0.2	0.9	0.04	0.7
95	10	0.05 < 0.051	1	3	17	0.8	14.4
90	5	0.02 < 0.031	6.3	0.5	3.1	0.1	3
90	5	0.03 < 0.031	2.8	4.9	28.1	1.3	23.9
90	10	0.024 < 0.061	4.3	0.1	0.6	0.02	0.5
90	10	0.06 < 0.061	0.7	2.5	14.2	0.6	12.1