Table 2.

Metal artifact reduction performances of tomosynthesis reconstruction algorithms. (dark artifact area).

ROI_Location_2
Variable	Difference	Standard Error	p	95% CI *
				Lower Limit	Upper Limit
CGpM-MAR vs. Original-FBP	−0.1726	0.02142	0.000	−0.2288	−0.1165
CGpM-MAR vs. DT-MAR	0.0836	0.02142	0.001	0.0274	0.1397
CGpM-MAR vs. CNNMAR	−0.1238	0.02142	0.000	−0.1800	−0.0677
Original FBP vs. DT-MAR	0.2562	0.02142	0.000	0.2001	0.3124
Original-FBP vs. CNNMAR	0.0488	0.02142	0.111	−0.0074	0.1049
Original-FBP vs. CGpM-MAR	0.1726	0.02142	0.000	0.1165	0.2288
DT-MAR vs. Original-FBP	−0.2562	0.02142	0.000	−0.3124	−0.2001
DT-MAR vs. CNNMAR	−0.2074	0.02142	0.000	−0.2636	−0.1513
DT-MAR vs. CGpM-MAR	−0.0836	0.02142	0.001	−0.1397	−0.0274
CNNMAR vs. Original-FBP	−0.0488	0.02142	0.111	−0.1049	0.0074
CNNMAR vs. DT-MAR	0.2074	0.02142	0.000	0.1513	0.2636
CNNMAR vs. CGpM-MAR	0.1238	0.02142	0.000	0.0677	0.1800
Source of variation	df *	sums of squares	mean square	F	p
Algorithm	3	0.979	0.326	59.273	0.000
Dose	2	0.154	0.077	13.960	0.000
Algorithm × Dose	6	0.351	0.058	10.618	0.000
Error	84	0.462	0.006	-	-

* CI: confidence interval; dependent variable: artificial index value. Tukey–Kramer test; p < 0.05 indicates a significant difference (without metal artifact reduction processing). * df: degree of freedom; dependent variable: artificial index value. Tukey–Kramer test; p < 0.05 indicates a significant difference (without metal artifact reduction processing).