. 2018 Feb 19;45(3):1178–1190. doi: 10.1002/mp.12763

Table 6.

Evaluation of the algorithm's ability to separate BI‐RADS density levels

View	BI‐RADS density level pairs	Proposed algorithm		LIBRA
View	BI‐RADS density level pairs	Mean difference and 95% CI	P‐value	Mean difference and 95% CI	P‐value
CC	1–2	−0.02 [−0.07, 0.02]	0.54	−0.05 [−0.15, 0.05]	0.64
	2–3	−0.12 [−0.16, −0.09]	<0.0001^a	−0.11 [−0.2, −0.02]	0.0067
	3–4	−0.14 [−0.18, −0.11]	<0.0001^a	−0.17 [−0.26, −0.09]	<0.0001^a
MLO	1–2	−0.12 [−0.17, −0.06]	<0.0001^a	−0.04 [−0.13, 0.05]	0.72
	2–3	−0.13 [−0.18, −0.09]	<0.0001^a	−0.16 [−0.23, −0.08]	<0.0001^a
	3–4	−0.08 [−0.13, −0.04]	<0.0001^a	−0.22 [−0.29, −0.14]	<0.0001^a
Averaged	1–2	−0.07 [−0.11, −0.03]	0.0001^a	−0.04 [−0.13, 0.04]	0.59
	2–3	−0.13 [−0.17, −0.09]	<0.0001^a	−0.14 [−0.21, −0.06]	<0.0001^a
	3–4	−0.11 [−0.15, −0.08]	<0.0001^a	−0.2 [−0.27, −0.12]	<0.0001^a

Statistically significant with significance level of 0.0014 with Bonferroni correction. The mean difference for all other pairs, 1–3, 1–4, and 2–4 were statistically significant with P < 0.0001.