Table 2.

Adjusted estimates of relative change¹ in percent BD by MRI at 6 and 12 months in AI Only and Sulindac Groups (All and stratified on baseline BD ≤ or >25%).

Relative % Change in BD (95% CI)
AI Only (n=46)
		N4	6- month	p-value	N5	12-month	p-value
All2		39	+1.6 (−2.8, +6.1)	0.486	39	−3.6 (−8.9, +2.0)	0.201
Adjust baseline BD category3	BD ≤ 25%	29	+2.3 (−2.8, 7.7)	0.388	28	−4.0 (−9.8, +2.2)	0.202
	BD > 25%	10	+1.0 (−7.4, +10.1)	0.829	11	−2.8 (−12.2, +7.6)	0.582
Sulindac plus AI (n=50)6
		N4	6- month	p-value	N5	12-month	p-value
All2		44	−4.2 (−8.1, −0.1)	0.046	43	−9.8 (−14.6, −4.7)	<0.001
Adjust baseline BD category3	BD ≤ 25%	34	−3.8 (−8.2, +0.9)	0.109	34	−8.1 (−13.2, −2.7)	0.004
	BD >25%	9	−6.2 (−14.5, +3.0)	0.178	8	−14.6 (−24.1, −3.9)	0.009

¹Estimated relative change in BD at 12 months is defined as BD at 12 months – BD at baseline/BD at baseline. Summary values are back-transformed estimates and 95% CI.

²Analysis based on pooled data from both groups (n=96) and P-values were based on t-test from linear mixed model adjusted for log-transformed baseline BD, time on aromatase inhibitor, BMI at each time point and study site.

³Analysis based on pooled data from both groups (n=96) and P-values were based on t-test from linear mixed model adjusted for baseline BD category, time on aromatase inhibitor, BMI at each time point and study site.

⁴Sample size with baseline and 6-month percent BD by MRI

⁵Sample size with baseline and 12-month percent BD by MRI

⁶Baseline BD was unavailable for 2 subjects who started sulindac (one image artifact, one image lost during IT security change on clinical instrument).