Table 2.

Inverse probability weighting-adjusted Cox regression for invasive disease-free survival (IDFS)

	Biomarker-assessment study (N = 302)					Biomarker-discovery study (N = 394)
		Unadjusted		IPW adjusted			Unadjusted		IPW adjusted
All patients	n/N	HR(95%CI)	p value	HR (95% CI)	p value	n/N	HR (95% CI)	p value	HR (95% CI)	p value
		Trend-test	0.013*	Trend-test	0.004*		Trend-test	0.007*	Trend-test	0.190
Low CPE	17/101	Reference		Reference		25/132	Reference		Reference
Intermediate CPE	15/100	0.84 (0.42-1.69)	0.63	0.84 (0.37-2.00)	0.69	15/132	0.55 (0.29-1.05)	0.070	0.66 (0.28-1.36)	0.266
High CPE	5/101	0.25 (0.09-0.68)	0.007*	0.27 (0.05-0.68)	0.004*	9/130	0.33 (0.15-0.70)	0.004*	0.59 (0.20-1.33)	0.189
Patients receiving endocrine therapy	n/N	HR (95% CI)	p value	HR (95% CI)	p value	n/N	HR (95% CI)	p value	HR (95% CI)	p value
		Trend-test	0.016*	Trend-test	0.006*		Trend-test	0.042*	Trend-test	0.022*
Low CPE	16/94	Reference		Reference		11/56	Reference		Reference
Intermediate CPE	15/93	0.92 (0.45-1.86)	0.808	0.97 (0.42-2.16)	0.951	7/56	0.67 (0.26-1.72)	0.404	0.91 (0.22-2.62)	0.79
High CPE	5/94	0.26 (0.09-0.72)	0.009*	0.26 (0.05-0.71)	0.009*	2/56	0.18 (0.04-0.79)	0.024*	0.22 (0.00-0.79)	0.019*

Contralateral parenchymal enhancement (CPE) was split on tertiles in low, intermediate and high CPE. Differences between these CPE groups were adjusted using inverse probability weighting (IPW) for age at diagnosis, histological grade, largest tumour diameter, progesterone receptor status, axillary lymph node status and administration of endocrine and chemotherapy

In each analysis, n represents the number of patients that had an event and N the number of patients in the group analysed

Pooled hazard ratios (HR), 95% confidence intervals (CI) and p values are reported

Results for the biomarker-discovery study are displayed on the right for comparison

*p < 0.05, significant