Table 5.

Local recurrence events in studies of preoperative MRI use and breast cancer outcomes^a.

	Number of events	Sample size	Median follow-up (month)b	IBTR (MRI vs. no MRI)	HR (95% CI) or P for event rate
Vapiwala et al.22c	49	755	166	8% vs. 8%	0.98 (0.52, 1.87)
Hill et al.19	78	1396	86 (MRI) and 77 (no MRI)	8 vs. 4%
Choi et al.21	66	1598	Over 60	4.0 vs. 4.3%	0.95 (0.58, 1.54)
Gervais et al.37	17d	470	85 (MRI) and 106 (no MRI)	1.6 vs. 4.2%	2.0 (0.45, 9.00)
Yi et al.15	27	936	74	1.7 vs. 4.1%	0.3 (0.1, 0.9)
Pilewskie et al.20 DCIS only	184	2321	59	8.5 vs. 7.2%	1.36 (0.78, 2.39)
Sung et al.16	22	348	96	5 vs. 9%	P = 0.33
Miller et al.44	9	414	29 (MRI) and 45 (no MRI)		P = 0.13
Hwang et al.17	11	472	54	1.8 vs. 2.5%	1.7 (0.2, 11.8)
Turnbull et al.9e	52	1569	24	1.5 vs. 1.5%	Not reported
Fischer et al.14f	10	346	41	1.2 vs. 6.5%	P = 0.001

^aThe meta-analysis of Houssami et al. is not included in the table, since the component studies are included.

^bFollow-up is presented separately by MRI status when so reported by authors.

^cUpdated analysis of Solin et al.¹⁸.

^dNumber of events estimated from reported rates at 8 years.

^eReported in Houssami meta-analysis.

^fOutcomes not adjusted for tumor or therapy characteristics.