Skip to main content
. 2017 Mar 30;3(8):1060–1068. doi: 10.1001/jamaoncol.2016.6948

Table 2. Cox Proportional Hazards Regression Multivariable Analysis of Ipsilateral Breast Tumor Recurrence.

Patient Characteristic HR (95% CI) P Value
Age, y
<50 1 [Reference] NA
≥50 0.54 (0.41-0.71) <.01
DCIS grade
I 1 [Reference] NA
II or III 1.59 (1.02-2.46) .04
Unknown 1.56 (0.93-2.59) .09
Tumor size (continuous variable) 1.15 (1.02-1.29) .03
Margin status
Negative 1 [Reference] NA
Positivea 1.72 (0.97-3.05) .06
Unknown 1.65 (1.13-2.41) .01
Comedo necrosis
No 1 [Reference] NA
Yes 1.12 (0.78-1.62) .53
Unknown 0.51 (0.34-0.77) <.01
Estrogen receptor status
Negative 1 [Reference] NA
Positive 0.7 (0.39-1.25) .23
Unknown 0.85 (0.52-1.38) .51
Tamoxifen citrate use
No 1 [Reference] NA
Yes 0.56 (0.36-0.87) .01
Unknown 0.93 (0.53-1.63) .81
RT boost
No 1 [Reference] NA
Yes 0.68 (0.5-0.91) .01

Abbreviations: DCIS, ductal carcinoma in situ; HR, hazard ratio; NA, not applicable; RT, radiotherapy.

a

Indicates the National Surgical Adjuvant Breast Project B-17 definition of ink on tumor or 0-mm margin width. The log-rank ratio tested the interaction between age and boost (P = .46), which was not statistically significant.