Skip to main content
. 2019 Jun 4;292(2):311–318. doi: 10.1148/radiol.2019182475

Table 4:

Logistic Regression Results Comparing Performance of Breast MRI Relative to Mammography Alone in Fully Adjusted Models

graphic file with name radiol.2019182475.tbl4.jpg

Note.—Data in parentheses are 95% confidence intervals. P values less than .05 were considered to indicate statistical significance. AJCC = American Joint Committee on Cancer, CI = confidence interval, OR = odds ratio, PPV1 = percentage of examinations with a positive end-of-day assessment that results in a second cancer diagnosis within 1 year, PPV3 = percentage of examinations with a positive final assessment and biopsy was known to be performed that results in a second cancer diagnosis within 1 year.

*Odds ratios adjusted for age at diagnosis, Breast Cancer Surveillance Consortium registry, AJCC stage, histologic findings, mode of detection, preoperative MRI, surgical procedure, radiation therapy, chemotherapy, estrogen receptor and progesterone receptor status (invasive only), grade, breast density, age at examination, year of examination, years since prior surveillance examination, and race/ethnicity.

Because of convergence issues with interval cancer rate and sensitivity, we used a reduced model, which did not adjust for surgical procedure, radiation, year of examination, or race/ethnicity.