Additional file 1 : Table S3 shows that fitting different models to different modes of detection made a highly significant improvement in fit compared with fitting one model to the combined data - which assumes that estimates are the same for both screen-detected and interval cancers - for all models (all P < 0.005).
← all excerpts
Predicting interval and screen-detected breast cancers from mammographic density defined by different brightness thresholds.
1
—
—