Skip to main content
. 2011 Nov 3;105(11):1669–1675. doi: 10.1038/bjc.2011.466

Figure 4.

Figure 4

Scatter plot of log probability of malignancy (y axis) vs mammographic extent of microcalcifications. Because the relationship between size and log of the probability of malignancy is not linear, the appropriate statistical analysis was for size to be categorised as ⩽15 mm or >15 mm rather than as a continuous variable.