Found small mismatch for a missing data.

In a column for "BI-RADS assessment" I found value which is not represented either by "?" mark or by discrete values from 1 to 5. It is "55", probably a mistype during the data gathering process.

In the tutorial there is no comment on this. So if we do a fit of our models the lower score is obtained for such case,
given same model parameters. If we assume that "55" is mistype of "5" and replace this value, more higher score could be obtained.

So this problem could be solved in two ways:

Change value from "55" either to "5" or "?" mark.
Include this into the future tutorials and ask to compare models if one of such outliers is found.