Oncotype and Her2 status

I've posted the following on two other threads so far because the issue keeps coming up.  The OncotypeDX score is calculated from the expression of 16 different genes and 5 reference genes so a high or low score isn't the result of any one thing in particular, rather the combined scores of ER-related expression (the higher the better) and invasion, HER2-related and proliferation (the lower the better) groups.  The post below is combined from the two other threads - one asking about Ki-67 and the other about PR...

Below is my post from the progesterone thread. You can see how genomic health uses Ki-67 as one of the 16 calculations. One thing I think is critical to remember when considering issues like this is that the pathologist (who would be the one to assign a Ki-67 measurement if done as well as ER, PR and HER2) is using a different tumor sample than genomic health would be using. Cancer is a heterogenous disease - it's not impossible to have very different results within the same tumor. Back to the old but original reason for this thread, the OP was concerned because her pathology didn't mention Ki-67 and that made her feel that she wasn't given all of the information she needed to make the chemo/no chemo decision. I too felt shorted because my pathology didn't mention Ki-67 but eventually realized that the oncotype test did measure it - as well as 4 other proliferation genes. Like the OP, I had to be convinced not to do chemo because I wanted to do everything possible to avoid a recurrence. Once I understood that chemo doesn't guarantee a cure and it has its own risks to consider, I agreed that the right plan for me was to forego it.

Here's my post...

the PR score is just one of the factors that goes into calculating the oncotype score. They break down the different genes into groups and multiply each group by a factor they somehow came up with to get the final score. PR is in the "ER related genes" group which is the only group to have a negative multiplier - the higher that group's score, the more it lowers the final result. So if you have low or absent PR expression, it will raise that part of the score. If the scores from the other groups are low, you can still end up with a low score of course - PR is just one piece of the puzzle.

The test was then validated using 675 node negative ER positive tamoxifen treated cases from NSABP B-14. In this analysis given classes of genes were given previously determined weighting factors and prognostic scores were calculated. The weightings would seem to make sense with proliferation, Her2 related, and invasion group genes increasing the score and ER related genes decreasing the score. The score was a continuous variable, but for the convenience of the presentation the patients were divided into 3 groups using cut points of 18 and 31.