the response rate in triple negative BC was 19% compared to 0% for the investigator’s choice (IC) arm.
Is there a biological explanation why CDX-011 is active in triple-negative breast cancer? Has this been discussed? Or why this response rate did not translate to an OS advantage? Thanks.
About the exceptionally low p value for "Triple Negative and High GPNMB Expression": IMO, one of the major differences between "High GPNMB Expression" and "Triple Negative and High GPNMB Expression" KM graphs is just the removal of the 3 patients who lived 11.4, 12.5 and 21m because they were not triple negative. See the last 2 graphs in
This is a little bit difficult to realize because the graphs are on a different time scale. However, thanks to PDF (which scales nicely), if you zoom, you can match the patients in two graphs. This removal alone should explain the very low p value for the "Triple Negative and High GPNMB Expression" subset.
As for an explanation, it seems like something else (IC drug?) worked for these 3 patients and the removal of these patients from the last subset ensured that patients who lived exceptionally long are not included in the last graph (no competing therapies for TNBC)