We used data obtained by Caris Life Sciences, to evaluate the benefits of tailoring
treatments for a breast carcinoma cohort by using tumor molecular profiles to inform
decisions. Data for 92 breast cancer patients from the commercial Caris Molecular
Intelligence database was retrospectively divided into two groups, so that the first
always followed treatment recommendations, whereas in the second group all
patients received at least one drug after profiling that was predicted to lack benefit.
The biomarker and drug associations were based on tests including fluorescent
in
situ
hybridization and DNA sequencing, although immunohistochemistry was the main
test used.
Patients whose drugs matched those recommended according to their tumor
profile had an average overall survival of 667 days, compared to 510 days for patients
that did not (P=0.0316). In the matched treatment group, 26% of patients were
deceased by the last time of monitoring, whereas this was 41% in the unmatched
group (P=0.1257). We therefore confirm the ability of tumor molecular profiling to
improve survival of breast cancer patients. Immunohistochemistry biomarkers for
the androgen, estrogen and progesterone receptors were found to be prognostic for
survival.