How are you different than Tempus/BostonGene?

Tempus, BostonGene, Cofactor Genomics, and others have all taken the approach where they take RNA data from FFPE samples and have built algorithms based on the expression profiles. It is akin to measuring PDL1 at baseline, but in a fancier way that instead of looking at 1 molecule that defines the TME, looks at multiple. It does not work for the same reason that PDL1 doesn't work. Interestingly, neither Tempus nor BostonGene have published their sensitivity or specificity, despite having data on thousands of patients. Cofactor did publish their sens/spec, and they showed that in Head & Neck cancer, which has a low specificity for PDL1, their algorithm could improve spec from 0.2 to 0.7; as a consequence, their sensitivity dropped far below PDL1 in H&N. We are seeing incredibly high sens and spec in our approach, we think because of the way we measure the functional response ex vivo. Immunotherapy is much more complex than other drugs like targeted therapy, as it activates a system that relies on a hugely complex interplay of not a single receptor or molecule, but many orders of magnitude more factors, which interact dynamically, not in a static way. It's like trying to predict the weather two years out from today be measuring the temperature and air pressure everywhere in your city. You can measure it accurately, but the interplay is so complex that you could never accurately predict so far out.