TMB - I'd interpret it in that way that Perlmutter looks for a biomarker when to deploy pembro, which makes more sense when the TME is inflamed. And here not all mutations are recognised as "not normal/out of equilibrium" by the immune system and drive an inflammatory response. So I believe there is a correlation between TMB and immune response, but it's not such a mechanistic one as PD-L1 expression.
As well, some mutations help to hide from the immune system, see here for an article I like to quote (here: mutation leads to deficiencies in MHC-I presentation of Ag by tumor)
https://www.nature.com/articles/s41467-017-01062-w "we find that B2M LOH is enriched threefold in non-responders (~30%) compared to responders (~10%) and associated with poorer overall survival. Loss of both copies of B2M is found only in non-responders. B2M loss is likely a common mechanism of resistance to therapies targeting CTLA4 or PD1."