Can you link to those articles you mention as showing relationship between TCR status and PD-1 inhibition prognosis?
Taking a guess on your question the reason could be that there is no well defined exhaustive set of TCR sequences related to cancer specific epitope and therefore it is hard to exclude an individual from PD-1 treatment on that piece of evidence.
On the other hand if tumor cells are expressing PD-L1 they are likely to respond to PD-1 inhibition and if they don't express PD-L1 at all, well then there is a clear cut case where PD-1 inhibition is useless.
From the other end of the spectrum, don't underestimate how conservative pharma/biotech are in their approaches. Rep-Seq of TCR is still a new technology in their eyes and therefore associated with significant risk and trouble (such as convincing FDA/EMA to use this instead of the well known immunohistochemistry that all the rest are using).