What brachial plexus dose constraints do you apply when treating conventionally fractionated NSCLC?

Yes. I am presuming you are discussing treating definitively with CRT for locally advanced NSCLC.

Since RTOG 0617, most people that treat NSCLC prescribe to a dose of 60-66 Gy in 30-33 fx, as this is a feared toxicity that is very painful and decreases QOL. We try to keep the brachial plexus dose to below RX dose. Even at 66 Gy, the rates of BP are exceedingly low. Thus, this data appears to make sense - it seems the risk is 10x higher when you exceed 69 Gy. Unless you are dose escalating based on a study or using older techniques, <69 Gy should not be a difficult constraint to meet.

I am much more interested in the auto-contouring of the brachial plexus that is described in the paper you have linked to.

Agreed with Dr. @Dr. First Last. Of note, if you are treating a superior sulcus tumor near the brachial plexus, you might consider the Intergroup 0160 regimen of pre-op chemoRT followed by surgery. In this trial, they used a dose of 45 Gy/25# with cisplatin/etoposide, but one could go higher to a definitive dose if there is some question about the patient's surgical candidacy.

NCCN lists median brachial plexus dose constraint as < 69Gy for conventionally fractionated 30-35Fx regimen. (When I inquired a while back, they confirmed this wasn't a typo).