Would you offer post-operative SRS/SBRT in the case of a resected brain metastasis from metastatic NSCLC with an EGFR mutation?

Absolutely. For symptomatic brain metastases, I would not rely on the efficacy of systemic therapy alone either in the intact or post-operative setting. CNS objective response rates are around 60% for osimertinib with DCR around 90% which is quite good. However, intracranial complete response rates are only around 10-15%. If your patient has a GTR of brain metastasis, it seems risky to compromise local control and potential survival by banking on control of residual microscopic disease with tumoristatic systemic therapy. GTR is associated with improved survival vs. STR and I would argue that the same compromise in survival would be seen in observing a GTR with systemic therapy and subsequently having to treat with salvage SRS to post-operative bed with gross disease recurrence. I would always rather be treating microscopic than macroscopic disease and given that SRS comes with a very reasonable safety profile, why take the risk here of observing? 

Slightly different scenario, I do routinely observe (short interval MRI) small asymptomatic brain metastases in these driver mutation populations as a fair amount are going to achieve excellent intracranial responses and we can either spare CNS radiation altogether or reserve if progression down the road. 

I agree. One reason is that after surgical resection, the vascularization of the surgical cavity is poor and drug penetration may be suboptimal.