Low-dose computed tomography (LDCT) Non-Small Cell Lung (NSCLC) screening is associated with high false-positive rates, leading to unnecessary expensive and invasive follow ups. There is a need for minimally invasive approaches to improve the accuracy of NSCLC diagnosis. In addition, NSCLC patients harboring sensitizing mutations in epidermal growth factor receptor EGFR (T790M, L578R) are treated with Osimertinib, a potent tyrosine kinase inhibitor (TKI). However, nearly all patients develop TKI resistance. The underlying mechanisms are not fully understood. Plasma extracellular vesicle (EV) and circulating microRNA (miRNA) have been proposed as biomarkers for cancer screening and to inform treatment decisions. However, the identification of highly sensitive and broadly predictive core miRNA signatures remains a challenge. Also, how these systemic and diverse miRNAs impact cancer drug response is not well understood. Using an integrative approach, we examined plasma EV and circulating miRNA isolated from NSCLC patients versus screening controls with a similar risk profile. We found that combining EV (Hsa-miR-184, Let-7b-5p) and circulating (Hsa-miR-22-3p) miRNAs abundance robustly discriminates between NSCLC patients and high-risk cancer-free controls. Further, we found that Hsa-miR-22-3p, Hsa-miR-184, and Let-7b-5p functionally converge on WNT/βcatenin and mTOR/AKT signaling axes, known cancer therapy resistance signals. Targeting Hsa-miR-22-3p and Hsa-miR-184 desensitized EGFR-mutated (T790M, L578R) NSCLC cells to Osimertinib. These findings suggest that the expression levels of circulating hsa-miR-22-3p combined with EV hsa-miR-184 and Let-7b-5p levels potentially define a core biomarker signature for improving the accuracy of NSCLC diagnosis. Importantly, these biomarkers have the potential to enable prospective identification of patients who are at risk of responding poorly to Osimertinib alone but likely to benefit from Osimertinib/AKT blockade combination treatments.