Background: Recommendation of adjuvant chemotherapy(ACT) is based on tumor size and lymph node invasion for early-stage surgically resected non-small cell lung cancer(NSCLC) patients. Specific criteria lack to define who will be benefit from ACT. Aim: To evaluate the reproducibility of a 12-gene panel to predict ACT benefit and prognosis for early-stage NSCLC patients. Methods: Formalin-fixed paraffin-embedded(FFPE) samples from surgically resected patients(n=139) were retrospectively included. RNA-based expression of 12 genes was evaluated by NanoString (ATP8A1, AURKA, C1orf116, COL4A3, DOCK9, HOPX, HSD17B6, IFT57, MBIP, NKX2-1, RRM2 and TTC37). Patients were stratified as low-risk(ACT non-benefit) and a high-risk (ACT-benefit), and overall(OS) and event-free survival(EFS) were assessed. Results: The RNA-based panel present a 0.8% failure rate. This panel successfully stratified patients(n=138) as low- and high-risk regardless of ACT treatment. Low-risk patients presented an improvement of 15 months in both OS and EFS compared with high-risk(pOS<0.0001, pEFS<0.001). Median survival was not reached. Regarding the predictive performance, ACT-treated high-risk(n=50) patients had an improvement of 5 months of OS(x?OS=22.6mos.) and EFS(x?EPFS=19.7mos.) compared with untreated patients(x?OS=17.5mos., x?EPFS=14.7mos.). Low-risk patients(n=88) had similar outcomes for ACT-treated and ACT-untreated patients for both OS and EFS. Conclusion: The 12-gene expression panel successfully stratify low- and high-risk patients, and may be a promising predictive tool for clinical management of early-stage NSCLC patients.