volume (MTV), and total lesion glycolysis (TLG). Percent changes in these variables from before to during or after treatment were also analyzed. Both the absolute values and changes in these variables were added to a clinical model (which included age, sex, race, tumor size, histology, disease stage, smoking history, receipt of induction chemotherapy, radiation technique, and total delivered dose) to assess added prediction power. Endpoints were recurrence within the planning target volume (PTV) and times to distant metastasis, disease-free survival, and OS, assessed by Cox regression analysis. Results: Median time to recurrence within the PTV was 12 months; time to distant metastasis, 9 months; disease-free survival, 9 months; and OS, 24 months. On the pretreatment PET/CT scans, SUVmax (hazard ratio [HR] 0.938, 95% confidence interval [CI] 0.887-0.991, PZ0.023) and SUVpeak (HR 0.927, 95% CI 0.870-0.988, PZ0.019) of the primary tumor correlated with distant metastasis in univariate analysis but not after adjustment for clinical factors in multivariate analysis (SUVmax: HR 0.954, 95% CI, 0.899-1.012, PZ0.117; SUVpeak: HR 0.944, 95% CI 0.881-0.987, PZ0.106). On the mid-treatment scans, MTV and TLG of lymph nodes were associated with time to local recurrence in both univariate and multivariate analyses after adjustment for clinical factors (MTV: HR 1.027, 95% CI 1.008-1.046, PZ0.005; TLG: HR 1.008, 95% CI 1.003-1.012, P<0.001). On the post-treatment PET/CT scans, no variables for primary tumor or lymph nodes, whether absolute values or percent changes were associated with clinical outcomes. Conclusion: Our results suggest that MTV and TLG in lymph nodes during chemoradiation therapy predicted local recurrence after concurrent chemoradiation therapy for locally advanced NSCLC. This information may help clinicians to modify treatment accordingly.