To compare the predicted response with observed response to treatment by measuring gross tumor volume-primary (GTVp) using onboard kilovoltage (kV) cone-beam computed tomography (CBCT), to analyze the serial tumor volumes during radiotherapy (RT) with serial tumor volumes during follow-up, and to identify the variables associated with survival outcomes. Between June 2017 and December 2019, 23 patients of histologically proven locally advanced nonsmall cell lung cancer (LA-NSCLC) received definitive chemoradiation. Serial kV-CBCT images X-ray volume imaging (XVI) were generated weekly for image guidance and were used to generate serial GTVp. Posttreatment follow-up images were used to generate follow-up GTVp. Relative volume regression (VR) during RT and relative response assessment (RA) during follow-up were defined from Avg Vol, of planning CT. The predicted progression model was generated from VR and analyzed against observed progression events. Regression-response model was generated to analyze VR against RA. The median XVI vol1, XVI vol2, and XVI vol3 were 78.123, 56.571, and 48.513 cc during the 2nd, 4th, and 6th weeks of RT, respectively. The median VR0 was 11.777% in the 2nd-week, VR1 was 20.959% in the 4th week, and 33.661% in the 6th week. The predicted responders and progression using the VR were similar to the observed response during the follow-up. The prediction of both RA0 and RA1 obtained from VR2 was statistically significant. Predication of RA0 from VR1 tended towards significant (P=0.084). VR2 was statistically significant in predicting RA2 (P = 0.04). The median progression-free survival (PFS) was not reached and the median overall survival (OS) was 24.2 months (95% confidence interval, 20.3-28.2 months). There was no statistically significant difference in PFS and OS between Avg Vol ≤ 99.5 cc and > 99.5 cc or other clinical parameters. Tumor regression during RT is a potential predictor of response in LA-NSCLC. kV-CBCT is a strong tool in assessing tumor regression during RT.