Background: Tumor growth rates (g) estimated using imaging measurements, have been associated with overall survival (OS) and progression-free survival (PFS) in patients with mNSCLC, including those treated with first-line immunotherapy (1L IO) or chemotherapy (chemo). Here, we evaluated whether early g estimates within 18 weeks of first treatment dose are associated with survival outcomes for 1L treatment in mNSCLC. Methods: This was a retrospective analysis of data from patients randomized to either nivolumab+ipilimumab (NIVO+IPI) or chemo in CheckMate 227 Part 1 (NCT02477826), or NIVO+IPI+chemo or chemo alone in CheckMate 9LA (NCT03215706). Tumor assessments were performed by blinded independent central review using RECIST v 1.1 at baseline, every 6 weeks for the first 48 weeks and then every 12 weeks until disease progression. The analysis included patients with at least 3 measurable timepoints, including baseline, week 6, week 12, and/or week 18. If a patient did not have a week 18 measurement, the first three measurements alone were used. To derive the median early g, sum of longest diameters (SLD), based on baseline, weeks 6, 12 and/or 18 assessments, and time relative to baseline were fitted to the model defined by sum of exponential growth (g) and decay (d): SLD (t) = exp (–d x t) + exp (g x t) – 1. OS and PFS were estimated using Kaplan-Meier methodology. Results: In the two studies, 865/1166 (75%) of randomized patients in CheckMate 227 Part 1 and 562/719 (78%) in CheckMate 9LA had evaluable tumor growth rate data (Table). The median early g at weeks 12 and 18 was numerically lower for the IO-containing arm vs chemo arm in both studies (Table). Patients with lower growth rate at week 12 or week 18 (g in first quartile [Q1]) had better OS relative to those with higher rate (g in fourth quartile [Q4]) across all treatment arms (Table). A similar trend was observed for PFS. Conclusions: Early g estimates based on 2 or 3 post-baseline tumor assessment timepoints were associated with longer-term survival outcomes for 1L treatment of mNSCLC and could discern efficacy outcomes. These findings provide the foundation for further research, which may incorporate volumetric segmentations of measurable lesions and radiomic feature changes to further explore indicators of patient outcomes that could inform future clinical trials and clinical practice.