Background: Adagrasib (ADA) is a potent covalent inhibitor of KRASG12C with favorable properties such as long half-life (23 h), dose-dependent pharmacokinetics, and brain penetrance. In the phase 1/2 KRYSTAL-1 trial, ADA demonstrated deep and durable responses with promising PFS and OS in patients (pts) with previously treated KRASG12C-mutated NSCLC. Here, we report the primary analysis from KRYSTAL-12 (NCT04685135), a randomized, open-label phase 3 trial of ADA compared with docetaxel (DOCE) in pts with KRASG12C-mutated locally advanced or metastatic NSCLC who had previously received a platinum-based chemotherapy, concurrently or sequentially with anti-PD-(L)1 therapy. Methods: Pts with KRASG12C-mutated locally advanced or metastatic NSCLC, previously treated with platinum-based chemotherapy and anti-PD-(L)1 therapy, were randomized 2:1 (stratified by region [non-Asia Pacific vs Asia Pacific] and sequential vs concurrent chemoimmunotherapy) to receive ADA (600 mg BID orally; tablet formulation) or DOCE (75 mg/m2 Q3W IV), with the ability to crossover to ADA upon disease progression (assessed by real-time blinded independent central review [BICR]). No washout period was required between prior anti-PD-(L)1 therapy and study treatment. Primary endpoint was PFS assessed per BICR according to RECIST v1.1. Secondary endpoints included ORR by BICR, duration of response (DOR), OS, 1-year OS rate, and safety. Results: In total, 301 pts were randomized to ADA and 152 to DOCE. Baseline characteristics were generally similar between treatment arms. With a median follow-up of 9.4 mo (data cutoff 31 Dec, 2023), the primary endpoint of PFS was significantly improved with ADA over DOCE (HR 0.58 [95% CI, 0.45-0.76]; P < 0.0001; median PFS 5.49 vs 3.84 mo). ORR by BICR was also significantly higher with ADA compared with DOCE (31.9% [95% CI, 26.7-37.5] vs 9.2% [95% CI, 5.1-15.0]; odds ratio 4.68 [95% CI, 2.56-8.56]; P < 0.0001); median DOR was 8.31 (95% CI, 6.05-10.35) vs 5.36 (95% CI, 2.86-8.54) mo, respectively. Treatment-related adverse events (TRAEs) were reported in 94.0% of pts treated with ADA and 86.4% with DOCE; grade >= 3 TRAEs occurred in 47.0% and 45.7% of pts, respectively. TRAEs led to discontinuation of ADA in 7.7% of pts and DOCE in 14.3%. Additional efficacy and safety analyses, including subgroup analyses, will be presented. Conclusions: In the phase 3 KRYSTAL-12 trial, ADA demonstrated a statistically significant and clinically meaningful improvement in PFS and ORR over DOCE in pts with previously treated KRASG12C-mutated NSCLC. Safety profile of ADA was consistent with previous reports and with no new safety signals. These results further support ADA as an efficacious treatment option for pts with previously treated KRASG12C-mutated locally advanced or metastatic NSCLC. Funding: Mirati, a Bristol Myers Squibb Company. Acknowledgements: KRYSTAL-12 was sponsored by Mirati, a Bristol Myers Squibb Company. Third-party medical writing support, under the direction of the authors, was provided by Flaminia Fenoaltea, MSc, of Ashfield MedComms, an Inizio company, and was funded by Mirati, a Bristol Myers Squibb Company. Clinical trial information: NCT04685135.