0Overall Survival Analysis of the Phase Ill CodeBreaK 300 Study of Sotorasib Plus Panitumumab Versus Investigator's Choice in Chemorefractory KRAs G12C Colorectal Cancer

Filippo Pietrantonio, MD' D; Lisa Salvatore, M D ^ { 2 , 3 } \oplus ; Taito Esaki, M D ^ { 4 } \oplus ; Dominik Paul Modest, M D ^ { 5 } \oplus ; David Paez Lopez-Bravo, { \mathsf { M D } } ^ { 6 } ... Julien Taieb, \mathsf { M D } ^ { 7 } \oplus ; Michalis V. Karamouzis, M D ^ { 8 } \oplus ; Erika Ruiz-Garcia, M D ^ { 9 } \oplus ; Tae Won Kim, M D ^ { 1 0 } \oplus : Yasutoshi Kuboki, MD11 \oplus Fausto Meriggi, M D ^ { 1 2 } : David Cunningham, M D ^ { 1 3 } \oplus ; Kun-Huei Yeh, MD, PhD 1 4 . 1 5 { / } ; Emily Chan, MD, \mathsf { P h D } ^ { 1 6 } ; Joseph Chao, M D ^ { 1 6 } \oplus . Qui Tran, \mathsf { P h D } ^ { \prime 6 } \oplus ; Chiara Cremolini, M D ^ { 1 7 } { \circ } ; and Marwan Fakih, M D ^ { 1 8 } \oplus

DOI https://doi.org/10.1200/JC0-24-02026

ABSTRACT

In the phase III CodeBreaK 3oo study, sotorasib 9 6 0 ~ { m g } -panitumumab significantly prolonged progression-free survival (PFS) versus investigator's choice (trifluridine/tipiracil or regorafenib) in patients with KRAs G12C-mutated chemorefractory metastatic colorectal cancer (mCRC). One hundred sixty patients were randomly assigned 1:1:1 to receive sotorasib 9 6 0 ~ { m g } -panitumumab ( \boldsymbol { { n } } = 5 3 ) , sotorasib 2 4 0 ~ { m g } -panitumumab { ( n } = 5 3 { ) } , or investigator's choice ( { \bf n } = 5 4 ; crossover permitted after primary analysis). Overall survival (os) analysis, a key secondary end point, although not adequately powered, was prespecified at 5 0 % maturity (after approximately 8o deaths). In this study, we report the OS, updated overall response rates (ORRs), and data for safety. After a median follow-up of 13.6 months, 24, 28, and 30 deaths occurred in the sotorasib 9 6 0 ~ { m g } -panitumumab, sotorasib 2 4 0 ~ { m g { - } } panitumumab, and investigator's choice arms, respectively; updated objective response rates (ORRs; 9 5 % CI) were 3 0 . 2 % - 9 5 % CI, 18.3 to 44-3), 7 . 5 % - 9 5 % CI, 2.1 to 18.2), and 1 . 9 % - 9 5 % CI, o.o to 9.9), respectively. Compared with investigator's choice, the hazard ratios (HRs [ 9 5 % CI]) for OS were 0.70 ( 9 5 % CI, 0.41 to 1.18; two-sided { \cal P } = . 2 0 \dot { \bf \cal _ { . } } with sotorasib 9 6 0 ~ { m g } -panitumumab and o.83 ( 9 5 % CI, 0.49 to 1.39; two-sided P = . 5 0 { { ~ } } with sotorasib 2 4 0 ~ { { m g } } . -panitumumab. No new safety signals were observed. Although not statistically significant, the observed OS HR and ORR along with prior PFS and safety findings support sotorasib 9 6 0 ~ { m g } -panitumumab as a standard of care in patients with chemorefractory KRAS G12C mCRC.

ACCOMPANYING CONTENT

Appendix

Data Sharing Statement

Protocol

Accepted February 21, 2025
Published April 11, 2025

J Clin Oncol 00:1-8
⊚ 2025 by American Society of
Clinical Oncology

View Online Article

Creative Commons Attribution Non-Commercial No Derivatives 4.0 License

INTRODUCTION

The randomized, phase III CodeBreaK 3oo study (ClinicalTrials.gov identifier: NcTo5198934) evaluated sotorasib, a specific and irreversible KRAsG12c inhibitor, at two doses (960 and 2 4 0 ~ { \ m g } , in combination with panitumumab, an epidermal growth factor receptor (EGFR) inhibitor, for the treatment of patients with KRAs G12Cmutated chemorefractory metastatic colorectal cancer (mcRc).' The primary analysis showed significantly prolonged progression-free survival (PFS) with sotorasib 9 6 0 ~ { \ m g } -panitumumab versus investigator's choice of trifluridine/tipiracil or regorafenib (median PFS, 5 . 6 \nu 2 . 0 months; hazard ratio [HR], o.48; 9 5 % CI, 0.30-0.78; P = .005).' In this study, we report findings from the protocolspecified final analysis of overall survival (Os), a key secondary end point.

METHODS

Trial Design and Patients

Details of the CodeBreaK 3oo trial have been published previously. Adult patients with KRAs G12C-mutated chemorefractory mCRC who were KRAsG12c inhibitor naive were randomly assigned 1:1:1 to receive sotorasib 9 6 0 ~ { \ m g ^ { - } } panitumumab, sotorasib 2 4 0 ~ { m g } -panitumumab, or investigator's choice (trifluridine/tipiracil or regorafenib). The protocol was approved by the institutional review boards or independent ethics committees at each site. All patients provided written informed consent as per the Declaration of Helsinki principles.

End Points and Assessments

OS (time from random assignment to death from any cause) and objective response rate (ORR) by blinded independent central review (BICR) per RECIST v1.1 were the key secondary end points. Other secondary end points included time to response, duration of response (DOR), and disease control (complete response, partial response, and stable disease) per BICR and safety. Criteria for tumor response assessment and adverse event (AE) severity grading have been described previously. Testing strategy (Appendix Fig A1) and statistical analysis methodology are reported in Appendix 1.

RESULTS

Patients and Treatment

Between April 19, 2022, and March 14, 2023, 160 patients (intention-to-treat [ITT] population) were randomly assigned to receive sotorasib 9 6 0 { m g } . -panitumumab ( \mathbf { n } = 5 3 ) , sotorasib 240 mg-panitumumab { I n } = 5 3 { . } , or investigator's choice \mathbf { \dot { n } } = \mathbf { \dot { \rho } } 54 [trifluridine/tipiracil, { \mathfrak { n } } = 3 7 regorafenib, \mathtt { n } = \mathtt { 1 4 } ] ; Fig 1). Baseline demographic and clinical characteristics of the treatment arms were generally balanced (Table 1).

The median duration of follow-up was 13.6 months. At data cutoff, 11 patients ( 2 0 . 8 % ) ,8 patients ( 1 5 . 1 % ) , and 1 patient ( 1 . 9 % ) remained on treatment in the sotorasib 9 6 0 ~ { { m g } - } panitumumab, sotorasib 2 4 0 { { ~ m g } } -panitumumab, and investigator's choice arms, respectively; the median (range) treatment duration was 6.0 months (1.0-16.3), 4.6 months (0.9-15.8), and 2.2 months (0.8-13.3), and after the randomly assigned treatment, 1 patient ( 1 . 9 % ) , 0 patients, and 17 patients ( 3 1 . 5 % ) received subsequent KRAsG12c inhibitor off-protocol.

Final OS Analysis

At data cutoff, 82 deaths occurred (Table 2). The estimated HR for OS was o.70 9 5 % CI, 0.41 to 1.18; two-sided P = . 2 0 ~ \nu statistically significance threshold o.o1875) with sotorasib 9 6 0 ~ { m g } -panitumumab and 0 . 8 3 C 9 5 % CI, o.49 to 1.39; descriptive two-sided \begin{array} { l l l } { P } & { = } & { . 5 0 { ⟩ } } \end{array} with sotorasib 240 mgpanitumumab (Fig 2 and Table 2). Median OS was not reached (not estimable [NE], 9 5 % CI, 8.61 to NE) with sotorasib 9 6 0 { m g } -panitumumab, and it was 11.9 months ( 9 5 % CI, 7.52 to NE) with sotorasib 2 4 0 ~ { m g } -panitumumab and 10.3 months ( 9 5 % CI, 7.0o to NE) with investigator's choice (Fig 2 and Table 2). The OS trend with sotorasib 9 6 0 { m g } -panitumumab in prespecified key subgroups was consistent with that in the ITT population; however, the small sample size of some subgroups caused wide CIs (Fig 2).

Subsequent Anticancer Therapy and OS Adjustment

While crossover was permitted after the primary analysis, no official on-protocol crossover had occurred at the time of the

FIG 1. consoRT diagram of patient disposition. For the sotorasib-panitumumab arms, data on patient numbers continuing treatment (treatment ongoing) represents continuing any component of the doublet for a patient. OS, overallsurvival.

NOTE. Data are No. ( % ) unless indicated otherwise. The intention-to-treat analysis set included allpatients who underwent random assignment. Abbreviations: OR durtion f reonse GFRmal gth fac retor, HR hazrd ratio; M KaplanMeer RAS,Kirt at saroma viral oncogene homolog; NE, not estimable; ORR, objective response rate; OS, overall survival; PFS, progression-free surival.
aHRs versus investigator's choice and 9 5 % CIs were estimated using a stratified Cox proportional hazards model.
{ ^ \mathsf { b } P } value was calculated using a stratified log-rank test.
9 5 % Cls were estimated using the method by Klein and Moeschberger with log-log transformation.
^ { \mathsf { d } } 0 \mathsf { S } rates and 9 5 % Cls were estimated using the method by Kalbfleisch and Prentice with log-log transformation.
9 5 % CIs were estimated using the Clopper-Pearson method.
fDifference of proportions versus investigator's choice and 9 5 % Cl were estimated using the stratified Cochran-Mantel-Haenszel method. DOR and time to response were estimated only for patints who achieved a confirmed best overal response of partial response or complete response.
hFor patients who crossed over from investigator's choice to sotorasib 9 6 0 { m g } -panitumumab, the data reported are from the date of the first dose of investigator's choice to the date of the first dose of sotorasib 960 mg-panitumumab.

OS final analysis. Details of receiving subsequent anticancer therapy (including { { K R A S } } ^ { { G 1 2 C } } inhibitors) are provided in Table 2. KRAsG12c inhibitor plus anti-EGFR antibody was received by 15 ( 2 7 . 8 % ) patients in the investigator's choice arm. An ad hoc sensitivity analysis to adjust for confounding effects of initiating subsequent therapies of interest { * K R A S ^ { G 1 2 C } } inhibitor plus EGFR inhibitor) showed a stratified HR of 0 . 6 5 - 9 5 % CI, 0.28 to 1.37) for the sotorasib 9 6 0 { m g } -panitumumab versus investigator's choice arm and 0 . 8 4 C 9 5 % CI, o.44 to 1.58) for the sotorasib 2 4 0 { m g } -panitumumab arm (Appendix Fig A2, Appendix Table A1).

Updated Efficacy Results

ORR ( 9 5 % CI) was 3 0 . 2 % - 9 5 % CI, 18.3 to 44-3), 7 . 5 % 9 5 % CI, 2.1 to 18.2), and 1 . 9 % C 9 5 % CI, 0.0 to 9.9) with sotorasib 9 6 0 ~ { \ m g } -panitumumab, sotorasib 240 mg-panitumumab, and investigator's choice, respectively. As statistical significance for OS was not achieved, ORR per BICR was not formally tested (Appendix Fig A3). Median DOR was 10.1 months 9 5 % CI, 3.9 to NE) in the sotorasib 9 6 0 ~ { m g } -panitumumab arm. Further efficacy updates are provided in Table 2.

Safety

AE profiles across arms were consistent with those previously reported (Appendix Table A2).' Grade \scriptstyle >= 3 treatmentrelated AEs (TRAEs) occurred in 4 5 . 3 % - 3 4 . 0 % , and 4 5 . 1 % of patients receiving sotorasib 9 6 0 ~ { m g } -panitumumab, sotorasib 240 mg-panitumumab, and investigator's choice, respectively; the most common >= 5 % patient incidence) were dermatitis acneiform, hypomagnesemia, and rash with sotorasib 96o mgpanitumumab; hypomagnesemia and diarrhea with sotorasib 240 mg-panitumumab; and neutropenia, anemia, and hypertension with investigator's choice. As events of interest, treatment-related grade \scriptstyle >= 3 hepatoxicity events occurred in 1 . 9 % 0 % , and 2 . 0 % of patients receiving sotorasib 960 mg-panitumumab, sotorasib 240 mgpanitumumab, and investigator's choice, respectively.

DISCUSSION

At this protocol-specified analysis of OS, sotorasib 9 6 0 { m g { - } } panitumumab, although not statistically significant, could suggest a potential 3 0 . 0 % relative reduction in the risk of death versus investigator's choice (HR, 0.70 [ 9 5 % CI, 0.41 to 1.18], P = . 2 ^ { * } in patients with chemorefractory KRAS G12Cmutated mcRc. This is promising for an incurable disease with a critical unmet need for molecularly selected therapies providing survival benefit. Updated ORRs across treatment arms were numerically slightly higher than those reported previously and remained higher with sotorasib 9 6 0 ~ { { m g } - } panitumumab versus investigator's choice. Despite higher use of subsequent therapies in the investigator's choice arm ( 6 1 . 1 % ) versus both sotorasib-panitumumab arms ( 4 3 . 4 % and 5 0 . 9 % , OS trends were similar after adjusting for subsequent medications such as a { { K R A S } } ^ { { G 1 2 C } } inhibitor plus EGFR inhibitor.

These findings support that the benefit provided by sotorasib 9 6 0 ~ { m g } -panitumumab in PFS and RECIST responses translated toward potential improvement in OS, even with > 3 0 % of patients in the control arm receiving KRAsGzc inhibitor

FIG 2. (Continued). on OS in subroup analyses. The dashed horizontal gray line indicates the median value. Vertical ars indicate censoring. Stratified Cox hazard ratio and stratified log-rank P value for sotorasib-panitumumab versus investigator's choice are provided. The intention-totreat population nudd al patients w undent rand assment. Had ratio was not estimatd whn thee wer ewer than 10 patients in either treatment arm in a subgroup. HR, hazard ratio; NE, not estimable; OS, overallsurvival.

combination as post-trial therapy. As the study was not powered for OS, the observed HR of o.70 and the median not being reached after 13.6 months of follow-up versus a median of 1o.3 months in the control arm indicate that even if not statistically different, OS improvement could be particularly meaningful, especially given that PFS is significantly longer.

The median OS with investigator's choice in this study was higher (1o.3 months) than that reported in similar phase III trials of trifluridine/tipiracil (7.1 months)? or regorafenib (6.4 months),3 which could be due to differences in patient populations and study designs (molecularly unselected patients and no crossover between treatment arms) and the availability of additional salvage therapies. Concurrent advances in late-line treatment of refractory mCRC showed significant Os improvements with regimens such as trifluridine/tipiracil plus bevacizumab (SUNLIGHT) and single-agent fruquintinib (FREscO-2), establishing both regimens as the new standard of care.4,5

Real-world data and the phase III RECOURSE trial showed that KRAS G12C mutations were predictive biomarkers for reduced OS benefit of trifluridine/tipiracil in late-stage mCRC, affecting patients eligible for treatment with these agents.6 A recent post hoc analysis of the SUNLIGHT trial showed that trifluridine/tipiracil plus bevacizumab improved OS versus trifluridine/tipiracil alone in patients with KRAS G12-mutated mCRC; OS trends were similar across all codon 12 KRAS mutations.7 Owing to differences in patient population and specific mutation, our study cannot be directly compared. While the benefit of adding bevacizumab to trifluridine/tipiracil in this molecular subgroup is unknown, targeting an identified molecular driver alteration is strongly preferred as a therapeutic strategy. As RAS testing, including KRAS G12C status, is routinely performed for patients with CRC to guide treatment with EGFR inhibitors, the sotorasib-panitumumab combination strategy can be readily integrated into the current treatment continuum for chemorefractory mCRC owing to its clinical benefits of a high response rate and long DORs.6 The KRYSTAL-1 trial used a combination of adagrasib and cetuximab for the treatment of mCRC and reported a median PFS of 6.9 months and ORR of 3 4 . 0 % which were comparable with our study despite the lack of a control arm and caveats of cross-trial comparison.8

The trial allowed crossover considering the unmet need in this chemorefractory patient population only after the primary analysis. However, patients in the control arm who subsequently received { { K R A S } } ^ { { G 1 2 C } } inhibitors had access to these options out of the trial, either in another clinical trial or as off-label usage, which might bias the interpretation of the OS results. As in the unadjusted ITT analysis (HR, 0.70 [ 9 5 % CI, 0.41 to 1.18], P = . 2 ^ { * } , the benefit with sotorasib 9 6 0 ~ { { m g } - } panitumumab treatment was generally comparable (HR, 0 . 6 5 [ 9 5 % CI, 0.28 to 1.37l) accounting for 15 of 54 patients in the investigator's choice arm who received a KRAsG12c inhibitor plus EGFR inhibitor as post-trial therapy. In addition, AE profiles across treatment arms remained consistent with the previous report with approximately 6 additional months of follow-up.

In conclusion, while to our knowledge this was the first phase III study of a KRAsG12c inhibitor plus an EGFR inhibitor to show a significant prolongation in PFS over standard therapies in chemorefractory mCRC, it was not powered to detect a statistically significant difference in OS. This was due to the small sample size in the study owing to the rarity of this patient population. Moreover, the post-trial use of KRASG12C plus EGFR inhibitors in 2 7 . 8 % of the patients in the investigator's choice arm outside of the trial could have further confounded the OS results. However, the analyses suggest a potential improvement in OS with sotorasib 9 6 0 { { ~ m g } } . -panitumumab versus investigator's choice.

Considering all the reported improved outcomes, this study supports a consistent treatment benefit of sotorasib 9 6 0 ~ { m g } -panitumumab as a new standard-of-care treatment for patients with KRAs G12C-mutated chemorefractory mCRC.

AFFILIATIONS

'Medical Oncology Department, Fondazione IRccs Istituto Nazionale dei Tumori, Milan, Italy 2Oncologia Medica, Universita Cattolica del Sacro Cuore, Rome, Italy 3Oncologia Medica, Comprehensive Cancer Center, Fondazione Policlinico Universitario Agostino Gemelli IRccs, Rome, Italy 4Department of Gastrointestinal and Medical Oncology, National Hospital Organization Kyushu Cancer Center, Fukuoka, Japan 5Medicine Department of Hematology, Oncology and Tumor Immunology, Charite - Universitaetsmedizin Berlin, Berlin, Germany

6Department of Medical Oncology, Hospital de Ia Santa Creu i Sant Pau,
Barcelona, Spain
7Universite Paris Cite, SIRIC CARPEM Comprehensive Cancer Center,
Department of Gastroenterology and Digestive Oncology, Hopital
Europeen Georges Pompidou, Paris, France
8Department of Biological Chemistry, National and Kapodistrian
University of Athens-School of Medicine, Athens, Greece
^ { 9 } { \sf G I } Oncology Department & Translational Medicine Laboratory, INCAN -
Instituto Nacional de Cancerologia, Ciudad de Mexico, Mexico
10Oncology Department, Asan Medical Center, University of Ulsan
College of Medicine, Seoul, Republic of Korea
11Experimental Therapeutics and GI Oncology Department, National
Cancer Center Hospital East, Kashiwa, Japan
12Oncology Department, Fondazione Poliambulanza Istituto
Ospedaliero, Brescia, Italy
13Medicine Department, The Royal Marsden Hospital, London and
Sutton, United Kingdom
14Department of Oncology, National Taiwan University Hospital, Taipei
City, Taiwan
15Graduate Institute of Oncology, National Taiwan University College of
Medicine, Taipei, Taiwan
16Amgen Inc., Thousand Oaks, CA
17Department of Translational Research and New Technologies in
Medicine and Surgery, University of Pisa, Pisa, Italy
18Medical Oncology & Therapeutics Research, City of Hope
Comprehensive Cancer Center, Duarte, CA

CORRESPONDING AUTHOR

Marwan G. Fakih, MD; e-mail: mfakih@coh.org.

DISCLAIMER

The sponsor was involved in the design and conduct of the study, the analysis and interpretation of the data, and the decision to submit this manuscript. Amgen managed patient data collected at the study sites and provided medical writing support.

PRIOR PRESENTATION

Presented as a rapid oral at the ASc0 2024 Annual Meeting, Chicago, IL, May 31-June 4, 2024 (abstr LBA3510).

SUPPORT

Supported by Amgen Inc.

CLINICAL TRIAL INFORMATION NCT05198934

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Disclosures provided by the authors are available with this article at DOI https://doi.org/10.1200/JC0-24-02026.

DATA SHARING STATEMENT

A data sharing statement provided by the authors is available with this article at D0I https://doi.org/10.1200/JC0-24-02026. Qualified researchers may request data from Amgen clinical studies. Complete details are available at the following link: http://www.amgen.com/ datasharing.

AUTHOR CONTRIBUTIONS

Conception and design: Emily Chan, Qui Tran
Provision of study materials or patients: Filippo Pietrantonio, Lisa
Salvatore, Taito Esaki, Dominik Paul Modest, David Paez Lopez-Bravo,
Julien Taieb, Michalis V. Karamouzis, Erika Ruiz-Garcia, Tae Won Kim,
Yasutoshi Kuboki, Fausto Meriggi, David Cunningham, Kun-Huei Yeh,
Chiara Cremolini, Marwan Fakih
Collection and assembly of data: All authors
Data analysis and interpretation: All authors
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors

ACKNOWLEDGMENT

We thank the patients, investigators, study staff, and the data monitoring committee who contributed to this study. Medical writing support was provided by Shubha Dastidar, PhD, CMPP (Cactus Life Sciences, part of Cactus Communications); Saimah Raza, PhD (Cactus Life Sciences, part of Cactus Communications); and Christopher Nosala, PhD, CMPP (Amgen Inc.) and was funded by Amgen Inc.

REFERENCES

1. Fakih are aki , t a ai us patab in rrcy cotl ccer wh mtte KA 12. Engl J Md 389:2125-239 2023
2. Mayer RVan tsem Falo A e al: Rndze til of TAS-102for ectoy mstatic cooctalcer. Eng Md 372:1909-1919 2015
3. phase 3 trial. Lancet 381:303-312, 2013
4 blind, phase 3 study. Lancet 402:41-53, 2023
5. Prager Taie , Faki t al Trdntipial d cizab n ray mestti ct cr.El Md 381657-1667, 2023
6an t a i t 64203
7 of the phase Il SUNLIGHT trial. ESM0 Open 9:102945, 2024
22

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

verall Survival nalysi of the Phas Il CodeBreaK 300Study of Storasib Plus Panitmumab Vrs Investigator's Choice in Chemorefractoy <RAS G12C Colorectal Cancer

The following represents disclosure information provided by authors of this manuscript. ll relationships are considere compensated unless otherwise noted. Relationships are self-held unless noted. \mid \mid = Immediate Family Member, Inst \ l = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about Asco's conflic of interest policy, please refer to ww.asco.org/wc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).

Filippo Pietrantonio
Honoraria: SERVIER, Bayer, MSD Oncology, Amgen, Pierre Fabre, Bristol
Myers Squibb, Merck Serono, Astellas Pharma, Takeda, Ipsen,
Johnson&Johnson, Rottapharm Biotech, Seagen, AstraZeneca, Daiichi
Sankyo, BeiGene
Consulting or Advisory Role: Amgen, SERVIER, MSD Oncology, Bayer,
Merck Serono, Takeda, GlaxoSmithKline, Rottapharm Biotech, Johnson
& Johnson/Janssen, Pfizer, Astellas Pharma, BMS, BeiGene, Agenus,
Gilead Sciences, Daiichi-Sankyo, Incyte, Jazz Pharmaceuticals, Pierre
Fabre, AstraZeneca
Research Funding: Bristol Myers Squibb (Inst), Astrazeneca (Inst),
Incyte (Inst),Agenus (Inst), Lilly (Inst), Amgen (Inst) Rottapharm
Biotech (Inst)
Travel, Accommodations, Expenses: Pierre Fabre, Merck Serono,
Takeda Science Foundation
Lisa Salvatore
Honoraria: Merck Serono, SERViER, Bayer, Amgen, AstraZeneca, Pierre
Fabre, MSD, GlaxoSmithKline, Incyte, Takeda
Consulting or Advisory Role: Merck Serono, SERVIER, Bayer, Amgen,
AstraZeneca, Pierre Fabre, MSD, GlaxoSmithKline, Incyte, Takeda
Research Funding: Merck Serono (Inst)
Travel, Accommodations, Expenses: Merck Serono, Bayer, SERVIER,
Pierre Fabre, Amgen

Taito Esaki

Honoraria: Taiho Pharmaceutical, Daiichi Sankyo, Chugai Pharma, Ono
Pharmaceutical, Hisamitsu Pharmaceutical, Roche Diagnostics
Solutions, Zeria Pharmaceutical, MSD, Lilly
Research Funding: MSD (Inst), Ono Pharmaceutical (Inst), Pfizer (Inst),
Chugai Pharma (Inst), Asahi Kasei (Inst),Amgen (Inst), Jazz
Pharmaceuticals (Inst), ALX Oncology (Inst), Seagen (Inst), Taiho
Pharmaceutical (Inst), Bristol Myers Squibb Japan (Inst), Nihonkayaku
(Inst)

Dominik Paul Modest

Honoraria: Merck Serono, Amgen, SERVIER, Bristol Myers Squibb, Taiho Pharmaceutical, Onkowissen, Merck, Pierre Fabre, AstraZeneca, Lilly, Takeda, GlaxoSmithKline, Seagen, Cor2Ed, Boehringer Ingelheim, Regeneron, Bicara Therapeutics, Rottapharm Biotech, IKF Klinische Krebsforschung, 21up
Consulting or Advisory Role: Merck Serono, Amgen, SERviER, Pierre Fabre, Lilly, Cor2Ed, Onkowissen, Regeneron, GlaxoSmithKline, Takeda, Incyte
Research Funding: Amgen (Inst), Servier (Inst)
Travel, Accommodations, Expenses: Amgen, Merck Serono, SERVIER
David Paez Lopez-Bravo
Consulting or Advisory Role: Amgen, Takeda, TERUmO, Advanced
Accelerator Applications
Speakers' Bureau: Amgen, MSD
Research Funding: Merck KGaA (Inst)
Travel, Accommodations, Expenses: Merck KGaA, Lilly, Amgen, Advanz
Pharma

Julien Taieb Consulting or Advisory Role: Merck KGaA, Amgen, SERVIER, MSD, Pierre Fabre, Novartis, AstraZeneca, BMS, takeda, Astellas Pharma, PROskOPE, Brenus Pharma, Boehringer Ingelheim, sanofi, natera Speakers' Bureau: SERVIER, Amgen, Merck, MSD, Pierre Fabre

Michalis V.Karamouzis
Honoraria: BMS, MSD, Astellas Pharma, Roche, Pfizer Hellas, SERVIER
Speakers' Bureau: BMSi
Travel, Accommodations, Expenses: Ipsen

Erika Ruiz-Garcia Consulting or Advisory Role: Roche/Genentech, Amgen, BMS, Bayer, Merck Serono, Astellas Pharma, AstraZeneca, SERVIER Travel, Accommodations, Expenses: Gilead Sciences

Tae Won Kim
Employment: Asan Medical Center
Research Funding: Roche/Genentech (Inst), Genome Insight (Inst)
Yasutoshi Kuboki
Honoraria: Taiho Pharmaceutical, Lilly Japan, Takeda, Kyowa Kirin
Consulting or Advisory Role: Takeda, Amgen, Incyte, Abbvie, Noile
Immune Biotech, Inc
Research Funding: Taiho Pharmaceutical (Inst), Takeda (Inst), Incyte
(Inst), Ono Pharmaceutical (Inst), Boehringer Ingelheim (Inst), Amgen
(Inst), Chugai Pharma (Inst), Astellas Pharma (Inst), Genmab (Inst),
Abbvie (Inst), Lilly (Inst), AstraZeneca (Inst), Merck Serono (Inst),
Jiangsu Hengrui Pharmaceuticals (Inst), Novartis (Inst), Carna
Biosciences (Inst), Daiichi Sankyo/UCB Japan (Inst), Kyowa Kirin (Inst),
Bristol Myers Squibb Japan (Inst)
Travel, Accommodations, Expenses: Amgen, Incyte, Chugai Pharma
David Cunningham
Stock and Other Ownership Interests: OVIBIO
Consulting or Advisory Role: OVIBIO
Research Funding: Bayer (Inst), 4SC (Inst), Clovis Oncology (Inst), Lilly
(Inst), Leap Oncology (Inst), Roche (Inst) Kun-Huei Yeh
Honoraria: TTY Biopharm, Amgen, Roche, Ono Pharmaceutical, Bristol Myers Squibb, MSD, Pfizer, Daiichi Sankyo, Novartis, Merck, Takeda Consulting or Advisory Role: Daiichi Sankyo/Astra Zeneca, Novartis, Daiichi Sankyo, MSD, AstraZeneca, Daiichi Sankyo/Astra Zeneca, Pierre Fabre, Bayer, Pfizer, MSD, DSI/AZ, Takeda
Travel, Accommodations, Expenses: Pfizer, Ono Pharmaceutical, Takeda
Chiara Cremolini
Honoraria Roche, Amgen, Bayer,SERViER,MSD, Merck, Pierre Fabre,
Merck, GlaxoSmithKline, Takeda
Consulting or Advisory Role: Roche, Bayer, Amgen, MSD, Pierre Fabre,
Nordic Bioscience, Rottapharm Biotech, Bicara Therapeutics, Takeda,
Revolution Medicines
Speakers' Bureau: SERVIER, Merck, Pierre Fabre, MSD
Research Funding:Merck, Bayer, Roche, SERVIER
Emily Chan
Employment: Amgen
Stock and Other Ownership Interests: Amgen
Patents, Royalties, Other Intellectual Property: Sotorasib patent
Joseph Chao
Employment: Amgen
Stock and Other Ownership Interests: Amgen
Qui Tran
Employment: Amgen
Stock and Other Ownership Interests: Amgen Marwan Fakih
Honoraria: Amgen
Consulting or Advisory Role: Taiho Pharmaceutical, Bayer, Pfizer, Roche/Genentech, Mirati Therapeutics, BMS, Eisai, Merck, Abbvie, Adagene, Delcath Systems, Entos, Janssen Research & Development, Microbial Machines, Nouscom, Tempus, Totus Medicines
Research Funding: Verastem (Inst), Roche/Genentech (Inst), Agenus (Inst)
No other potential conflicts of interest were reported.

APPENDIX 1. STATISTICAL ANALYSIS

Statistical analysis methods for this trial have been previously reported.' All randomly assigned patients were included in efficacy analyses (intention-to-treat). Safety was assessed in allrandomly assigned patients who received at least 1 dose of assigned treatment. The sample size was estimated on the basis of the primary end point of blinded independent central review-assessed progression-free survival (PFS) and has been reported previously.' Overall survival (OS) was tested only if PFS results for the sotorasib combination arm (960 or 2 4 0 ~ \mathsf { m g } versus control arm reached statistical significance according to the prespecified conditions in the protocol.

A hierarchical testing strategy between PFS, OS, and overall response rate (ORR) was used to control the overall type I error rate at two-sided 5 % alpha using the MaurerBretz testing procedure (Appendix Fig A1). Because statistical significance was achieved for sotorasib 960 mg-panitumumab at the PFS primary analysis, 2 . 5 % ^ { \star } ( 3 / 4 ) alpha was propagated to test 0S for sotorasib 960 mg-panitumumab versus investigator's choice at its final analysis using the stratified log-rank test, which was specified to occur after 5 0 % of patients had observed events (approximately 80 deaths). 0S for sotorasib 240 mg-panitumumab is included descriptively. The study sample size was not powered to demonstrate a statistically significant difference in OS. This OS analysis timing aimed to provide more mature OS data subsequent to the previous analysis triggered by PFS. A later timing for additional OS events would not substantially increase the OS power because of sample size constraint. If OS achieved statistical significance at its final analysis, ORR would be tested.

The Kaplan-Meier method was used to estimate median OS, and a stratified Cox proportional hazards model was used to estimate the hazard ratio and 9 5 % Cl for death in the OS analysis. Random assignment stratification factors were applied to the stratified Cox model. The median follow-up time for OS was estimated using the reversed Kaplan-Meier method.

Data are No. ( % ) unless indicated otherwise.

Abbreviatins: GFR demal gowth factr recetor; R had ratio RAS,Kirste rat saoma viral ocogene hmolog; S, veral surival, RPSFTM, rank preserving structural failure time model.
aA g-estimation procedure was used to find the value of the acceleration factor and its 9 5 % Cl such that the counterfactual OS times were balanced across the treatment groups. Recensoring was applied to the counterfactual OS times.
bHRs were estimated using a stratified Cox proportional hazards model. 9 5 % Cls were estimated using bootstrapping (1,000 samples) for RPSFTM. NOTE. Data are No. ( % ) . AEs were coded using MedDRA version 26.1. The CTCAE version 5.0 was used to grade severity of AEs. Abbreviatios: A, adverse event; CTAE, Cmmon Tenology Criteri for Adverse Events; eCRF, electronic case report for; MedDRA, Medal Dictionary for Regulatory Activities; TEAE, treatment-emergent adverse event; TRAE, treatment-related adverse event.
aThe safety analysis set included all randomly assigned patients who received at least 1 dose of their assigned treatment.
bTRAEs are those TEEs demed reatedto stuy treatment s per the invetigator as reported on the eCRF. TEAE was consideed aTRA if there was a reasonable possiliy that the evet may hae be cased by the stdy tratment. I the unlikely eent that the relationship was missing, the TEAE was considered treatment related.
cSotorasib dose reduction was not allowed for the 2 4 0 - \mathsf { m g } group. Patients could resume the same dose after interruption or permanently discontinue.
dData denote the "Skin and subcutaneous tissue disorders" system organ class.
eAny-grade events listed are those that occurred in at least 5 % of patients in any arm. Grade ^ { \ge 3 } events listed are those that occurred in at least 3 % of patients in any arm.