Efficacy and Safety of Adagrasib plus Cetuximab in Patients with KRAS^G12C-Mutated Metastatic Colorectal Cancer

KRAS^G12C mutations occur in 3% to 4% of colorectal cancer cases and are associated with poor prognosis (1, 2). Adagrasib is a selective, irreversible inhibitor of KRAS^G12C and has favorable properties, including long half-life (23 hours), dose-dependent pharmacokinetics, and central nervous system penetration (3–5). Adagrasib was granted accelerated approval by the FDA for the treatment of adult patients with locally advanced or metastatic KRAS^G12C-mutated non–small cell lung cancer who have received at least one prior systemic therapy (6).

In colorectal cancer, adaptive resistance to KRAS^G12C inhibition can occur via EGFR-mediated reactivation of the MAPK pathway (7). Preclinical data suggest that dual EGFR/KRAS^G12C blockade mitigates adaptive resistance in KRAS^G12C-mutated colorectal cancer cell lines (7). In a phase I cohort (n = 32) of KRYSTAL-1, adagrasib plus cetuximab yielded higher clinical activity than adagrasib monotherapy in patients with KRAS^G12C-mutated metastatic colorectal cancer. Objective response rate (ORR) was 46% [95% confidence interval (CI), 28–66) and 19% (95% CI, 8–33), respectively; median progression-free survival (PFS) was 6.9 months (95% CI, 5.4–8.1) and 5.6 months (95% CI, 4.1–8.3), respectively; per investigator assessment (8). Adagrasib, in combination with cetuximab, has received FDA breakthrough therapy designation for the treatment of patients with KRAS^G12C-mutated advanced colorectal cancer that has progressed after previous treatment (9). Adagrasib, in combination with cetuximab or panitumumab, is recommended for patients with previously treated advanced or metastatic KRAS^G12C-mutated colorectal cancer by the National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (NCCN Guidelines); adagrasib may also be considered as monotherapy if a patient is unable to tolerate an anti-EGFR antibody due to toxicity (10, 11).

Here, we report pooled data from the phase I and phase II cohorts of the KRYSTAL-1 trial for previously treated patients with KRAS^G12C-mutated unresectable or metastatic colorectal cancer who received adagrasib plus cetuximab. To contextualize the results obtained with combination therapy, we include an additional phase II adagrasib monotherapy cohort distinct from the adagrasib monotherapy cohort previously described (8). Finally, we provide exploratory evidence supporting a role for circulating tumor DNA (ctDNA) as a tool to define biomarkers of response and molecular resistance.

As of June 30, 2023, 94 patients with KRAS^G12C-mutated colorectal cancer had received ≥1 dose of study treatment across phase I (n = 32) and phase II (n = 62) combination cohorts. Median follow-up was 11.9 months (30.0 months for the phase I cohort and 10.0 months for the phase II cohort). Median (range) age was 57 (24–75) years; median (range) prior lines of systemic therapy was 3 (1–9). All patients received prior fluoropyrimidine, 98.9% prior oxaliplatin, 95.7% prior anti-VEGF treatment, and 94.7% prior irinotecan (Table 1). Median (range) duration of treatment with adagrasib was 7.4 (0.03–36.4) months. Seventeen patients started cetuximab intravenously weekly (400 mg/m² loading dose followed by 250 mg/m²), seven of whom later switched to treatment every 2 weeks (500 mg/m²). Seventy-seven patients started with cetuximab treatment every 2 weeks (500 mg/m²), consistent with the cetuximab U.S. prescribing information (12). Median (range) duration of treatment with cetuximab was 7.1 (0.2–34.5) months. As of data cutoff, treatment was ongoing in 18 of 94 (19.1%) patients. A total of 38 of 94 (40.4%) patients continued adagrasib plus cetuximab beyond disease progression, 10 of 94 (10.6%) patients continued adagrasib only beyond disease progression, and one patient continued treatment with cetuximab only beyond disease progression. The participants in this trial were representative of the broader population of patients with KRAS^G12C-mutated colorectal cancer (Supplementary Table S1). See Supplementary Appendix for details of the adagrasib monotherapy cohort (Supplementary Table S2).

Overall, 94 patients with measurable disease at baseline received ≥1 dose of study treatment. The ORR as assessed by blinded independent central review (BICR) was 34.0% (95% CI, 24.6–44.5) and disease control rate (DCR) was 85.1% (95% CI, 76.3–91.6; Table 2; Fig. 1A–C). Median duration of response (DOR) was 5.8 months (95% CI, 4.2–7.6) and median PFS was 6.9 months (95% CI, 5.7–7.4; Table 2; Fig. 1D). The ORR as assessed by investigator was 42.6% (95% CI, 32.4–53.2), DCR was 86.2% (95% CI, 77.5–92.4), and median DOR was 5.9 months (95% CI, 5.5–7.6; Table 2). Median overall survival (OS) was 15.9 months (95% CI, 11.8–18.8; Table 2; Fig. 1E). Separate efficacy data, per BICR, for the phase I and phase II cohorts, as well as for the pooled data from both cohorts are shown in the Supplementary Appendix (Supplementary Table S3). In total, 36 patients (38.3%) received subsequent anticancer therapy. Of 49 patients who continued study treatment after disease progression, postprogression scans were available for 34 (33 who continued adagrasib and cetuximab and one who continued adagrasib alone). Continued tumor control was observed in many patients receiving treatment beyond progression (Supplementary Fig. S1), with tumor shrinkage observed versus baseline at the first postprogression scan for 18 of 33 patients (54.5%) continuing adagrasib plus cetuximab and 1 of 1 continuing adagrasib only.

Subgroup analyses of ORR (per BICR) for the pooled phase I and phase II adagrasib plus cetuximab combination data are shown in Fig. 2. For all subgroups, the range of 95% CIs overlapped the ORR for the overall population (34.0%). Subgroup analyses based on the presence of brain, bone, or adrenal metastases at baseline were omitted from the forest plot due to small patient numbers (less than five) in one of the subgroup categories.

For the adagrasib monotherapy cohort, ORR per BICR was 21.4% (95% CI, 10.3–36.8), median DOR was 15.2 months (95% CI, 2.8–not evaluable), median PFS was 4.1 months (95% CI, 2.8–6.5), and median OS was 12.2 months (95% CI, 8.1–15.2; see Supplementary Appendix for further detail; Supplementary Table S4; Supplementary Fig. S2A–S2D).

No new safety signals were observed in this larger dataset. The most frequent treatment-related adverse events (TRAEs) with adagrasib plus cetuximab included nausea (60.6%), vomiting (51.6%), diarrhea (48.9%), and dermatitis acneiform (47.9%; Table 3). Grade 3–4 TRAEs occurred in 27.7% of patients; there were no grade 5 TRAEs. TRAEs led to adagrasib dose reductions in 29.8% of patients and cetuximab dose reductions in 6.4% of patients. One patient required dose reduction of adagrasib and cetuximab. TRAEs led to discontinuation of cetuximab in eight patients (8.5%; Table 3), six of whom continued with adagrasib as a single agent; no TRAEs led to discontinuation of adagrasib. See Supplementary Appendix for the adagrasib monotherapy cohort safety data (Supplementary Table S5).

All 94 patients had a KRAS^G12C mutation detected in tumor tissue and 83 patients had available baseline plasma ctDNA for analysis. The concordance of KRAS^G12C mutation–positive status at baseline for patients with paired tumor and plasma samples was 83% (69 of 83) with the KRAS^G12C alteration detected in tissue but not ctDNA for 14 of 83 cases. ORR was 35% in the 83 patients with paired ctDNA/tissue, 39% in those with KRAS^G12C detected in ctDNA, and 14% in those where KRAS^G12C was not detected in ctDNA. Clearance of ctDNA of the targeted oncogene has been proposed as a potential pharmacodynamic marker, predictive of response to the matched therapy and of OS (13, 14). We therefore evaluated KRAS^G12C ctDNA at cycle 1 day 1 (C1D1) (baseline), C2D1 (4 weeks), and C4D1 (8 weeks), and associations of ctDNA clearance with response (Fig. 3). Clearance of KRAS^G12C ctDNA was defined as more than 90% decrease in the variant allele fraction based on prior defined thresholds (15). These analyses were only performed in phase I patients. Samples at baseline, C2D1, and C4D1 were available in 15 of 32 of these patients. ORR was 60% (9 of 15) in the evaluable population. At C4D1, 12 of 15 (80%) patients had a decrease in KRAS^G12C ctDNA by at least 90% and responses were seen in 67% (8 of 12) of these patients. Three patients had less than 90% clearance of KRAS^G12C at C4D1 and one response (1 of 3; 33%) was seen among these patients (Fig. 3).

To identify potential mechanisms underlying acquired cli­nical resistance, we compared genomic alterations in plasma ctDNA at baseline with those in ctDNA collected at end of therapy (EOT). Paired plasma samples were available for 34 patients across the phase I and II adagrasib plus cetu­ximab combination therapy cohorts. A total of 25 patients (74%) had at least one acquired pathogenic alteration, with a total of 116 alterations identified overall. In 21 of 25 cases, multiple acquired alterations were observed in the same patient (Fig. 4). Acquired alterations included RTK amplifications/fusions (12 of 34 patients; 35%), EGFR amplification (3 of 34; 9%), secondary mutations in KRAS (14 of 34; 41%), KRAS amplification (10 of 34; 29%), and mutations/fusions in other MAPK pathway genes (13 of 34; 38%).

Dual KRAS/EGFR inhibition with adagrasib plus cetuximab demonstrated promising clinical activity in patients with heavily pretreated (≥3 lines of therapy) unresectable or metastatic KRAS^G12C-mutated colorectal cancer from the KRYSTAL-1 trial. ORR was 34.0%, median DOR was 5.8 months, median PFS was 6.9 months, and median OS was 15.9 months. Consistent with a previous report (8), in this study, ORR, median PFS, and median OS with adagrasib plus cetuximab were numerically higher than observed with adagrasib monotherapy. The combination of adagrasib plus cetuximab was tolerable, with no new safety concerns identified; most TRAEs were grade 1–2 in severity. Adagrasib-related AEs resulted in adagrasib dose reduction in 29.8% of patients and many patients continued to derive benefit after dose reduction. Some patients continued study treatment beyond RECIST progression and, among these, some appeared to obtain a degree of ongoing disease control. However, there are no prospective data on this treatment approach and it remains investigational.

The efficacy data reported here are consistent with data showing synergistic activity of KRAS^G12C and EGFR inhibitors in KRAS^G12C-mutated colorectal cancer cell lines (7, 16). Similarly, phase I/II data for other KRAS^G12C inhibitors in colorectal cancer demonstrated greater activity in combination with anti-EGFR antibodies versus monotherapy (17–21). Recent findings from a phase III study in the third-line setting (CodeBreaK 300) showed a PFS benefit with sotorasib (960 mg once daily) when combined with panitumumab versus standard of care [median PFS 5.6 months vs. 2.2 months; HR, 0.49 (95% CI, 0.30–0.80; P = 0.006); ref. 22]. In addition, ORR was higher in patients treated with sotorasib plus panitumumab (26.4%; 95% CI, 15.3–40.3) than in those treated with standard of care (0%; 95% CI, 0.0–6.6; ref. 22).

These data support adagrasib plus cetuximab as a potential therapy for heavily pretreated patients with unresectable or metastatic KRAS^G12C-mutated colorectal cancer, where treatment options are limited (8), and are currently under review by the FDA. The low efficacy of current late-line standard of care trifluridine/tipiracil plus bevacizumab (SUNLIGHT trial; median PFS 5.6 months; ORR 6%), fruquintinib (median PFS 3.7 months; ORR 2%), or regorafenib (median PFS 1.9 months; ORR 1%) underscores the high unmet need in heavily pretreated patients (23–25). However, it should be noted that these trials included a molecularly unselected population, while this study only includes patients with KRAS^G12C-mutated colorectal cancer, a population with a generally worse prognosis (26). No subgroup analysis has been presented for trifluridine-tipiracil plus bevacizumab, fruquintinib, or regorafenib in patients with KRAS^G12C-mutated colorectal cancer.

There are multiple series highlighting the potential of ctDNA for both baseline genotyping in colorectal cancer and defining resistance mechanisms to targeted therapies (27–29). In our study population, baseline KRAS^G12C was detected in ctDNA for 69 of 83 patients, with 83% concordance between paired tumor samples. This highlights the possible utility of liquid biopsies for molecular genotyping, facilitating more rapid patient access to targeted therapy options in the future (30). In an exploratory analysis of plasma ctDNA kinetics, among patients treated with adagrasib plus cetuximab in the phase I portion, the ORR for those with a ≥90% decrease in KRAS^G12C ctDNA at 8 weeks (67%) was higher than that for those with <90% clearance of KRAS^G12C ctDNA at this timepoint (33%). While several variables can influence ctDNA detection and the optimal early timepoint is not known, our data further support that early, profound decrease is a favorable feature. In an exploratory analysis, we observed a diverse spectrum of acquired genomic alterations in patients treated with adagrasib in combination with cetuximab. This is in line with previous reports of acquired KRAS mutations and acquired RTK/RAS/MAPK/PI3K pathway alterations following concurrent KRAS^G12C and EGFR inhibition in colorectal cancer (31, 32).

Limitations of this study include the nonrandomized, open-label trial design as well as the absence of a comparator arm. The proportion of patients with lung metastases was high in our cohort (71.3%) and higher than that for liver metastases (63.8%), although RAS mutations have been associated with increased tendency for lung involvement in colorectal cancer (33–35). In the context of a non-randomized study, it is not possible to determine whether factors such as this may have impacted the outcomes observed, however, the patient population in this study was representative of later-line colorectal cancer in the United States and responses were observed across all subgroups analyzed.

As observed in large genomic studies (36), KRAS mutations are enriched in microsatellite stable (MSS) tumors relative to microsatellite instability-high (MSI-H) tumors and rarely cooccur with actionable oncogenic drivers for which approved immunotherapy and targeted therapy options exist. Although our study did not exclude patients with MSI-H or deficient mismatch repair (dMMR) colorectal cancer, all patients enrolled had MSS tumors. While we were not able to specifically test the efficacy of adagrasib plus cetuximab in KRAS^G12C-mutated colorectal cancer with MSI-H/dMMR, we expect that this subset would also benefit from this combination, similar to the success of BRAF^V600E-targeted therapy in MSI-H and MSS colorectal cancer. The efficacy of second-line adagrasib plus cetuximab versus standard of care is being investigated in the phase III KRYSTAL-10 trial (NCT04793958; ref. 16). These findings emphasize the clinical importance of testing for KRAS^G12C mutations in patients with unresectable or metastatic colorectal cancer to identify those eligible for this combination therapy.

In heavily pretreated patients with unresectable or metastatic KRAS^G12C-mutated colorectal cancer, dual blockade of KRAS^G12C/EGFR by adagrasib plus cetuximab demonstrated encouraging clinical activity, was well tolerated, and is now recommended in the NCCN Guidelines. These data support potential registration of adagrasib plus cetuximab as a late-line treatment for patients with unresectable or metastatic KRAS^G12C-mutated colorectal cancer.

This open-label, nonrandomized, multicenter, multiple-expansion cohort phase I/II study was designed by employees of the sponsor, Mirati Therapeutics, a wholly owned subsidiary of Bristol Myers Squibb Company, and the investigators. The study was conducted in accordance with the principles of the Declaration of Helsinki and International Council for Harmonisation Good Clinical Practice guidelines (37, 38). The protocol was approved by the relevant institutional review boards or ethics committees; all patients provided written informed consent. Trial oversight was provided by the sponsor, the investigators, local institutional review boards, a specifically commissioned central institutional review board, and an independent data monitoring committee.

Full study details have been published previously (4, 5, 8, 39). The KRYSTAL-1 colorectal cancer cohorts included phase I and phase II cohorts in which the safety and efficacy of adagrasib plus cetuximab were assessed in patients with previously treated unresectable or metastatic KRAS^G12C-mutated colorectal cancer. Patients received adagrasib 600 mg orally twice daily and cetuximab intravenously weekly (400 mg/m² loading dose followed by 250 mg/m²; phase I) or every 2 weeks (500 mg/m²; phase I and phase II) until disease progression, unacceptable toxicity, withdrawal of consent, or death. Patients experiencing clinical benefit per the investigator could continue treatment despite confirmed radiographic progression. For the phase II adagrasib monotherapy cohort, study design, eligibility criteria, and endpoints are detailed in the Supplementary Methods.

Patients (aged ≥18 years of age) with previously treated, histologically confirmed, unresectable or metastatic KRAS^G12C-mutated colorectal cancer and Eastern Cooperative Oncology Group performance status (ECOG PS) of 0–1 were eligible for inclusion. Patients enrolled in the phase I and phase II cohorts had no available treatment with curative intent and were ineligible for, declined, or had no available standard of care treatment options; patients enrolled in the phase II expansion cohort had previously had to have received fluoropyrimidine, irinotecan, oxaliplatin, and a VEGF/VEGFR inhibitor. Patients with untreated or active brain metastases were excluded; patients with stable, treated brain metastases were eligible. Patients with prior therapy targeting KRAS^G12C mutations were excluded in the phase II portion.

The primary endpoint of the phase I portion was safety. The primary endpoint of the phase II portion was confirmed ORR as assessed by BICR according to RECIST v1.1. Secondary endpoints of the phase I portion included ORR, DOR, and PFS, all per BICR, as well as OS. Secondary endpoints of the phase II portion included DOR and PFS, both per BICR, OS, and safety. KRAS^G12C mutations were identified in tumor tissue (phase I and phase II) and/or ctDNA (phase I) by preapproved methods (PCR or NGS platforms) and laboratories. Additional baseline genomic alterations were assessed from anonymized testing reports that had been uploaded into the electronic data capture system. The reports were reviewed by the sponsor for the presence of certain genomic alterations, namely TP53, PIK3CA, BRAF, EGFR, NTRK, and HER2 (ERBB2), and for MSI/MMR status. Disease was evaluated at baseline and every 6 weeks during the study by CT scan or MRI. Patients with tumor response had a confirmatory assessment at least 4 weeks after the initial response. AEs were monitored throughout the study until 28 days after the last dose of study treatment and were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events version 5.0.

Serial plasma samples were collected (before C1D1 treatment and at the beginning of C2 and C4) and analyzed using droplet digital PCR to assess the effect of treatment on ctDNA clearance of KRAS^G12C-variant alleles. Cycle length was defined as 28 days. Clearance of KRAS^G12C ctDNA was defined as more than 90% decrease in the variant allele fraction based on prior defined thresholds (15). These data were only collected for the phase I cohort and are not available for the phase II cohort.

Targeted hybrid capture NGS (Agilent Resolution ctDx FIRST 109 gene panel; https://www.resolutionbio.com/assays/Resolution-ctDx-FIRST.html) was perfor­med on ctDNA obtained at baseline and EOT. Acquired genomic alterations were characterized for pathogenicity and defined as absent at baseline and present at EOT. All clearly inactivating mutations for known tumor suppressor genes were included. Well-established, annotated, clearly recurrent mutations were confirmed by COSMIC. For inclusion, point mutations that were potential variants of unknown significance required evidence of recurrence in COSMIC (≥5 instances) plus structural impact assessment by SIFT and mutation assessor.

Subgroup analyses of ORR (per BICR) were conducted for the following baseline characteristics: sex; age (<65 vs. ≥65); number of prior systemic therapies (≤2 vs. >2); baseline ECOG PS (0 vs. 1); presence of liver, lung, brain, bone, or adrenal metastases at baseline (yes vs. no for each metastatic site); PIK3CA and TP53 status (positive vs. negative); and patients treated with prior fluoropyrimidine, oxaliplatin, irinotecan, and VEGF/VEGFR inhibitor (all treatments; yes vs. no).

In the phase I cohort, seven patients were initially enrolled to receive adagrasib plus cetuximab; if the number of patients with dose-limiting toxicity observed among these patients warranted further enrollment, approximately 30 patients could be enrolled to fully characterize safety. In the phase II combination cohort, a sample size of 60 patients was planned to demonstrate a target ORR of ≥30% with a 95% CI excluding the ORR of <5% for standard of care in this setting (Clopper–Pearson method). The safety population included all patients who received ≥1 dose of any study treatment. The full analysis set included all patients with measurable disease at baseline who received ≥1 dose of adagrasib plus cetuximab. ORR was determined by BICR according to RECIST v1.1. DOR, PFS, and OS were estimated using the Kaplan–Meier method.

The Bristol Myers Squibb data sharing policy (https://www.bms.com/researchers-and-partners/independent-research/data-sharing-request-process.html) is compliant with ICMJE guidelines. Bristol Myers Squibb will honor legitimate requests for clinical trial data from qualified researchers. Data will be shared with external researchers whose proposed use of the data has been approved. Complete deidentified patient data sets will be eligible for sharing 2 years after completion of the study. Before data are released, the researcher(s) must sign a Data Sharing Agreement, after which the deidentified and anonymized datasets can be accessed within a secured portal.

R. Yaeger reports grants and personal fees from Mirati Therapeutics; grants from Pfizer, Boehringer Ingelheim, Daiichi Sankyo, and Boundless Bio; personal fees from Zai Lab, Loxo@Lilly, and Revolution Medicine; and other support from Amgen outside the submitted work. N. Uboha reports grants and personal fees from Ispen and Arcus; grants from Gilead and EMD Serono; personal fees from AstraZeneca, Astellas, Bristol Myers Squibb, Eli Lilly, Elevation Oncology, Strata, and personal fees from Eisai outside the submitted work. M.S. Pelster reports grants and other support from Arcus Biosciences, CytomX, and Elevation Oncology; grants from Actuate Therapeutics, Agenus, Astellas, BeiGene, BioNTech, Bristol Myers Squibb, Codiak Biosciences, Compass Therapeutics, Eisai, Exelixis, Gilead, HiberCell, Immune-Onc Therapeutics, Impact Therapeutics, Leap Therapeutics, Novartis, OncXerna Therapeutics, Panbela Therapeutics, Revolution Medicines, Surface Oncology, SQZ Biotechnologies, Tachyon Therapeutics, Translational Genomics, TransThera Sciences, ZielBio, 1200 Pharma; and grants from Bristol Myers Squibb during the conduct of the study; other support from AstraZeneca, Daiichi Sankyo, EMD Serono, Ipsen Biopharmaceuticals, Jazz Pharmaceuticals, Pfizer, SeaGen, and other support from Stemline Therapeutics outside the submitted work. T.S. Bekaii-Saab reports research funding (to institution) from Agios, Arys, Arcus, Atreca, Boston Biomedical, Bayer, Eisai, Celgene, Eli Lilly, Ipsen, Clovis, Seattle Genetics, Genentech, Novartis, Mirati, Merus, Abgenomics, Incyte, Pfizer, and Bristol Myers Squibb; consulting (to institution) from Servier, Ipsen, Arcus, Pfizer, Seattle Genetics, Bayer, Genentech, Incyte, Eisai, Merus, Merck KGaA, and Merck; consulting (to self) from Stemline, AbbVie, Blueprint Medicines, Boehringer Ingelheim, Janssen, Daiichi Sankyo, Natera, TreosBio, Celularity, Caladrius Biosciences, Exact Science, Sobi, BeiGene, Kanaph, AstraZeneca, Deciphera, Zai Labs, Exelixis, MJH Life Sciences, Aptitude Health, Illumina, Foundation Medicine; Sanofi; GSK, and Xilio; IDMC/DSMB from The Valley Hospital, Fibrogen, Suzhou Kintor, AstraZeneca, Exelixis, Merck/Eisai, PanCan, and 1Globe; is a scientific advisory board member for Imugene, Immuneering, Xilis, Replimune, Artiva, and Sun Biopharma; royalties: from UptoDate; in addition, T.S. Bekaii-Saab has patent for Human PD1 peptide vaccines and uses thereof (WO/2018/183488) licensed to Imugene; and a patent for Methods and compositions for the treatment of cancer cachexia (WO/2019/055687) licensed to Recursion. M. Barve reports personal fees from Tempus Labs and personal fees from Texas Oncology outside the submitted work. J.K. Sabari reports personal fees from AstraZeneca, Abbvie, Genentech, and F. Hoffmann-La Roche; grants and personal fees from Janssen, Loxo Lilly, Mirati, and Regeneron; personal fees from Jazz, Pfizer, Sanofi Genzyme, and personal fees from Takeda during the conduct of the study. J.A. Peguero reports grants and non-financial support from Mirati outside the submitted work. A.S. Paulson reports other support from Mirati during the conduct of the study; and other support from Mirati outside the submitted work; in addition, A.S. Paulson has an immediate family member employed by Day One Biopharmaceuticals; has stock ownership in Actinium; has received honoraria from Curio Science and Cardinal Health; has advisory role or received consulting fees from Amgen, Bristol Myers Squibb, Ipsen, Advanced Accelerator Applications, Incite, Exelixis, Pfizer, QED Therapeutics, Lilly Pharmaceuticals, Mirati, HutchMed, Astellas, Aadi, EMD Serono, AstraZeneca, Servier, Novartis, Array, Bayer, Jazz Pharmaceuticals, Takeda, Seagen, Agenus, Ideo Oncology; has received institutional research funding from Buzzard Pharmaceuticals, Ipsen, Bristol Myers Squibb, Exelixis, HutchMed, Taiho Pharmaceutical, Eli Lilly, AstraZeneca, Incyte, Deciphera, G1 Therapeutics, Zentalis, Tempus, Camurus, Relay Therapeutics, Nucana, Merck, Bayer, Seattle Genetics, Sotio, Innovative Cellular Therapeutics Co., Regenxbio, Gilead Sciences, Gritstone Bio, BioNTech SE, Novartis, TransThera Biosciences, ITM Oncologics, Inspirna; and has received travel, accommodations, expenses from Pfizer, Camurus, Mirati Therapeutics, Nucana, and AADi. P.A. Jänne reports grants and personal fees from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Daiichi Sankyo, and Takeda Oncology; personal fees from Pfizer, F. Hoffmann-La Roche/Genentech, Chugai Pharmaceuticals, SFJ Pharmaceuticals, Voronoi, Biocartis, Novartis, Sanofi Oncology, Transcenta, Silicon Therapeutics, Syndax, Nuvalent, Bayer, Eisai, Alllorion Therapeutics, Accutar Biotech, AbbVie, Mone Rosa, Scorpion Therapeutics, Merus, Frontier Medicines, Hongyun Biotechnology, Duality Biologics, Blueprint Medicines; and personal fees from Mirati Therapeutics during the conduct of the study; grants from Revolution Medicines, and grants from PUMA outside the submitted work; in addition, P.A. Jänne has a patent for EGFR mutations issued and licensed to Lab Corp. M. Cruz-Correa reports grants from Mirati during the conduct of the study; other support from Pan American Center for Oncology Trials; grants from Pfizer, AbbVie, Bristol Myers Squibb, Merck, Genentech, Huyabio, Janssen, Natera, and grants from SeaGen outside the submitted work; is a governing board member of the American Association for Cancer Institute; and is an executive board member of Precision Oncology Alliance. K. Anderes reports personal fees from Mirati Therapeutics outside the submitted work. H. Der-Torossian reports other support from Mirati Therapeutics/BMS outside the submitted work. S.J. Klempner reports personal fees from Astellas, Novartis, Amgen, Pfizer, Natera, AstraZeneca, Sanofi-Aventis, Daiichi-Sankyo, Mersana, Merck, and personal fees from Bristol Myers Squibb outside the submitted work; and is an NCCN Guideline committee member (no compensation). S.E. Kopetz reports other support from Lutris, Frontier Medicines, Xilis, Navire, Genentech, EMD Serono, Merck, Holy Stone Healthcare, Novartis, Eli Lilly, Boehringer Ingelheim, AstraZeneca/MedImmune, other support from Bayer Health, Redx Pharma, Ipsen, HalioDx, Lutris, Jacobio, Pfizer, Repare Therapeutics, Inivata, GSK, Jazz Pharmaceuticals, Iylon, Xilis, AbbVie, Amal Therapeutics, Gilead Sciences, Mirati Therapeutics, Flame Biosciences, Servier, Carina Biotech, Bicara Therapeutics, Endeavor BioMedicines, Numab, Johnson & Johnson/Janssen, Genomic Health, Frontier Medicines, Replimune, Taiho Pharmaceutical, Cardiff Oncology, Ono Pharmaceutical, Bristol Myers Squibb-Medarex, Amgen, Tempus, other support from Foundation Medicine, Harbinger Oncology, Takeda, CureTeq, Zentalis, Black Stone Therapeutics, NeoGenomics, Accademia Nazionale Di Medicina, Tachyon Therapeutics, Sanofi, Biocartis, Guardant Health, Array BioPharma, F. Hoffmann-La Roche/Genentech, EMD Serono, MedImmune, Novartis, Amgen, Eli Lilly, and other support from Daiichi Sankyo outside the submitted work. No disclosures were reported by the other authors.

R. Yaeger: Data curation, investigation, writing–original draft, writing–review and editing. N. Uboha: Resources, data curation, investigation, writing–original draft, project administration, writing–review and editing. M.S. Pelster: Data curation, investigation, writing–original draft, writing–review and editing. T.S. Bekaii-Saab: Conceptualization, resources, writing–original draft, writing–review and editing. M. Barve: Resources, investigation, writing–original draft, writing–review and editing. J. Saltzman: Resources, writing–original draft, writing–review and editing. J.K. Sabari: Conceptualization, supervision, investigation, writing–original draft, writing–review and editing. J.A. Peguero: Investigation, writing–original draft, writing–review and editing. A.S. Paulson: Resources, data curation, supervision, validation, investigation, writing–original draft, project administration, writing–review and editing. P.A. Jänne: Resources, writing–original draft, writing–review and editing. M. Cruz-Correa: Supervision, investigation, writing–original draft, project administration, writing–review and editing. K. Anderes: Conceptualization, investigation, writing–original draft, project administration, writing–review and editing. K. Velastegui: Data curation, investigation, writing–original draft, writing–review and editing. X. Yan: Formal analysis, writing–original draft, writing–review and editing. H. Der-Torossian: Conceptualization, visualization, methodology, writing–original draft, writing–review and editing. S.J. Klempner: Resources, data curation, investigation, writing–original draft, writing–review and editing. S.E. Kopetz: Conceptualization, resources, data curation, supervision, investigation, writing–original draft, writing–review and editing.

We thank the patients, their families, their caregivers, the study investigators, and their team members at each site for participation in the ongoing KRYSTAL-1 trial, Skye Sully, Senior Director, Clinical Operations, of Mirati Therapeutics, a wholly owned subsidiary of Bristol Myers Squibb Company, for her support in data delivery and Cloud Paweletz, Head of Research, Belfer Center for Applied Cancer Science, Dana Farber Cancer Institute for his support in ctDNA data analysis and delivery.