Advances on drug therapy for KRAS^G12C-mutant non-small-cell lung cancer

Highlights

• •This article highlights the research progress and clinical trial results of KRAS^G12C targeted inhibitors.
• •The resistance mechanism and research progress of KRAS^G12C targeted inhibitors are described.
• •Combination treatment strategies and novel therapies for KRAS^G12C-mutant NSCLC.

Abstract

Lung cancer has an extremely high mortality rate among malignant tumors, posing a significant threat to human health Among all lung cancer cases, non-small cell lung cancer (NSCLC) accounts for a significant proportion and has become a hot topic in clinical research and treatment. The Kirsten rat sarcoma viral oncogene homolog (KRAS) is one of the most common oncogenic drivers in NSCLC, closely associated with tumor initiation, treatment response, and prognosis. However, due to the relatively smooth surface of the KRAS protein and the lack of drug-binding pockets, it has long been regarded as an "undrugable target". With further research, recently, targeted drugs targeting the KRAS^G12C gene mutation have achieved significant breakthroughs in clinical trials, especially the application of KRAS^G12C-specific inhibitors adagrasib and sotorasib, which has changed the treatment landscape for NSCLC patients. To address challenges such as tumor heterogeneity, the complexity of the tumor microenvironment, interpatient variability, and acquired drug resistance mechanisms, combination therapy strategies involving KRAS^G12C inhibitors have emerged sequentially. This article systematically reviews the progress of targeted therapy for KRAS^G12C-mutant NSCLC and the results of related clinical trials, while exploring novel therapeutic strategies for patients with KRAS^G12C mutations, aiming to provide a reference for the selection of clinical treatment regimens.

Introduction

Lung cancer is the leading cause of mortality among all solid tumors worldwide, with approximately 1.6 million deaths annually. NSCLC accounts for about 80–85 % of all lung cancer pathological types[1]. Currently, the main treatment modalities for NSCLC include surgery, chemotherapy, radiotherapy, immunotherapy, and targeted therapy. However, nearly 75 % of NSCLC patients are diagnosed at an advanced stage upon initial presentation, and the 5-year relative survival rate for patients diagnosed with metastatic lung cancer is only about 6 %[2]. Traditional surgery, radiotherapy, and chemotherapy have limited efficacy in these patients. With advances in molecular biology and next-generation sequencing (NGS) technology, a series of therapeutic genetic targets have been identified, such as EGFR, ALK, ROS1, MET, PIK3CA, RET, KRAS, BRAF. This targeted therapy against the driving oncogenes has changed the treatment landscape of NSCLC and improved the survival rate of lung cancer patients.

The RAS oncogene is a common oncogenic driver factor across multiple cancer types. KRAS is its most frequently mutated subtype (about 86 %), predominantly occurring in patients with NSCLC (about 20.4 %), pancreatic ductal adenocarcinoma (PDAC) (about 67.6 %), and colorectal cancer (CRC) (about 35.7 %), and represents a major oncogenic driver in NSCLC. KRAS mutations are primarily associated with smoking history and correlate with poor prognosis[3]. However, due to the structure of the encoded KRAS protein being approximately spherical with a molecular weight of only 21KD and having no obvious binding pockets, and having a high affinity for GTP (picomolar level), making it difficult to interfere with GTP binding using GTP analogs[4], Consequently, the development of targeted therapies against KRAS mutations has progressed extremely slowly. Therefore, the breakthrough discovery of KRAS inhibitors is an important milestone. Among these, sotorasib (AMG-510) and adagrasib (MRTX-849) have received the U.S. Food and Drug Administration (FDA) approval and demonstrated clinical efficacy in patients with KRAS^G12C-mutant NSCLC. These agents covalently and irreversibly bind to the cysteine residue of KRAS^G12C, thereby locking the KRAS protein in its inactive or "OFF" GDP-bound state and preventing the activation of downstream signaling pathways. However, with advancing research, their resistance mechanisms and limitations have become increasingly apparent. In this narrative review, we outline the structure and function of the KRAS protein, summarize the current progress in targeted therapy for KRAS^G12C-mutant NSCLC, and discuss future therapeutic prospects and limitations, aiming to provide a reference for clinical decision-making.

1. Structure and function of KRAS proteins

The KRAS gene was first discovered in 1982 and belongs to the RAS oncogene family along with the neuroblastoma rat sarcoma virus (NRAS) and Harvey rat sarcoma virus (HRAS), located on the short arm of chromosome 12. It encodes two alternatively spliced proteins: KRAS 4A and KRAS 4B Structurally, the KRAS protein consists of two main regions: the N-terminal G-domain and the C-terminal hypervariable region (HVR). The G-domain includes the effector region and the allosteric region, which contain the binding sites for nucleotides, effector proteins, and regulatory proteins. The effector region is composed of the P-loop and the switch regions (Switch I and II), the latter being an important area for mediating protein-protein interactions[5].

Under physiological conditions, the KRAS protein can be activated by various stimuli including epidermal growth factor (EGF), fibroblast growth factor (FGF), platelet-derived growth factor (PDGF), insulin-like growth factor (IGF), and many other cytokines. It encodes a G protein with intrinsic guanosine triphosphatase (GTPase) activity, acting downstream of receptor tyrosine kinases (RTKs) and inducing the activation of mitogen-activated protein kinase (MAPK/ERK), phosphatidylinositol 3-kinase (PI3K) pathways, and many other keys signaling pathways. The RAF-MEK-ERK pathway is a classic downstream signaling pathway of KRAS. The activation of KRAS can induce the phosphorylation of MEK1/2 and ERK1/2, thereby regulating the transcription and translation of specific genes and influencing cell proliferation and differentiation. The PI3K-AKT-mTOR signaling pathway is crucial for cell growth, glucose transport and metabolism, and apoptosis. KRAS can activate AKT and mTOR through PI3K. The activation of mTOR protein can drive cell proliferation. The phosphorylation of AKT and the activation of Bcl-XL/Bcl-2 associated death promoter can inhibit cell apoptosis. Additionally, RAL guanine nucleotide dissociation stimulator, as a downstream signaling protein of KRAS, acts as a GTP/GDP exchange factor, promoting the conversion from RAL GDP to RAL GTP. KRAS can also regulate TIAM1 and RAC-specific guanine nucleotide exchange factors, activating the RAC1 signaling pathway, which affects cell shape, migration, adhesion, actin cytoskeleton formation, endocytosis, and membrane transport[6].

Evidence suggests that different KRAS isoforms exhibit varying degrees of dependence on these major downstream pathways. It has been reported that KRAS-mediated lung tumorigenesis requires elevated KRAS expression and/or altered expression/activity of key tumor suppressors (such as APC, ARF, INK4A, LKB1, TP53, or PTEN). In a genetically engineered mouse (GEM) model of KRAS^G12D-induced lung cancer, researchers demonstrated that PI3Kα (PIK3CA) acts as a direct effector of KRAS, and can enhance KRAS-driven lung tumorigenesis through additional activation of PI3′-lipid signaling[7]. Furthermore, an analysis of molecularly targeted clinical trial data from 215 patients with refractory NSCLC revealed that the phosphatidylinositol 3-kinase (PI3K) and mitogen-activated protein/extracellular signal-regulated kinase (MEK) signaling pathways are activated in KRAS^G12D-mutant NSCLC cell lines. In contrast, the RAL signaling pathway is activated and Akt activation is reduced in KRAS^G12C- or KRAS^G12V-mutant NSCLC cell lines[8]. Researchers propose that this preference for specific downstream pathways holds important clinical implications for the selection of monotherapy or combination therapy regimens in cancer treatment[9].

Under physiological conditions, the KRAS protein is in a dynamic cycle of "on/off" states on the cell membrane through the exchange and hydrolysis of GTP/GDP, and it regulates cell proliferation, apoptosis, differentiation, survival, and migration. Inside the cell, the KRAS protein releases GDP under the action of guanine nucleotide-exchange factor (GEF) and rapidly binds to GTP, existing in the active KRAS-GTP form. The KRAS protein itself has GTPase activity, and its activity is enhanced under the regulation of GTPase-activating proteins (GAPs), thereby converting KRAS-GTP to the inactive KRAS-GDP. However, KRAS point mutations lead to a decrease in the intrinsic GTPase and GAPs activities, causing the gradual accumulation of KRAS-GTP, over-activating downstream signaling pathways, and thus driving tumor growth[10].

2. Treatment progress of KRAS^G12C inhibitors in NSCLC

According to epidemiological data, among patients with NSCLC in China, KRAS missense mutations represent the second most prevalent oncogenic driver mutation following epidermal growth factor receptor (EGFR) mutations, observed in approximately 25 % of lung adenocarcinoma cases and in about 3 % of squamous cell carcinoma patients[11]. In Western populations, KRAS mutations are detectable in approximately one-quarter to one-third of NSCLC patients, whereas the overall prevalence of KRAS mutations in the Chinese population ranges from 9.8 % to 12.3 %. The most frequently affected codons are G12, G13, and Q61, among which KRAS^G12C is the most common mutation, accounting for approximately 40 % of all KRAS mutations, followed by G12D, G12V, and G12A[12]. However, due to the absence of classic drug-binding pockets on the KRAS protein, the development of targeted therapies against KRAS has been historically challenging. Over the past 40 years, numerous studies have focused on exploring pathways that directly or indirectly target KRAS-mutated cancers, including KRAS expression, processing, upstream regulators or downstream effectors, and RAF, MEK, or PI3K/AKT inhibitors, but with little success.

2.1. Direct targeted therapy of KRAS^G12C

With the in-depth study of the molecular structure and biological properties of the KRAS protein, in 2013, researchers such as Shokat obtained small molecule compounds targeting KRAS^G12C after screening and optimizing disulfide fragments. This compound can covalently and irreversibly bind to the switch-II site which exposed in the non-active KRAS-GDP configuration, locking KRAS in its inactive GDP-bound state, thereby preventing its binding to GTP and inactivating its downstream effectors[13]. Subsequently, through a series of structural optimizations of this small molecule compound, KRAS^G12C inhibitors gradually entered clinical research. Currently, sotorasib and adagrasib have been approved by FDA as second-line treatments for patients with locally advanced or metastatic NSCLC (mNSCLC)[14,15].

2.1.1. Sotorasib (Lumakras/AMG-510)

Sotorasib, as the first clinically effective targeted inhibitor for KRAS^G12C mutations approved by the FDA, is an oral small molecule drug that can covalently bind to the cysteine at the switch II site, locking the KRAS protein in the inactive KRAS-GDP state and inhibiting the signal transduction mediated by KRAS. CodeBreak 100 is an I/II phase clinical study aimed at evaluating the safety, pharmacokinetics, and efficacy of sotorasib in patients with locally advanced or metastatic KRAS^G12C tumors who have previously received first-line treatment (see Table 1 for details).In the I phase trial, a total of 129 patients with advanced solid tumors were enrolled in dose-escalation and dose-expansion cohorts, with safety as the primary endpoint. As of the data cutoff date of June 1, 2020, no dose-limiting toxicities (DLTs) or treatment-related deaths were observed. Adverse events (AEs) of any cause during treatment were reported in 96.9 % (125/129) of patients. The most common AEs included diarrhea (29.5 %), fatigue (23.3 %), and nausea (20.9 %). Treatment discontinuation occurred in 82.9 % (107/129) of patients, with disease progression being the most frequent reason for discontinuation. Pharmacokinetic results showed that in the 960 mg dose group, the maximum plasma concentration (Cmax) was 7.50 μg/mL (coefficient of variation (CV): 98.3 %), and the time to reach maximum concentration (Tmax) was 2.0 h (range: 0.3–6.0 h). The elimination half-life was 5.5 ± 1.8 h. A total of 59 patients were enrolled in the NSCLC subgroup. The study demonstrated a response rate of 32.2 % across all dose levels and 35.3 % at the target dose of 960 mg. However, disease progression occurred shortly after the initial response in some patients[16].

Table 1.

Clinical trials of marketed KRAS^G12C inhibitors in NSCLC.

Clinical trial name	Study phase	KRASG12C inhibitor	NSCLC patients	Median follow-up time (months)	Current or former smoker ( %)	Previous anti–PD-1 or anti–PD-L1 therapies ( %)	Previous platinum-based chemotherapy ( %)	Median duration of treatment (months/weeks)	DCR ( %)	ORR ( %)	mDOR (months)	mPFS (months)	mOS (months)	CNS activity (yes/no)	TRAEs( %)	TRAEs ≥3( %)	FDA (other) approval date
CodeBreak 100[16] NCT03600883	Phase 1	Sotorasib, 180/360/720/960 mg, QD	59	11.7 (4.8–21.2)	89.8 % (53/59)	89.8 % (53/59)	100.0 %	3.9 months (0–16.6)	88.1 (95 %CI: 77.07–95.09)	32.2 (95 %CI: 20.62–45.64)	10.9 (1.1+ - 13.6)	6.3 (0+ – 14.9)	NA	NA	56.6 %	11.6 %	NA
CodeBreak 100[17] NCT03600883	Phase 2	Sotorasib, 960 mg, QD	126	15.3 (1.1–18.4+)	92.9 % (117/126)	92.1 % (116/126)	89.7 % (113/126)	5.5 months (0.2–17.8)	80.6 (95 %CI: 72.6–87.2)	37.1 (95 %CI: 28.6–46.2)	11.1 (95 %CI: 6.9–NE)	6.8 (95 %CI: 5.1–8.2)	12.5 (95 %CI: 10.0–NE)	NA	69.8 %	20.6 %	FDA, USA, 2021/5/28
CodeBreak 100[18] NCT03600883	Phase I II	Sotorasib, 960 mg, QD	174	NA	NA	90 % (157/174)	93 % (161/174)	5.6 months (0.2–35.9)	84 (95 %CI: 77.3–88.9)	41 (95 %CI: 33.3–48.4)	12.3 (95 %CI: 7.1–15.0)	6.3 (95 %CI: 5.3–8.2)	12.5 (10.0–17.8)	Yes, 16 evaluable patients with brain metastases, 3 achieved CR, 11 achieved SD	70 %	21 %	NA
CodeBreak 200[19] NCT04303780	Phase 3	Sotorasib, 960 mg, QD vs docetaxel, 75 mg/m2, Q3W	171 vs 174	17.7 (IQR: 16.4–20.1)	97.1 % (166/171) vs 95.4 % (166/174)	21.1 % (36/171) vs 12.1 % (21/174)	8.2 % (14/171) vs 5.7 % (10/174)	19.9 weeks (0.4–101.3) vs 12.0 weeks (3.0–101.0)	82.5 (95 %CI: 75.9–87.8) vs 60.3 (95 %CI: 52.7–67.7)	28.1 (95 %CI: 21.5–35.4) vs 13.2 (95 %CI: 8.6–19.2)	8.6 (95 %CI: 7.1–18.0) vs 6.8 (95 %CI: 4.3–8.3)	5.6 (95 %CI: 4.3–7.8) vs 4.5 (95 %CI: 3.0–5.7)	10.6 (95 %CI: 8.9–14.0) vs 11.3 (95 %CI: 9.0–14.9)	Yes, median time to CNS disease recurrence 15.8 months (95 %CI: 9.7–NE) vs 10.5 (95 %CI: 5.8–NE)	98 % vs 98 %	33 % vs 40 %	NA
KRYSTAL-1[22] NCT03785249	Phase 1/1b	Adagrasib, 300/600/1200 mg (QD) or 600 mg (BID)	18	19.6	NA	NA	NA	NA	NA	53.3 (95 %CI: 26.6–78.7)	16.4+ (95 %CI: 3.1–NE)	11.1 (95 %CI: 2.6–NE)	NR	NA	92 %	36 %	NA
KRYSTAL‐1[23] NCT03785249	Phase 2	Adagrasib, 600 mg, BID	116	12.9 (95 %CI: 11.8–13.5)	95.7 % (111/116)	98.3 % (114/116)	100 %	5.7 months (0.03–19.6)	79.5 (95 %CI: 70.8–86.5)	42.9 (95 %CI: 33.5–52.6)	8.5 (95 %CI: 6.2–13.8)	6.5 (95 %CI: 4.7–8.4)	11.7 (95 %CI: 9.2–NE)	Yes; ORR: 33.3 % (95 %CI: 18.0–51.8)	97.4 %	44.8 %	FDA, USA, 2022/12/12
KRYSTAL‐12[25] NCT04685135	Phase 3	Adagrasib, 600 mg, BID vs docetaxel, 75 mg/m2, Q3W	301 vs 152	7.2 (95 %CI: 5.8–8.7)	94 % (284/301) vs 93 % (142/152)	100 %	100 %	3.9 months (IQR: 1.8–7.6) vs 2.7 months (1.4–4.2)	78 vs 59	31.9 (95 %CI: 26.7–37.5) vs 9.2 (95 %CI: 5.1–15.0)	8.3 (95 %CI: 6.1–10.4) vs 5.4 (95 %CI: 2.9–8.5)	5.5 (95 %CI: 4.5–6.7) vs 3.8 (95 %CI:2.7–4.7)	NA	Yes; ORR: 24 % vs 11 %	94 % vs:86 %	47 % vs 46 %	NA
NCT05005234[27]	Phase Ia/Ib	Fulzerasib, 250/450/700/900 (QD) or 450/600/750 mg (BID)	166	8.2 (95 %CI: 6.9–8.3)	82.5 % (137/166)	58.4 % (97/166)	81.3 % (135/166)	NA	92.1 (95 %CI: 86.9–95.7)	45.5 (95 %CI: 37.7–53.4)	NR	9.6 (95 %CI: 6.8–16.6)	NR	Yes; ORR: 27.6 % (95 %CI:12.7–47.2), CR: 17.2 %	95.8 %	36.7 %	NA
NCT05005234[28]	Phase 2	Fulzerasib, 600 mg, BID	116	15.1 (95 %CI: 14.8–15.7)	79.3 % (92/116)	87.1 % (101/116)	97.4 % (113/116)	201 days (range: 1–525)	90.5 (95 %CI: 83.7–95.2)	48.3 (95 %CI: 38.9–57.7)	NR	9.7 (95 %CI: 5.6–11.0)	NA	Yes; ORR: 48.6 % (95 %CI:31.4–66.0)	92.2 %	41.4 %	NDA, China, 2024/8/21
NCT05383898[30]	Phase 1	Garsorasib, 600/800/1200 mg (QD) or 400/600 mg (BID)	79	8.8 (95 %CI: 0.7–14.9)	70.9 % (56/79)	69.6 % (55/79)	93.7 % (74/79)	NA	91.9 (95 %Cl: 83.2–97.0)	40.5 (95 %Cl: 29.3–52.6)	7.1 (95 %CI: 6.2–NA)	8.2 (95 %CI: 7.5–NA)	NA	Yes; ORR: 17 %, DCR: 100 %.	94.9 %	38.0 %	NA
NCT05383898[31]	Phase 2	Garsorasib, 600 mg, BID	123	7.9 (IQR: 6.3–10.4)	80 % (98/123)	96 % (118/123)	98 % (120/123)	6.2 months (IQR: 3.2–8.6)	89 (95 %CI: 82–94)	50 (95 %CI: 41–59)	12.8 (95 %CI: 6.2–NE)	7.6 (95 %CI: 5.6–9.7)	NR	NA	95 %	50 %	NDA, China, 2024/11/8
NCT05009329[32]	Phase 2	Glecirasib, 800 mg, QD	119	10.4 (95 %CI: 9.5–11.0)	59.7 % (71/119)	94.1 % (112/119)	94.1 % (112/119)	NA	86.3 (95 %CI: 78.7–92.0)	47.9 (95 %CI: 38.5–57.3)	NE (95 %CI: 7.2–NE)	8.2 (95 %CI: 5.5–13.1)	13.6 (95 %CI: 10.9–NE)	NA	97.5 %	38.7 %	NDA, China, 2025/5/22

The phase II trial focused on evaluating the efficacy and safety of sotorasib as monotherapy in pretreated patients with KRAS^G12C-mutant NSCLC (see Table 1 for details). The primary endpoint was the objective response rate (ORR). As of the data cutoff date of March 15, 2021, a total of 126 patients were enrolled. Treatment discontinuation occurred in 81.7 % (103/126) of patients, with disease progression (65.9 %) and AEs (8.7 %) being the most common reasons for discontinuation. AEs were observed in 99.2 % of patients, with the most frequent including diarrhea (50.8 %), nausea (31.0 %), fatigue (25.4 %), and arthralgia (21.4 %). Additionally, 69.8 % of patients received sotorasib for 3 months or longer, and 47.6 % for 6 months or longer. Among 124 patients evaluable for response, tumor shrinkage of any degree was observed in 102 patients (82.3 %); the tumor burden (defined as the sum of the longest diameters of all target lesions) of all responders decreased by 60 % relative to baseline. These data confirm that sotorasib confers durable clinical benefits in patients with KRAS^G12C-mutant NSCLC.

Furthermore, in exploratory analyses, potential associations between efficacy and PD-L1 expression levels, tumor mutational burden (TMB), as well as co-mutations in STK11, KEAP1, and TP53—among the most common co-mutated genes in KRAS-mutant NSCLC—were evaluated. Among 86 evaluable patients, objective responses and tumor shrinkage were observed across all baseline PD-L1 expression levels. Among 84 evaluable patients, the response rates were 42 % (95 % confidence interval (CI): 30–55) in the low TMB group (<10 mutations per megabase) and 40 % (95 %CI: 16–68) in the high TMB group (≥10 mutations per megabase), with no significant difference. Among 104 evaluable patients, the ORR was 50 % (95 %CI: 28–72) in the STK11-mutant/KEAP1-wild-type subgroup. In contrast, the KEAP1-mutant subgroups exhibited poorer efficacy: the ORR was 23 % (95 %CI: 5–54) in patients with concurrent STK11 and KEAP1 mutations, and 14 % (95 %CI: 0–58) in the STK11-wild-type/KEAP1-mutant subgroup. However, these exploratory analyses did not reach statistical significance[17].

In 2023, the research team conducted a pooled analysis of data from this 2-year clinical trial: this represents the largest sample size and longest follow-up dataset to date for KRAS^G12C inhibitor therapy¹⁸ (see Table for details). The trial demonstrated that the long-term benefits of sotorasib (defined as progression-free survival (PFS) ≥ 12 months) were associated with low baseline circulating tumor DNA (ctDNA) levels. Additionally, benefits were observed across patients with varying KRAS^G12C variant allele frequencies and PD-L1 expression levels, with some patients harboring concurrent STK11 and/or KEAP1 co-mutations also deriving benefit. Furthermore, sotorasib was well-tolerated, with almost no late-onset treatment-related toxicities, none of which led to treatment discontinuation. These findings further confirm the favorable safety profile and durable efficacy of sotorasib in patients with KRAS^G12C-mutant NSCLC. Based on the CodeBreak 100 trial series, the FDA granted accelerated approval for sotorasib in May 2021 for the treatment of patients with advanced KRAS^G12C -mutant NSCLC who have received prior therapy. As the world’s first targeted agent against KRAS, this approval marks a landmark advancement.

As part of the post-marketing requirements to confirm the clinical benefits of sotorasib, CodeBreak 200—a randomized, open-label phase III clinical trial conducted across 148 centers in 22 countries—was initiated to compare the efficacy and safety of sotorasib versus docetaxel in patients with KRAS^G12C -mutant NSCLC who had progressed following prior treatment (see Table 1 for details). Study data demonstrated that the sotorasib group significantly improved PFS compared with the docetaxel group, reducing the risk of disease progression or death by 34 %. This benefit was consistent across subgroups defined by different demographic characteristics, Eastern Cooperative Oncology Group (ECOG) performance status, lines of prior therapy, PD-L1 expression levels, and presence or absence of central nervous system (CNS) involvement. In exploratory analyses, sotorasib delayed the median time to CNS disease recurrence compared with docetaxel (15.8 months (95 %CI: 9.7–not estimable) vs 10.5 months (5.8–not estimable)). Additionally, sotorasib showed clinically meaningful advantages in delaying the deterioration of overall health status, physical function, and cancer-related symptoms (dyspnea and cough)[19].

Despite the improved PFS demonstrated by sotorasib, analyses revealed that these results were subject to multiple systematic biases. Consequently, the FDA Oncologic Drugs Advisory Committee stated in October 2023 that the PFS benefit of sotorasib over docetaxel could not be reliably confirmed. Additionally, studies by Stratmann, Passiglia, and colleagues have reported findings consistent with the core results of the CodeBreak trials[20,21]. Currently, the first-line treatment potential of sotorasib is being further evaluated in the phase II CodeBreak 201 clinical trial (NCT04933695), which will provide critical evidence for its positioning in the first-line treatment of KRAS^G12C-mutant NSCLC.

2.1.2. Adagrasib (MRTX-849)

Adagrasib is another covalent inhibitor targeting the KRAS^G12C-mutant protein. It can selectively bind to the GDP-KRAS protein and locks it in an inactive state. Compared with sotorasib, adagrasib exhibits distinct pharmacokinetic advantages, including a long half-life (approximately 23 h, while the drug half-life of sotorasib is about 5 h), dose dependence, and CNS penetration. These properties render it a potential therapeutic option for patients with KRAS^G12C-mutant NSCLC and concurrent brain metastases. In the KRYSTAL-1 I/II[22], pharmacokinetic, efficacy, and safety data from the dose-escalation and dose-expansion cohorts demonstrated that adagrasib was well-tolerated with a favorable safety profile, exhibiting promising antitumor activity in patients with advanced solid tumors harboring KRAS^G12C mutations (see Table 1 for details). Pharmacokinetic results showed that following a single oral dose of 600 mg adagrasib, Tmax was 4.17 h (range: 2.0–10.10 h), and the elimination half-life was 23.0 h (range: 16.3–27.9 h). Additionally, preliminary data from 2 treatment-naive patients with CNS metastases in the phase 1b cohort revealed that the concentration of adagrasib in cerebrospinal fluid exceeded the half-maximal inhibitory concentration (IC50) for tumor cells, with a mean unbound brain-to-plasma concentration ratio of 0.47. These findings confirm its clinically meaningful CNS penetration.

Based on the integrated phase I trial results, the recommended phase II dose (RP2D) of adagrasib was determined to be 600 mg twice daily. The phase II trial[23] enrolled patients with KRAS^G12C-mutant NSCLC who had previously received platinum-based chemotherapy and immune checkpoint inhibitor therapy. The primary endpoint was ORR. As of October 15, 2021, AEs of any cause were reported in all 116 enrolled patients (100 %). The most common AEs included diarrhea (70.7 %), nausea (69.8 %), fatigue (59.5 %), vomiting (56.9 %), anemia (36.2 %), dyspnea (35.3 %), increased serum creatinine (34.5 %), and decreased appetite (31.9 %). Treatment was ongoing in 33.3 % (39/116) of patients, while 73.3 % (85/116) had discontinued treatment. The most frequent reasons for discontinuation were disease progression (26.7 %) and AEs (13.8 %). No new safety signals were identified, indicating a favorable safety profile of adagrasib. Among 48 patients who achieved a response, the median time to response was 1.4 months (range: 0.9–7.2 months), and the Kaplan-Meier-estimated duration of response (DOR) at 9 months was 48.4 % (95 %CI: 32.5–62.6), demonstrating durable antitumor activity.

In exploratory analyses, the confirmed objective response rates (cORR) in subgroups with co-mutations in STK11, KEAP1, TP53, and CDKN2A were 40.5 %, 28.6 %, 51.4 %, and 58.3 %, respectively. The response rates ranged from 35.7 % to 55.9 % in subgroups with STK11 mutations (regardless of KEAP1 status) and concurrent STK11/KEAP1 mutations. In contrast, the cORR was only 14.3 % in the subgroup with wild-type STK11 and KEAP1 mutations. Among 86 evaluable patients, the cORR was comparable across subgroups with different PD-L1 expression levels (41.7 % to 46.8 %), indicating that efficacy was not influenced by PD-L1 expression status. Consequently, adagrasib received accelerated approval from FDA on December 12, 2022.

Based on the aforementioned results, the KRYSTAL-12 phase III clinical trial was designed to evaluate the efficacy of adagrasib versus docetaxel in previously treated patients with advanced KRAS^G12C-mutant NSCLC (see Table 1 for details)[24,25]. As of the data cutoff date of December 31, 2023, 37 % (111/300) of patients in the adagrasib group remained on treatment, compared with 11 % (17/150) in the docetaxel group. Study data demonstrated that the adagrasib group had a significantly longer time to deterioration in Lung Cancer Symptom Scale (LCSS) scores than the docetaxel group, with a median of 3.0 months (95 %CI: 2.7–4.1) versus 1.5 months (1.3–1.9) (hazard ratio (HR) = 0.57 (95 %CI: 0.45–0.74)). Adagrasib showed clinically meaningful advantages in delaying the deterioration of core symptoms such as dyspnea and cough. The study found that regardless of whether prior chemotherapy and immunotherapy were administered concurrently or sequentially, the adagrasib group achieved a statistically significant improvement in PFS compared with the docetaxel group, significantly reducing the risk of disease progression or death. No new safety signals were observed, and adagrasib was well-tolerated, including in patients with baseline brain metastases (BM). Additionally, the median time to intracranial progression (TIP) was 18.6 months (95 %CI: 9.6–NE) in the adagrasib group (78 patients with baseline BM) versus NE (95 %CI: 4.2–NE) in the docetaxel group (36 patients with baseline BM) (HR = 0.60 (95 %CI: 0.26–1.40)). The HR for intracranial PFS was 0.93 (95 %CI: 0.50–1.73), indicating a potential advantage in intracranial disease control. Currently, the clinical development of adagrasib focuses on dose optimization and first-line treatment exploration: KRYSTAL-21 (NCT05853575) aims to compare the efficacy and safety of 600 mg twice daily versus 400 mg BID; KRYSTAL-7 (NCT04613596) evaluates the potential of adagrasib monotherapy or combination with pembrolizumab as first-line treatment; and KRYSTAL-4 (NCT06875310) explores the efficacy of adagrasib in combination with pembrolizumab and chemotherapy for first-line therapy.

Notably, given the high incidence of BM (approximately 40 %) in patients with KRAS^G12C-mutant NSCLC, a post-hoc analysis of 42 patients with baseline CNS metastases was conducted in the KRYSTAL-1 trial using the modified Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) criteria. Among 33 radiologically evaluable patients, the cORR was 33.3 % (95 %CI: 18.0–51.8), the median intracranial DOR was 11.2 months (95 %CI: 2.99–NE), and the median intracranial PFS was 5.4 months (95 %CI: 3.3–11.6). These results are consistent with the phase 1b data (intracranial ORR = 42 %, iDOR = 12.7 months), confirming adagrasib’s favorable intracranial activity in patients with untreated brain metastases. In contrast, the CNS activity of sotorasib remains unclear, as patients with active, untreated brain metastases were excluded from its initial trials. However, a post-hoc analysis of CodeBreaK 100 showed that 88 % (n = 16) of patients achieved intracranial disease control, with 2 patients achieving complete response (CR). Differently, Divyan and colleagues conducted an indirect comparison of the safety and efficacy of sotorasib and adagrasib through matching key clinical trial data. The results demonstrated that both agents were similarly effective in slowing cancer progression and shrinking tumors. However, sotorasib had a lower incidence of AEs, and in patients with brain metastases, sotorasib reduced the risk of progression by 39 % compared with adagrasib[26].

2.1.3. Fulzerasib (IBI351)

Despite the FDA having approved two KRAS^G12C inhibitors, the clinical needs of Chinese patients with KRAS^G12C-mutated tumors remain insufficiently addressed. As the first China - approved covalent KRAS^G12C inhibitor, fulzerasib is supported by a single - arm clinical trial (NCT05005234)conducted in China and published in the Journal of Thoracic Oncology. It shares the same mechanism of action as other agents in its class: it covalently and irreversibly modifies the cysteine residue of the KRAS^G12C protein, inhibits GTP/GDP exchange, and locks the protein in an inactive state. This trial aimed to evaluate the efficacy and safety of fulzerasib in previously treated patients with KRAS^G12C-mutated solid tumors. During the phase Ia dose - escalation stage[27], fulzerasib was evaluated at doses of 250/450/700/900 mg once daily and 450/600/750 mg twice daily (see Table 1 for details). The study found that fulzerasib was well - tolerated in patients with advanced solid tumors and exhibited promising antitumor activity. Data showed that no DLTs were reported across all dose groups, and the most common treatment - related adverse events (TRAEs) were increased gamma - glutamyl transferase (10.2 %) and anemia (6.8 %). In addition, fulzerasib was rapidly absorbed, with a time to maximum concentration of approximately 2 h and an elimination half - life of about 5 h. Based on the comprehensive results of safety, efficacy, and pharmacokinetics, the RP2D was determined to be 600 mg twice daily. As of the data cutoff date of June 13, 2023, 47.2 % of patients remained on treatment, while 52.8 % discontinued treatment, with disease progression (39.8 %) being the primary reason. Among the 166 patients in the NSCLC subgroup, 48 had baseline intracranial tumor lesions. In 29 patients who had not received prior intracranial radiotherapy, the intracranial ORR was 27.6 % (95 %CI: 12.7–47.2), 10 patients (34.5 %) experienced intracranial progression or death events, and the 6 - month PFS rate was 65.0 % (95 %CI: 43.3–80.1). Notably, the study observed a higher incidence of hepatotoxicity in NSCLC patients who had previously received anti-PD-L1 therapy, so more frequent monitoring and strict liver function management are required for this subset of patients. These encouraging data demonstrate fulzerasib’s safety and rapid, durable clinical responses.

In the phase II trial (see Table 1 for details)[28], as of December 13, 2023, 92.2 % (107/116) of patients experienced TRAEs, with anemia (44.8 %) being the most common. Among biomarker - evaluable patients (n = 95), all tissue samples were KRAS^G12C- positive, 72 blood samples were positive, and 23 blood samples were negative. Data indicated that patients with both positive blood and tissue samples had a significantly higher baseline tumor burden (p < 0.05) and shorter PFS (p < 0.05). Tumor mutation analysis revealed that the most common genes co-mutated with KRAS^G12C were Tp53 (45.3 %), STK11 (30.5 %), and KEAP1 (21.1 %). Among the 13 genes with a mutation frequency of 5 % or higher, mutations in 6 genes (STK11, KEAP1, PIK3CG, POLE, SMAD4, and BMPR1B) were significantly associated with shorter PFS (p < 0.05). Specifically, STK11 mutation was significantly correlated with high baseline tumor burden and low response rate (p < 0.05), serving as a potential marker for predicting clinical efficacy. Brain metastases occur frequently in patients with advanced NSCLC and are associated with a poor prognosis. Fulzerasib achieved an intracranial ORR of 22.6 %, with all responses being CR. Post - hoc analysis showed that the ORR was 48.6 % in patients with baseline brain metastases and 49.4 % in those without. Compared with patients with baseline brain metastases, patients without baseline brain metastases had numerically longer median PFS (10.4 months vs 6.4 months) and median overall survival (mOS: 14.1 months vs 10.9 months). These results indicate that fulzerasib monotherapy exhibits durable clinical efficacy with a manageable safety profile, providing strong support for its use in the treatment of patients with KRAS^G12C - mutated NSCLC. The 2025 Chinese Society of Clinical Oncology (CSCO) guidelines have newly added fulzerasib for the second - line treatment of patients with advanced NSCLC[29]. Currently, two key studies are ongoing in China: one is a clinical trial (NCT05005234) further confirming the efficacy and safety of fulzerasib (600 mg BID) in patients with advanced NSCLC, and the other is a phase III study (NCT05497336) investigating fulzerasib in combination with cetuximab for the treatment of patients with advanced CRC.

2.1.4. Garsorasib (D-1553)

Garsorasib exhibited potent tumor growth inhibitory effects in preclinical studies, with high oral bioavailability and CNS penetrability. A phase I multicenter clinical trial evaluated the safety, pharmacokinetics, and efficacy of garsorasib in patients with advanced KRAS^G12C-mutated NSCLC through dose escalation and frequency adjustment (see Table 1 for details)[30]. As of September 12, 2022, 79 previously treated patients with KRAS^G12C - mutated NSCLC were enrolled. No DLTs were observed in the dose - escalation cohort. Across all dose levels, the time to maximum concentration (Tmax) was 1–4 h after a single oral dose and 1–2 h at steady - state following repeated administration, with an elimination half - life of approximately 3–6 h. No drug accumulation was observed after 14 days of administration. In addition, TRAEs occurred in 94.9 % (75/79) of patients, with the most common being liver function abnormalities and gastrointestinal events, indicating that garsorasib has a favorable safety and tolerability profile. Notably, the incidence of grade≥3 liver function abnormalities was 37.8 % (14/37) in patients whose last anti-PD-L1 treatment was within 90 days before the initiation of garsorasib, compared with 16.7 % (7/42) in those who had not received prior anti-PD-L1 treatment or whose last anti-PD-L1 treatment was at least 90 days before garsorasib initiation. These data suggest that this subset of patients has a higher potential risk of garsorasib-related hepatotoxicity, requiring more frequent liver function monitoring. Data showed that 76.0 % (60 patients) received treatment for 3 months or longer, and 59.5 % (47 patients) for 6 months or longer. Treatment was discontinued in 53.2 % of patients, with the most common reason being disease progression (64.3 %). At the first assessment at 6 weeks of treatment, a significant reduction in tumor size was observed in 86.5 % of patients, indicating a rapid and effective therapeutic effect. It is worth noting that the median PFS across all dose groups in this trial was 8.2 months, and the mPFS in the 600 mg twice daily group was 7.6 months. In contrast, the reported mPFS for sotorasib and adagrasib was 6.8 months (with a follow - up period of 15.3 months) and 6.5 months (with a follow - up period of 12.9 months), respectively. The study indicated that garsorasib exerts a relatively durable antitumor effect, and a longer follow - up period is needed to assess the long - term therapeutic outcomes.

The phase II clinical trial[31] aimed to evaluate the efficacy of garsorasib in patients with advanced or metastatic KRAS^G12C-mutated NSCLC. This study was conducted in an exclusively Asian population, supplementing the lack of data on the Asian population in the studies of sotorasib and adagrasib. A total of 123 patients were enrolled. As of November 17, 2023, 67 % (82/123) of patients had discontinued treatment, with the most common reason being disease progression (37 %). Grade≥3 adverse events occurred in 50 % (61/123) of patients, mainly including liver function abnormalities and gastrointestinal events. No new safety signals emerged, indicating a good tolerability of the drug. At the first assessment at 6 weeks of treatment, tumor responses were observed in 79 % (48/61) of the 61 responders. As of the data cutoff date, 25 % (31/123) of patients remained on treatment, among whom 29 maintained a durable response within 6 months.

Furthermore, subgroup analysis revealed no statistically significant differences in efficacy among different co - mutation subgroups (such as STK11, TP53, and KEAP1) and subgroups with different PD - L1 expression levels. Notably, the ORR in patients with KRAS^G12C co-mutated with STK11/TP53/KEAP1 was slightly lower than the overall level, while the ORR in patients with KRAS^G12C co-mutated with PIK3CA was slightly higher than the overall level. In conclusion, garsorasib has demonstrated highly encouraging antitumor activity and favorable tolerability, providing reliable support for the treatment of patients with KRAS^G12C-mutated NSCLC. Currently, ongoing trials include those evaluating garsorasib as monotherapy and in combination with various drugs for the treatment of KRAS^G12C-mutated NSCLC, CRC, and other solid tumors, as well as a phase III trial (NCT06300177) comparing garsorasib with docetaxel in patients with KRAS^G12C- mutated NSCLC.

2.1.5. Glecirasib (JAB-21822)

Glecirasib is a novel covalent oral KRAS^G12C inhibitor approved for marketing in China in 2025. Through novel scaffold modification to regulate overall lipophilicity, construction of a new hydrogen bond network to enhance target-binding affinity, and introduction of multi-halogen groups to block metabolically labile sites, glecirasib exhibits significantly improved drug stability and oral bioavailability. Meanwhile, it redirects the primary elimination pathway to glucuronidation, distinguishing itself from other agents in its class: this modification not only reduces cytochrome P450 (CYP)-mediated toxicities and drug-drug interaction risks but also identifies elevated bilirubin as a key adverse event signal requiring close monitoring. Such structural optimizations achieve a balance between high potency, favorable pharmacokinetics, and low CYP dependency, providing support for safe clinical use. Glecirasib demonstrated potent antitumor efficacy as a monotherapy both in vitro and in vivo in preclinical models. A multicenter, single-arm phase 2b clinical trial conducted in China[32] evaluated the efficacy and safety of oral glecirasib (800 mg once daily, qd) in patients with locally advanced or metastatic KRAS^G12C-mutant NSCLC (see Table 1 for details). A total of 119 patients were enrolled. As of March 28, 2024, safety data showed that glecirasib (JAB-21,822) was well-tolerated, with only 5.0 % (6/119) of patients discontinuing treatment due to TRAEs. No DLTs or treatment-related deaths were reported. Notably, its gastrointestinal toxicity profile was significantly superior to that of other KRAS^G12C inhibitors: the incidences of diarrhea and vomiting were only 5.6 % and 6.9 %, respectively, with no grade ≥2 gastrointestinal adverse events. This favorable safety profile may be attributed to its unique molecular structure and mechanism of action. Compared with other targeted agents, glecirasib exhibits distinct pharmacokinetic characteristics, with an elimination half-life of approximately 4.99–5.54 h and a recommended once-daily dosing schedule, which improves patient adherence.

Furthermore, study results showed that 36 patients achieved a tumor reduction of >50 %. The DOR rates at 6 and 12 months were 73.6 % and 56.6 %, respectively, suggesting a median DOR exceeding 12 months. The 12-month OS rate was 54.6 %, indicating that glecirasib confers significant efficacy in prolonging patient survival. Notably, the ORR was 46.2 % in patients with STK11 co-mutation and 40.0 % in those with wild-type STK11. Although this difference was not statistically significant, the numerical advantage suggests potential therapeutic benefit in the STK11 co-mutation subgroup. Therefore, a phase Ib/II clinical trial is ongoing to further validate the first-line treatment potential of glecirasib in patients with KRAS^G12C and STK11 co-mutations plus wild-type KEAP1 NSCLC. Additionally, the study identified RTK-mediated wild-type RAS activation as the primary resistance mechanism to glecirasib (JAB-21,822). Based on this mechanism, a combination therapy of glecirasib with the SHP2 inhibitor JAB-3312 is currently under investigation in clinical trials (NCT06416410).

2.2. Indirect targeted therapy of KRAS^G12C

Although KRAS^G12C (OFF) direct-targeting inhibitors show good anti-tumor activity, recent clinical data have revealed that almost all patients who benefited clinically from sotorasib and adagrasib treatment developed resistance. In the 2-year data analysis of the CodeBreaK 100 trial, 36 % of the patients experienced primary resistance or early disease progression (PFS < 3 months) after sotorasib treatment[18]. In the NSCLC analysis data, patients with co-mutations of KEAP1, SMARCA4, and CDKN2A tended to have poorer clinical outcomes, and the biological mechanisms of resistance mediated by these mutations still need to be explored[33]. The acquired resistance mechanisms currently discovered can be roughly classified into: (1) Acquired resistance caused by gene mutations: KRAS^G12C codon mutations to another mutant (G12X) or secondary activating mutations on the previous WT KRAS allele (G12D/R/V, G13D, Q61H), Switch II site mutations (R68, H95 or Y96) bound to KRAS^G12C inhibitor, KRAS^G12C gene amplification or copy number increase, upstream receptor tyrosine kinase (RTK) amplification or mutation (such as EGFR, FGFR), NRAS or HRAS activating mutations, downstream effector activating mutations, such as BRAF, MEK, or PIK3CA[34]; (2) Adaptive resistance: activation of WT RAS and reactivation of upstream RTK,KRAS and its downstream signaling pathways (RAF/MEK and PI3K/AKT); (3) Tissue/cell state transformation: in rapid autopsy cases of lung adenocarcinoma patients who relapsed during sotorasib treatment, epithelial-mesenchymal transition (EMT) was found to be the main mechanism of resistance[35]; In addition, new preclinical studies have also observed that lung adenocarcinoma transformed into type I alveolar-like state[36]or squamous cell state[37]after treatment with KRAS inhibitors.

Therefore, given the diversity of tumor resistance mutations, treatment sequencing has emerged as a core strategy for optimizing subsequent therapeutic regimens. A large-scale genomic analysis of 143 patients with disease progression following STK11 co-mutations plus wild-type KEAP1 NSCLC. inhibitor treatment (68 with NSCLC, 58 with CRC)[38] revealed that 46 % of patients harbored acquired genomic alterations, predominantly concentrated in the RAS/MAPK signaling axis: KRAS activating mutations (25 %), KRAS amplification (22 %), RAF/MAPK mutations/fusions (21 %), and NRAS/HRAS mutations (8 %). Notably, the resistance mutation landscape exhibited significant cancer-type specificity—69 % of CRC patients carried ≥1 RAS/MAPK pathway alteration, compared with only 26 % of NSCLC patients—further underscoring the necessity of cancer-type-specific sequencing analyses. Single-cell sequencing further confirmed that secondary RAS/BRAF mutations coexist with the original KRAS^G12C mutation in the same cells, enabling circumvention of targeted inhibition without impairing target inactivation. This "heterogeneous resistance" highlights the critical importance of treatment sequencing, providing guidance for the development of personalized therapeutic strategies.

The choice of subsequent sequencing technology varies across clinical scenarios: the NGS, leveraging its high-throughput advantage, can simultaneously detect multiple genes (such as RAS, BRAF, EGFR) and unknown alterations, and is recommended as a core tool by NCCN/ESMO guidelines. For dynamic monitoring and efficacy assessment, digital droplet PCR or BEAMing technology are preferred due to their ultra-high sensitivity (limit of detection: 0.03 %–0.05 %), enabling early identification of new subclonal mutations (such as KRAS^G12V) 2–4 weeks prior to radiological progression[39]. For patients with insufficient tissue samples, liquid biopsy based on circulating tumor DNA offers substantial value. The ultimate goal of sequencing is to match patients with personalized therapeutic regimens based on the detected genomic alterations.Therefore, for acquired resistance mechanisms, new KRAS-targeting drugs such as pan-RAS ON inhibitors or those targeting upstream and downstream signaling molecules of the RAS signaling pathway seem to be able to bring more durable clinical benefits to patients.

2.2.1. KRAS^G12C ON inhibitor

The KRAS^G12C OFF inhibitor relies on the covalent interaction between a small molecule inhibitor and the mutated cysteine residue. However, this strategy is not applicable to other amino acids. The tri-complex inhibitor is an allele-specific and pan-allele-targeting drug that can combine with KRAS-GTP through a unique mechanism and has potential theoretical advantages in clinical treatment to overcome drug resistance. It was found that in cases of resistance to KRAS^G12C inhibitors, secondary Y96D (Switch II site) mutations were detected. However, in the G12C/Y96D double-mutant cancer cell line model, the KRAS^G12C ON inhibitor RM-018 maintained anti-tumor efficacy, demonstrating that this drug can overcome the resistance mechanism of KRAS OFF inhibitors and provides a new treatment strategy for KRAS-driven cancers[40]. Similarly, the compounds RMC-4998 and RMC-6291 developed in clinical trials showed consistent clinical efficacy. These compounds are natural products of a macrolide class discovered from actinomycete strains, sanglifehrin A, which have high affinity for the holoprotein A (CYPA). RMC-6291 (Elironrasib) is an oral inhibitor that can combine with CypA to reshape its surface to form a binary complex and bind with KRAS protein with high affinity, ultimately generating a CypA-compound-KRAS(GTP) complex, blocking downstream oncogenic signaling pathways[41]. Preliminary research data indicate that RMC-6291 can produce deep and persistent inhibition of the RAS pathway activity in KRAS^G12C tumor models and has good anti-tumor activity, and may overcome the resistance of the first-generation KRAS^G12C (OFF) inhibitors[42]. Currently, newly developed drugs such as FMC-376 and D3S-001 can directly combine with KRAS-GTP, similar to KRAS^G12C OFF inhibitors, have good clinical data, and overcome some secondary resistance changes.

2.2.2. Pan-RAS/Pan-KRAS inhibitor

The Pan-RAS inhibitor that targets all mutant and wild-type subtypes has a theoretical advantage over targeting individual mutant alleles alone. Firstly, the relative rarity of certain KRAS alleles (such as Q61, G13X, etc.) makes it impractical to generate specific inhibitors for each point mutation, thus broadening the therapeutic potential of pan-RAS/KRAS drugs. Secondly, pan-targeted drugs may block the compensatory activation of the wild-type RAS subtype. Thirdly, pan-RAS/KRAS drugs may prevent the emergence of at least one acquired resistance pattern (targeted mutations within KRAS), and thus can also be used for treatment after directly targeting specific alleles.

RMC-6236 is a RAS MULTI (ON) trimeric complex inhibitor, similar to the KRAS^G12C (ON) inhibitor. This drug forms a trimeric complex between the active RAS and CypA proteins. It is active against mutant and wild-type KRAS, NRAS, and HRAS, and has shown effective preclinical activity in tumors with various RAS genotypes, including cancer models resistant to KRAS^G12C OFF inhibitors due to secondary NRAS mutations, KRAS amplification, and RTK amplification[43]. Early data presented at the 2023 ESMO conference reported the clinical efficacy of this compound (NCT05379985). In 40 NSCLC patients, the ORR was 38 %, DCR was 85 %, and 1/40 showed CR. In 46 PDAC patients, the ORR was 20 % (all PR), and DCR was 87 %. The most common grade 3 TRAEs was rash; only 1 patient had a grade 4 intestinal perforation at the tumor site.

BI-2493 and BI-2865 are novel pan-KRAS inhibitors that selectively bind to KRAS-GDP. They prevent the activation of wild-type and mutant KRAS by blocking nucleotide exchange, while retaining other RAS family proteins, thereby minimizing damage to normal cells. The researchers such as Antonio, found that both BI-2493 and BI-2865 exhibited good tumor suppression effects in vitro and vivo experiments. BI 3706,674, as its further optimized product, has entered the clinical trial stage[44]. Pan-KRAS inhibitors provide new treatment options for KRAS mutant patients.

2.3. Combination treatment strategies for KRAS

Chemotherapy can rapidly reduce tumor burden and mitigate early resistance; anti-angiogenic therapy can induce durable tumor regression by restricting tumor vasculature and promoting vascular normalization within the tumor microenvironment (TME). Vascular normalization upregulates the expression of leukocyte adhesion molecules (such as ICAM1 and VCAM1), improves perfusion and oxygenation, thereby enhancing T-cell infiltration and converting the TME from immunosuppressive to immunostimulatory. Consequently, administration of ICIs during this vascular normalization window can improve clinical outcomes[45,46]. Moreover, studies have shown that KRAS mutations can improve the immune-mediated landscape by affecting immune cell infiltration, increasing immunosuppressive properties, and promoting the upregulation of immune checkpoints (especially PD-L1 expression)[47,48], These changes interact with the tumor mutation burden (TMB), tumor-infiltrating lymphocytes (TIL) within the TME, chemokines, and other oncogenic drivers, influencing the clinical efficacy of ICIs therapy. Therefore, combined ICIs can improve the immunogenicity and sensitivity of tumors with low immunogenicity. therefore, chemotherapy combined with immune checkpoint inhibitors (ICIs) or anti-angiogenic therapy remains the first-line treatment for patients with KRAS^G12C-mutant NSCLC. However, its clinical efficacy is limited. In the KEYNOTE-189 trial, data demonstrated that the first-line regimen of pemetrexed plus carboplatin and pembrolizumab achieved a median PFS of 9.0 months (95 %CI: 8.1–9.9) and a median second progression-free survival of 17.0 months (95 %CI: 15.1–19.4)[49]. In contrast, KRAS-targeted monotherapy has shown a modest median PFS of only approximately 5–8 months. This therapeutic disparity indicates that while KRAS inhibitor monotherapy exhibits preliminary efficacy in clinical practice, its action is restricted to a single target (KRAS^G12C mutation). Consequently, it is highly susceptible to tumor heterogeneity and the emergence of resistant clones, which further limits the full realization of its monotherapeutic potential.

Accordingly, current research on first-line clinical regimens has shifted toward combination therapeutic strategies involving KRAS inhibitors, such as combinations with chemotherapy, immunotherapy, or anti-angiogenic agents, aiming to synergistically enhance efficacy through multiple mechanisms and overcome the limitations of monotherapy. A phase Ⅱ study[50] demonstrated that first-line treatment with sotorasib combined with carboplatin/pemetrexed yielded an ORR of 88.9 %, a median PFS of 6.6 months, and an mOS exceeding 20 months, with particularly significant benefits observed in patients with PD-L1 expression < 1 %. Grade ≥ 3 TRAEs were predominantly hematological toxicities. The favorable efficacy and safety profile of this regimen provide a crucial therapeutic option for patients with low PD-L1 expression or those requiring rapid tumor shrinkage.

The 2025 ASCO conference reported partial results from the KRYSTAL-7 phase II trial, which aimed to compare the safety and efficacy of adagrasib combined with pembrolizumab versus monotherapy immunotherapy for advanced or metastatic KRAS^G12C-mutant NSCLC patients. As of August 23, 2024, a total of 149 patients were included, with an ORR of 44.3 % (95 %CI: 36.2–52.7), a mDOR of 26.3 months (95 %CI: 14.9 to uncertain), and a mPFS of 11.0 months (95 %CI: 5.8–14.0). The current trial is still recruiting participants[51]. However, the co-mutation of tumor suppressor genes such as STK11 and KEAP1 and the resistance of ICIs enable tumor immune escape. Studies have found that in NSCLC mouse models, dual immune checkpoint blockade (targeting PD-(L)1 and CTLA4 pathways) can regulate the TME by enhancing T cell activation and infiltration within the TME to overcome resistance. In ICIs-resistant NSCLC patients carrying STK11 and/or KEAP1 mutations, the OS is higher when treated with anti-PD-L1 antibody durvalumab combined with anti-CTLA4 antibody tremelimumab compared to chemotherapy with docetaxel or chemotherapy alone[52].

Given the acquired resistance mechanism of KRAS^G12C inhibitors, a treatment strategy that targets both upstream and downstream signaling molecules of the RAS signaling pathway seems to offer more durable benefits to patients. Currently, ongoing clinical trials mainly involve treatment strategies using KRAS^G12C inhibitors combination with SHP2 inhibitors, SOS1 inhibitors, EGFR and pan-ERBB inhibitors, MEK inhibitors, ERK inhibitors, and mTOR inhibitors. A single-arm, multi-center clinical trial reported at the 2025 ELCC conference demonstrated the considerable efficacy of the combination treatment of fulzerasib and cetuximab. SOS1/SHP2 are key downstream signaling molecules of RTKs, they can keep KRAS in an inactive state. In preclinical models, SHP2 inhibitors and KRAS^G12C selective inhibitors have a synergistic effect. In an I/IIa study, the combined treatment of glecirasib (KRAS^G12C inhibitor) and JAB-3312 (SHP2 inhibitor) showed that the ORR of KRAS^G12C-primarily treated NSCLC patients was 50 % (14/28), and the DCR was 100 %. In NSCLC patients who had previously received KRAS^G12C inhibitor treatment, the ORR was 14.3 % (1/7), and the DCR was 57.1 %[53]. In addition, in the ongoing CodeBreak 101 trial, which evaluated the safety and efficacy of sotorasib combined with MEK inhibitor trametinib, mTOR inhibitor everolimus, pan-KRAS inhibitor BI 1701,963 and son on for KRAS^G12C advanced solid tumors.

Cyclin-dependent kinase (CDK) is the second potential co-target for KRAS inhibitors. The classical MAPK pathway drives the transcription of cyclin D1, leading to the heterodimerization of CDK4/6, as well as CDK-dependent phosphorylation of RB and cell cycle transition. In preclinical studies, combined inhibition of CDK and KRAS can prevent cell cycle progression. Moreover, this combination may cause cancer cell arrest and enhance immune-mediated monitoring of residual tumor cells, suggesting the existence of an autonomous co-regulatory mechanism. Currently ongoing clinical trials combinations include sotorasib and palbociclib (CodeBreak101), adagralib and palbociclib (KRYSTAL-16), and JDQ443 and ribociclib(NCT04185883, NCT05178888, NCT05358249).

In addition to reactivation from the MAPK signaling axis, resistance can also be obtained by reactivating parallel signaling pathways, enabling tumors to break free from dependence on the mutated KRAS. In recent preclinical studies, members of the Hippo pathway, YAP and TAZ, have emerged as the main regulators of KRAS inhibitor resistance[54], regarding this therapeutic strategy, it is still in the preliminary clinical trials.

Notably, as a core driver of NSCLC, KRAS mutation not only promotes tumor cell proliferation and survival but also triggers oncogene-induced senescence (OIS), a key tumor-suppressive program[55]. This process is primarily regulated by the p53 pathway, which blocks the proliferation of damaged cells through the p53/p21 and p16/RB pathways in the early stages of tumorigenesis. Its protective effects stem from both cell-autonomous growth arrest and non-cell-autonomous mechanisms mediated by the senescence-associated secretory phenotype (SASP). However, KRAS-mutant NSCLC cells can escape OIS through multiple mechanisms. Clinical studies have shown that approximately 45 % of patients treated with adagrasib develop resistance due to the senescence escape mechanism of tumor cells. Up to 20 % of patients with KRAS-mutant NSCLC harbor loss-of-function mutations in cell cycle regulators and the tumor suppressor gene CDKN2A (encoding p16), leading to sustained phosphorylation of retinoblastoma (RB) protein mediated by CDK4/6, which bypasses cell cycle arrest and drives abnormal cell proliferation.

Furthermore, KRAS can indirectly regulate the senescence program by modulating autophagy (such as the USP5-Beclin 1 axis): KRAS activates USP5 (ubiquitin-specific protease 5) through reactive oxygen species (ROS), which further stabilizes Beclin 1 (autophagy-related protein 1). Beclin 1 promotes autophagy and degrades p53 (tumor suppressor protein), thereby inhibiting p53-dependent OIS. Inhibition of USP5 or Beclin 1 can restore the senescence effect and block KRAS-mediated NSCLC growth[56]. The transactivation domain of the transcription factor Twist1 is a key element in its inhibition of OIS; it suppresses KRAS oncogene-induced OIS by upregulating O-GlcNAcylation, thereby promoting NSCLC tumorigenesis. Preclinical studies have confirmed that inhibiting O-GlcNAcylation, using the novel MYC inhibitor MYCi975, or knocking down the MYC gene can restore OIS and inhibit tumors[57].

Notably, senescent cells in the TME, predominantly macrophages and endothelial cells, also contribute to KRAS-driven lung cancer progression: these cells promote tumorigenesis in a paracrine manner through SASP. In KRAS-driven lung cancer models, genetic approaches, senolytics (senescent cell-clearing agents), or macrophage depletion significantly reduced tumor burden and prolonged survival, revealing the protumorigenic role of senescent cells in the TME[58]. Additionally, a recent phase I clinical trial of the novel MYC inhibitor OMO-103 enrolled 22 patients with various cancer types, among whom 9 (including the only NSCLC patient) achieved stable disease at efficacy assessment[59]. This combination therapeutic strategy provides a new direction for overcoming senescence escape and targeted therapy resistance in KRAS-mutant NSCLC, but its clinical translation value requires further verification in future clinical trials.

2.4. Novel treatment strategy of targeted KRAS

Antibody-drug conjugates (ADCs) are a promising therapeutic agent that specifically combine cytotoxic drugs with cancer cells through monoclonal antibodies. In NSCLC, ADCs targeting HER2 (such as trastuzumab deruxtecan) and TROP2 (such as datopotamab deruxtecan) have been found that not only eliminate cancer cells but also enhance the potential of antigen release and immune activation within the TME. In the II phase DESTINY-Lung01 trial, trastuzumab deruxtecan showed anti-tumor activity in patients with HER2-overexpressing metastatic NSCLC, with an ORR was up to 34 %[60]. In the II phase TROPION-Lung 05 study, the ORR for patients receiving datopotamab deruxtecan was 36 %, and even 44 % in patients with EGFR mutations[61].

Adoptive cell therapy (ACT), including chimeric antigen receptor T cells (CART), T cells receptor-engineered T cells, and tumor-infiltrating lymphocytes (TIL), is currently undergoing multiple early clinical trials in NSCLC[62,63]. Despite challenges such as TME-induced immunosuppression, poor cell trafficking, and limited persistence that hinder the efficacy of ACT in patients with solid tumors, innovations in engineering and delivery methods are being further developed.

Personalized cancer vaccines targeting patient-specific tumor neoantigens have emerged as a promising strategy for treating patients with solid tumors[64]. Oncolytic viruses selectively lyse cancer cells, releasing tumor antigens and promoting immune cell infiltration, while cancer vaccines activate T cells responses to tumor neoantigens and reconfigure the immune environment within the TME. In NSCLC patients, personalized mRNA vaccines combined with pembrolizumab have been proven to expand T cell responses and restructure the TME[65].

Clinical trials are testing the combination of cytokines and ICIs to stimulate T cell and NK cell functions, thereby improving the anti-tumor immune response in NSCLC patients. Although preclinical studies and the CANTOS trialindicate that exposure to anti-IL-1β antibody canakimumab can reduce the incidence of lung cancer, the therapeutic benefits of this approach for NSCLC patients remain unclear[66,67]. Moreover, the CANOPY trial suggests that canakimumab is ineffective when combination with chemotherapy or ICI. These results still require further research to determine the optimal patient population, timing, or combination strategies to fully utilize the therapeutic potential of IL-1β blockers in NSCLC.

3. Summary and outlook

Lung cancer is one of the cancers with the highest mortality rate worldwide. KRAS mutations were previously regarded as "undrugable" targets. However, with technological advancements and in-depth research, significant breakthroughs have been achieved. Sotorasib achieves high selectivity, while adagrasib significantly enhances blood-brain barrier penetration by increasing hydrophobicity and inhibiting P-glycoprotein efflux. Fulzerasib adopts a rigid scaffold to reduce cytochrome P450 (CYP) enzyme-dependent metabolism, resulting in the longest median progression-free survival (mPFS) among these agents. Garsorasib, by decreasing lipophilicity and increasing the proportion of free drug, emerges as a potential therapeutic option for patients with brain metastases. Collectively, these advances have shaped a new therapeutic landscape for patients with advanced or metastatic KRAS^G12C-mutant NSCLC. Nevertheless, due to the limited long-term survival benefits of single-agent treatment and the emergence of drug resistance, combination therapies targeting different resistance mechanisms, such as combination with PD-1/PD-L1/CTLA4, SHP2, SOS1, EGFR and pan-ERBB, MEK, ERK, mTOR, CDK antibody therapies, as well as MYC inhibitor,ADCs, ACTs, and malignant tumor vaccines, have brought new treatment ideas to clinical patients. Moreover, KRAS mutant tumors have high heterogeneity, and treatment strategies need to be individualized based on their subtypes, co-mutated genes, upstream and downstream pathway activation mechanisms, immune microenvironment characteristics, and resistance mechanisms. In conclusion, current research has made significant progress in treating KRAS^G12C-mutant NSCLC, this review has the potential to serve as a valuable resource for clinicians as it offers insights into NSCLC treatment, but this also may only touch the potential surface of treatment possibilities. Further in-depth exploration is still needed to better balance the long-term survival benefits for patients.

CI: Confidence Interval; CNS: Central Nervous System; DCR: Disease Control Rate; FDA: Food and Drug Administration; HR: Hazard Ratio; mOS: Median Overall Survival, mPFS: Median Progression Free Survival; NA: Not Applicable; NDA: New Drug Application; NSCLC: Non-Small Cell Lung Cancer; NE: Not evaluated; ORR: Objective Response Rate; TRAE:Treatment-Related Adverse Event; +:with + indicating that the value includes patient data that were censored at data cutoff; mDOR:Median duration of response; NR:not reached

Glossary

NSCLC	non-small cell lung cancer
KRAS	Kirsten rat sarcoma virus
NGS	next-generation sequencing
PDAC	pancreatic ductal adenocarcinoma
CRC	colorectal cancer
NRAS	neuroblastoma rat sarcoma virus
HRAS	Harvey rat sarcoma virus
HVR	hypervariable region
EGF	epidermal growth factor
FGF	fibroblast growth factor
PDGF	platelet-derived growth factor
IGF	insulin-like growth factor
GTPase	guanosine triphosphatase
MEK	mitogen-activated protein/extracellular signal-regulated kinase
GEM	genetically engineered mouse
RTK	receptor tyrosine kinases
MAPK	mitogen-activated protein kinase
PI3K	phosphatidylinositol 3-kinase
GEF	guanine nucleotide-exchange factor
GAPs	GTPase-activating proteins
EGFR	epidermal growth factor receptor
AEs	Adverse events
CV	coefficient of variation
Tmax	the time to reach maximum concentration
ORR	the objective response rate
Cmax	the maximum plasma concentration
TMB	tumor mutational burden
PFS	progression-free survival
ctDNA	circulating tumor DNA
ECOG	Eastern Cooperative Oncology Group
RP2D	the recom mended phase II dose
mNSCLC	metastatic NSCLC
CI	Confidence Interval
CNS	Central Nervous System
DOR	duration of response
cORR	confirmed objective response rates
LCSS	Lung Cancer Symptom Scale
BM	baseline brain metastases
TIP	intracranial progression
CR	complete response
TRAEs	treatment - related adverse events
mOS	median overall survival
CSCO	Chinese Society of Clinical Oncology
DLTs	dose - limiting toxicities
ICIs	immune checkpoint inhibitors
TME	tumor microenvironment
TMB	tumor mutation burden
CDK	Cyclin-dependent kinase
OIS	oncogene-induced senescence
SASP	senescence-associated secretory phenotype
ROS	reactive oxygen species
FDA	Food and Drug Administration
ADCs	Antibody-drug conjugates
ACT	Adoptive cell therapy
CART	chimeric antigen receptor T cells

Funding statement

Funding: This work was supported by Clinical Research Project of Air Force Medical University (No. 2022LC2256).

CRediT authorship contribution statement

Ting Tian: Writing – review & editing, Writing – original draft. Wangping Li: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.