KRAS G12C inhibitors versus chemotherapy in second line and beyond in adults with advanced or metastatic non-small cell lung cancer (NSCLC) harbouring the KRAS G12C mutation.
1/5 보강
This is a protocol for a Cochrane Review (intervention).
APA
Noordhof A, Asmara OD, et al. (2026). KRAS G12C inhibitors versus chemotherapy in second line and beyond in adults with advanced or metastatic non-small cell lung cancer (NSCLC) harbouring the KRAS G12C mutation.. The Cochrane database of systematic reviews, 2(2), CD016054. https://doi.org/10.1002/14651858.CD016054
MLA
Noordhof A, et al.. "KRAS G12C inhibitors versus chemotherapy in second line and beyond in adults with advanced or metastatic non-small cell lung cancer (NSCLC) harbouring the KRAS G12C mutation.." The Cochrane database of systematic reviews, vol. 2, no. 2, 2026, pp. CD016054.
PMID
41700598 ↗
Abstract 한글 요약
This is a protocol for a Cochrane Review (intervention). The objectives are as follows: To assess the benefits and harms of G12C-inhibitors compared to chemotherapy in second line and beyond in adults with advanced/metastatic non-small cell lung cancer (NSCLC) with a Kirsten rat sarcoma (KRAS) G12C mutation.
🏷️ 키워드 / MeSH 📖 같은 키워드 OA만
📖 전문 본문 읽기 PMC JATS · ~34 KB · 영문
Background
Background
Description of the condition
Lung cancer is the leading cause of cancer‐related death worldwide [1, 2]. Annually, 2.2 million people are diagnosed with lung cancer [1, 2, 3, 4, 5]. Almost half of these people already have advanced/metastatic disease at the time of their diagnosis [6]. The incidence of lung cancer varies across the world, due to differences in tobacco smoking habits, which is the leading risk factor for the development of lung cancer [7].
The two main types of lung cancer are small‐cell lung cancer (SCLC) and non‐small cell lung cancer (NSCLC). NSCLC accounts for approximately 80% of all types of lung cancer, and can be further divided into different histological subtypes, of which adenocarcinoma (55%) and squamous cell carcinoma (30%) are the most common [6, 8].
In NSCLC, several oncogenic driver mutations (genetic alterations that boost the growth of cancer cells) can occur. Kirsten rat sarcoma (KRAS) is the most common oncogenic driver mutation and is seen in 30% to 40% of all people with the adenocarcinoma type in western populations [9, 10]. In Asian populations, the KRAS mutation is less frequently seen, with an incidence of only 5% to 15% [11, 12]. The KRAS mutation is associated with tobacco smoking and also with a higher tumour‐programmed death‐ligand 1 (PD‐L1) expression on tumour cells [13, 14, 15].
KRAS is part of the rat sarcoma virus (RAS) oncogene family and is an essential membrane‐bound guanosine triphosphatase (GTP‐ase), involved in cell proliferation and survival via intracellular downstream signalling pathways such as the mitogen‐activated protein kinase (MAPK), phosphatidylinositol‐3‐kinase (PI3K) and RAS‐like pathways. The KRAS protein can switch between a guanosine triphosphate (GTP)‐bound active state promoting proliferation and a guanosine diphosphate (GDP)‐bound inactive state. When mutated, KRAS remains in its GTP‐bound active state, leading to uncontrolled cell growth [9, 15, 16]. There are different KRAS subtype mutations; the most common one is the G12C subtype, which is seen in 40% of all KRAS mutations [15, 17, 18].
The first line of treatment for people with metastatic NSCLC (or advanced NSCLC requiring systemic treatment) is immunotherapy with checkpoint inhibitors (ICI) with or without platinum doublet chemotherapy according to the PD‐L1 expression [8]. The introduction of ICI in 2015 has led to significantly improved survival for people with NSCLC [8, 19, 20, 21, 22].
However, for people with NSCLC harbouring an actionable driver mutation (i.e. a mutation for which targeted therapy is available), such as epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK), amongst others, targeted therapy is the preferred first‐line treatment, and this generally offers these people a better prognosis than ICI with or without chemotherapy [23].
For decades, KRAS was considered “undruggable”, meaning that efforts to develop targeted therapy for KRAS were unsuccessful [24]. However, this has recently changed, shortly after the discovery of a specific groove in the KRAS G12C molecule. Since this discovery, several small molecules (G12C‐inhibitors) have been designed to fit into this specific groove, keeping KRAS in its GDP‐bound inactive state and thus inhibiting downstream signalling and preventing survival and proliferation [25, 26, 27].
The first line of treatment for people with NSCLC with a KRAS G12C mutation is currently still ICI with or without chemotherapy. In the setting of first‐line mono‐immunotherapy, it was shown that KRAS did not have a prognostic impact on survival compared to people without the mutation [10]. Furthermore, in the setting of first‐line (chemo‐)immunotherapy, people with the specific KRAS G12C mutation did not have significantly different survival than those with a KRAS non‐G12C mutation [17].
For people with progressive disease despite first‐line treatment, G12C‐inhibitors such as (but not limited to) sotorasib and adagrasib are now available as second‐line treatment and beyond in daily clinical practice.
Description of the intervention and how it might work
In the CodeBreaK 100 phase II study, the G12C‐inhibitor sotorasib showed an overall response rate (ORR) of 41% and a median progression‐free survival (PFS) of 6.3 months [28]. For the G12C‐inhibitor adagrasib, the ORR was 43% and the PFS was 6.5 months in the KRYSTAL‐1 phase II trial [29]. These trials both included participants that had all received at least one line of previous systemic therapy. People with active untreated brain metastases were excluded.
Both sotorasib and adagrasib have now also been studied in phase III trials, where participants were all pretreated with at least one line of systemic treatment. In the phase III CodeBreaK 200 trial, participants treated with sotorasib had a higher median PFS (5.6 months) versus the pretreated participants who received chemotherapy through docetaxel (4.5 months), hazard ratio (HR) 0.66, P = 0.0017 [30]. Similar results were seen in the primary analysis from the KRYSTAL‐12 phase III study, where participants treated with adagrasib had a higher median PFS (5.5 months) versus the participants treated with docetaxel (3.8 months), HR 0.58, P < 0.0001 [31, 32].
Why it is important to do this review
To date, trials investigating the effects of sotorasib and adagrasib have been conducted within highly selective study populations, all of whom had undergone prior systemic treatment. Data from individual trials showed some benefit, but failed to meet reimbursement criteria in many countries, and many questions remain about G12C‐inhibition effectiveness.
Consequently, although the published trials with sotorasib and adagrasib have shown promising results, a systematic review of pooled randomised controlled trials (RCTs) is needed to provide proper judgement regarding the outcomes of G12C‐inhibitors. Now that G12C‐inhibitors are available (in some countries) in a second‐line setting and beyond, we aim to assess the outcomes of G12C‐inhibitors in the second line and beyond in people with advanced NSCLC with a KRAS G12C mutation using pooled RCT data.
With pooled data from these studies, the effects of G12C‐inhibitors can be better assessed. More phase III trials including new G12C‐inhibitors are expected in the upcoming period, so we anticipate an even higher yield of trials to be included. This need for pooled data holds particular significance as the optimal treatment strategy for KRAS G12C‐mutated NSCLC remains unclear, including the optimal positioning of G12C inhibitors within the treatment paradigm.
Description of the condition
Lung cancer is the leading cause of cancer‐related death worldwide [1, 2]. Annually, 2.2 million people are diagnosed with lung cancer [1, 2, 3, 4, 5]. Almost half of these people already have advanced/metastatic disease at the time of their diagnosis [6]. The incidence of lung cancer varies across the world, due to differences in tobacco smoking habits, which is the leading risk factor for the development of lung cancer [7].
The two main types of lung cancer are small‐cell lung cancer (SCLC) and non‐small cell lung cancer (NSCLC). NSCLC accounts for approximately 80% of all types of lung cancer, and can be further divided into different histological subtypes, of which adenocarcinoma (55%) and squamous cell carcinoma (30%) are the most common [6, 8].
In NSCLC, several oncogenic driver mutations (genetic alterations that boost the growth of cancer cells) can occur. Kirsten rat sarcoma (KRAS) is the most common oncogenic driver mutation and is seen in 30% to 40% of all people with the adenocarcinoma type in western populations [9, 10]. In Asian populations, the KRAS mutation is less frequently seen, with an incidence of only 5% to 15% [11, 12]. The KRAS mutation is associated with tobacco smoking and also with a higher tumour‐programmed death‐ligand 1 (PD‐L1) expression on tumour cells [13, 14, 15].
KRAS is part of the rat sarcoma virus (RAS) oncogene family and is an essential membrane‐bound guanosine triphosphatase (GTP‐ase), involved in cell proliferation and survival via intracellular downstream signalling pathways such as the mitogen‐activated protein kinase (MAPK), phosphatidylinositol‐3‐kinase (PI3K) and RAS‐like pathways. The KRAS protein can switch between a guanosine triphosphate (GTP)‐bound active state promoting proliferation and a guanosine diphosphate (GDP)‐bound inactive state. When mutated, KRAS remains in its GTP‐bound active state, leading to uncontrolled cell growth [9, 15, 16]. There are different KRAS subtype mutations; the most common one is the G12C subtype, which is seen in 40% of all KRAS mutations [15, 17, 18].
The first line of treatment for people with metastatic NSCLC (or advanced NSCLC requiring systemic treatment) is immunotherapy with checkpoint inhibitors (ICI) with or without platinum doublet chemotherapy according to the PD‐L1 expression [8]. The introduction of ICI in 2015 has led to significantly improved survival for people with NSCLC [8, 19, 20, 21, 22].
However, for people with NSCLC harbouring an actionable driver mutation (i.e. a mutation for which targeted therapy is available), such as epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK), amongst others, targeted therapy is the preferred first‐line treatment, and this generally offers these people a better prognosis than ICI with or without chemotherapy [23].
For decades, KRAS was considered “undruggable”, meaning that efforts to develop targeted therapy for KRAS were unsuccessful [24]. However, this has recently changed, shortly after the discovery of a specific groove in the KRAS G12C molecule. Since this discovery, several small molecules (G12C‐inhibitors) have been designed to fit into this specific groove, keeping KRAS in its GDP‐bound inactive state and thus inhibiting downstream signalling and preventing survival and proliferation [25, 26, 27].
The first line of treatment for people with NSCLC with a KRAS G12C mutation is currently still ICI with or without chemotherapy. In the setting of first‐line mono‐immunotherapy, it was shown that KRAS did not have a prognostic impact on survival compared to people without the mutation [10]. Furthermore, in the setting of first‐line (chemo‐)immunotherapy, people with the specific KRAS G12C mutation did not have significantly different survival than those with a KRAS non‐G12C mutation [17].
For people with progressive disease despite first‐line treatment, G12C‐inhibitors such as (but not limited to) sotorasib and adagrasib are now available as second‐line treatment and beyond in daily clinical practice.
Description of the intervention and how it might work
In the CodeBreaK 100 phase II study, the G12C‐inhibitor sotorasib showed an overall response rate (ORR) of 41% and a median progression‐free survival (PFS) of 6.3 months [28]. For the G12C‐inhibitor adagrasib, the ORR was 43% and the PFS was 6.5 months in the KRYSTAL‐1 phase II trial [29]. These trials both included participants that had all received at least one line of previous systemic therapy. People with active untreated brain metastases were excluded.
Both sotorasib and adagrasib have now also been studied in phase III trials, where participants were all pretreated with at least one line of systemic treatment. In the phase III CodeBreaK 200 trial, participants treated with sotorasib had a higher median PFS (5.6 months) versus the pretreated participants who received chemotherapy through docetaxel (4.5 months), hazard ratio (HR) 0.66, P = 0.0017 [30]. Similar results were seen in the primary analysis from the KRYSTAL‐12 phase III study, where participants treated with adagrasib had a higher median PFS (5.5 months) versus the participants treated with docetaxel (3.8 months), HR 0.58, P < 0.0001 [31, 32].
Why it is important to do this review
To date, trials investigating the effects of sotorasib and adagrasib have been conducted within highly selective study populations, all of whom had undergone prior systemic treatment. Data from individual trials showed some benefit, but failed to meet reimbursement criteria in many countries, and many questions remain about G12C‐inhibition effectiveness.
Consequently, although the published trials with sotorasib and adagrasib have shown promising results, a systematic review of pooled randomised controlled trials (RCTs) is needed to provide proper judgement regarding the outcomes of G12C‐inhibitors. Now that G12C‐inhibitors are available (in some countries) in a second‐line setting and beyond, we aim to assess the outcomes of G12C‐inhibitors in the second line and beyond in people with advanced NSCLC with a KRAS G12C mutation using pooled RCT data.
With pooled data from these studies, the effects of G12C‐inhibitors can be better assessed. More phase III trials including new G12C‐inhibitors are expected in the upcoming period, so we anticipate an even higher yield of trials to be included. This need for pooled data holds particular significance as the optimal treatment strategy for KRAS G12C‐mutated NSCLC remains unclear, including the optimal positioning of G12C inhibitors within the treatment paradigm.
Objectives
Objectives
To assess the benefits and harms of G12C‐inhibitors compared to chemotherapy in second line and beyond in adults with advanced/metastatic non‐small cell lung cancer (NSCLC) with a Kirsten rat sarcoma (KRAS) G12C mutation.
To assess the benefits and harms of G12C‐inhibitors compared to chemotherapy in second line and beyond in adults with advanced/metastatic non‐small cell lung cancer (NSCLC) with a Kirsten rat sarcoma (KRAS) G12C mutation.
Methods
Methods
Criteria for considering studies for this review
Types of studies
We will include phase III randomised controlled trials (RCTs) studying the effects of G12C‐inhibitors in a second‐line setting and beyond, versus chemotherapy, in adults with advanced or metastatic lung adenocarcinoma.
We will apply no restriction on language or publication status. We will also include unpublished online data and meeting abstracts, but only when sufficient data are available and a proper risk of bias assessment can be made.
Cross‐over RCTs will not be eligible in this review because of possible bias by 'carry‐over' effects. Cluster‐randomised RCTs will not be eligible because, in these trials, the unit of randomisation is typically a group, rather than individual participants. This contrasts with the focus of this review, which aims to assess outcomes for individual participants. We will also exclude non‐randomised studies of interventions (NRSIs) and trials that have used inappropriate strategies to allocate interventions (i.e. quasi‐randomised studies) because these designs carry a higher risk of bias and do not ensure true random assignment of participants to intervention groups.
Types of participants
We will include participants (> 18 years old) with advanced NSCLC with a KRAS G12C mutation who have already been treated with checkpoint inhibitors, platinum doublet chemotherapy, or a combination of both, and who receive either a G12C‐inhibitor or chemotherapy in a second‐line setting or beyond.
Advanced NSCLC is defined as pathologically confirmed, either stage III disease not suitable for chemoradiation and thus requiring systemic treatment or stage IV according to the ninth edition of the TNM (tumor, node, metastasis) classification of lung cancer [33], or corresponding stages from previous editions.
In the case of a study with only a subset of eligible participants, we will contact the study's authors to retrieve the raw data of these participants. If this is not possible, or if the data cannot be disaggregated, we will exclude the study.
Types of interventions
The planned comparison for this review is a G12C inhibitor versus chemotherapy (any, not specifically specified beforehand) in a second‐line setting and beyond.
We will include any type of G12C inhibitor (regardless of dosage/duration/frequency). Sotorasib and adagrasib are G12C inhibitors that are already available; in the near future, we expect that more trial data with other G12C inhibitors will become available.
Outcome measures
Critical outcomes
Overall survival (OS) at 3, 6 and 12 months, defined as the time from start of treatment until death from any cause.
Progression‐free survival (PFS) at 3, 6 and 12 months, defined as the time from start of treatment until disease progression.
Important outcomes
Survival rate at specific time points: 3, 6, 12 and 24 months
Overall response rate
Adverse events (AE), defined as grade 3 to 5 toxicities according to the Common Terminology Criteria for Adverse Events (CTCAE) [34]. We will include all reported grade 3 to 5 AEs occurring at any time during the study period, including follow‐up, up to at least 28 days after the final study dose.
Patient‐reported health‐related quality of life (HRQoL) by validated scales: EORTC QLQ‐C30 and EORTC QLQ‐LC13 (questionnaires from the European Organisation For Research and Treatment of Cancer [35]) at 12 weeks. We will specifically assess global health, physical functioning, and dyspnoea, ranked in this order.
We searched the Core Outcome Measures in Effectiveness Trials (COMET) initiative database to see whether there was a core outcome set relevant to this patient population, but this was not found [36]. However, these outcomes are in line with trials already published in this population.
Search methods for identification of studies
We will search for all possible comparisons formed by the interventions of interest according to the methods in the Cochrane Handbook for Systematic Reviews of Interventions [37].
Electronic searches
We will conduct searches to identify RCTs in the following electronic databases.
MEDLINE (accessed via PubMed) from 1951 to date
Embase (accessed via Elsevier) from 1947 to date
Cochrane Central Register of Controlled Trials (CENTRAL), the Cochrane Library, current issue
We will search these databases according to the search strategies proposed in Supplementary material 1, which were designed by the Information Specialists of the Cochrane Lung Cancer Group and by author YH.
We will search the databases using appropriate controlled vocabulary (e.g. MeSH, Emtree) and free‐text terms. We will perform the MEDLINE searches according to the Cochrane Highly Sensitive Search Strategy, as described in chapter 4 of the Cochrane Handbook for Systematic Reviews of Interventions [38].
Searching other resources
We will search the reference lists of the included trials by hand to identify any additional eligible studies.
We will conduct searches (from 1947 to date) in the US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov) and the EU Clinical Trials Register at www.clinicaltrialsregister.eu to identify any ongoing trials.
We will also check meeting libraries of the following conferences (from 2023 to date) for eligible studies.
American Society of Clinical Oncology (ASCO)
European Society for Medical Oncology (ESMO)
World Conference on Lung Cancer (WCLC)
We will also examine any postpublication amendments, relevant retraction statements/notices and errata of studies for information, as described in section 4.4.6 of the Cochrane Handbook for Systematic Reviews of Interventions [38]. We will do this for studies that we consider eligible for inclusion in the review. For future updates of the review, we will do this for all studies.
Data collection and analysis
Selection of studies
Two review authors (AN and OA) will independently screen titles and abstracts retrieved from our search strategies for possible inclusion in the proposed review, according to the instructions in chapter 4 of the Cochrane Handbook for Systematic Reviews of Interventions [38]. We will retrieve full‐text articles of eligible or potentially eligible studies, which AN and OA will then independently screen. Any disagreement in this process will be resolved through discussion, and if necessary, we will consult a third author (WvG) to reach consensus. The results of our search will be managed in Rayyan [39]. If translation is needed, we will ask a native speaker or use AI‐based translation software.
Review authors who are involved in potential included studies will not be involved in study selection, data extraction or risk of bias assessment for their own study or be involved in the GRADE assessment that uses outcome data from their own study.
Data extraction and management
With the use of an electronic data collection form that we will pilot‐test ahead of the extraction, two authors (AN, OA) will independently document study characteristics and outcomes from the included studies for this review. We will cross‐check the data and, if necessary, a third author (WvG) will be consulted to reach consensus. We will then transfer data to the Cochrane Review Manager (RevMan) software [40].
We will extract the following data from each included study.
General information: authors, journal of publication, year of publication
Methods: study design, study setting, inclusion and exclusion criteria, duration of trial (start and end dates) and duration of follow‐up, number of trial centres and their locations
Participants: inclusion and exclusion criteria, number of participants, gender, age, performance status; and if available: prior treatment(s), treatment line
Interventions: type of G12C‐inhibitor, agent used as comparator; with no limitation on duration or number of cycles
Outcomes: results for primary and secondary endpoints, with time points
Miscellaneous: funding sources, notable conflicts of interest of investigators
Risk of bias assessment in included studies
Two review authors (AN, OA) will independently evaluate the risk of bias of each individual study included in the review, according to chapters 7 and 8 of the Cochrane Handbook for Systematic Reviews of Interventions [41, 42], using the RoB 2 guidelines [43]. We will consult a third author (WvG) if necessary.
We will assess the following bias domains.
Bias arising from the randomisation process
Bias due to deviations from intended interventions
Bias due to missing outcome data
Bias in measurement of the outcome
Bias in selection of the reported result
In the domain ‘bias due to deviations from intended interventions’, we will qualify the effect of adhering to the intervention as specified in the trial protocol (‘per‐protocol effect’).
We will assess the different domains as either ‘low risk of bias’, ‘some concerns’ or ‘high risk of bias’ at the study level and for each outcome. We will use signalling questions/tool algorithms. We will reach the overall risk of bias for each outcome within a study by considering all domains relevant to the outcome (i.e. both study‐level entries, such as allocation sequence concealment, and outcome‐specific entries, such as blinding), and we will summarise our judgement in a risk of bias table. In the ‘characteristics of included studies’ table, we will provide a brief statement per included study regarding bias.
Measures of treatment effect
For time‐to‐event data (OS, PFS and time‐specific survival rate), we will use hazard ratios (HRs) to measure treatment effect. We will report HRs with a 95% confidence interval. We will also report median OS, and median PFS if applicable.
If HRs are not reported, we will contact the study authors to try to retrieve the original individual participant data and use these data to calculate the HR, with the help of methods suggested in chapter 6 (section 6.8.2) of the Cochrane Handbook for Systematic Reviews of Interventions [44].
For continuous outcomes (HRQoL), we will use the mean difference (MD) when similar outcome scales are used. When different outcome scales are used, we will use the standardised mean difference (SMD).
For dichotomous outcomes (AE), we will use the risk ratio (RR) and a 95% confidence interval.
Unit of analysis issues
For studies with more than one intervention arm, we plan to only include the relevant study arms and analyse these groups separately. For studies with multiple relevant intervention groups versus the same control group, we will combine these intervention groups to perform a single pair‐wise comparison.
In the case of repeated observations for participants (e.g. OS at different times), we would define several time‐based outcomes, and plan separate analyses for these outcomes. We do not expect to include trials using a cross‐over or cluster‐RCT design.
Dealing with missing data
In the case of missing data, we will contact the trial investigators to obtain this data. If the missing data cannot be obtained, we will analyse only the available data and use chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions as guidance regarding how to deal with the data [45]. When missing data are thought to introduce serious bias (we will use chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions as guidance for this [42]), we will explore the impact of including such studies in the overall assessment of results through a sensitivity analysis. In the Discussion section, we plan to address the possible impact of missing data on the results of the review.
Reporting bias assessment
Review authors AN and OA will both, independently, investigate the risk of reporting bias/bias due to missing results for each included study. If needed, we will consult a third author (WvG) to reach consensus.
To minimise this risk, we will compare the published study data with the original study protocol to identify possible unreported data. If applicable, we will make attempts to obtain these unpublished data.
If more than 10 studies are included in the meta‐analysis, we will use funnel plots to explore the possibility of small study bias and non‐reporting bias.
Synthesis methods
If a sufficient number of studies can be identified that are eligible for the review, we will pool their results in a meta‐analysis, according to recommendations in chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions [45].
We will use the fixed‐effect model/method for meta‐analysis. We chose this method with the assumption that the included participants represent the same population (i.e. all participants have progressive disease after the standard first line of treatment) and undergo the same type of intervention. In other words, we expect that the included studies will be conducted similarly in the same patient groups.
For continuous outcomes and time‐to‐event outcomes, we will use the inverse‐variance method. For dichotomous outcomes, the Mantel‐Haenszel method will be used. However, in the case of rare events, we will use the Peto odds ratio method [46].
If the included studies are too dissimilar, we will not proceed with the meta‐analysis, but will offer a narrative synthesis of the results instead. If, for any other reason, a meta‐analysis cannot be performed, we will summarise the results using tables or describe study outcomes in text for each individual included study.
Investigation of heterogeneity and subgroup analysis
For the assessment of heterogeneity, we will follow the recommendations in chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions [45].
We will check for heterogeneity by visually inspecting funnel plots for asymmetry, and by using the Chi2 test. Within this test, we will use a significance level of 0.10 (rather than 0.05) to determine statistical significance, accounting for the possibility of a small number of included studies. We will use the I2 statistic to assess the impact of heterogeneity on the meta‐analysis.
Using the recommendations in chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions, we will apply the following thresholds for the interpretation of the I2 statistic [45].
0% to 40%: might not be important
30% to 60%: may represent moderate heterogeneity
50% to 90%: may represent substantial heterogeneity
75% to 100%: considerable heterogeneity
If we find substantial heterogeneity, we will then investigate whether clinical or methodological heterogeneity could potentially be the cause for this.
The following factors/variables are considered to be potential causes of heterogeneity, and will therefore be explored by subgroup analyses if we find significant heterogeneity.
G12C inhibitor used in second‐line treatment versus third‐line treatment or beyond
Known brain metastases at baseline versus no known brain metastases at baseline
We will compare subgroups with the use of the formal test for subgroup differences in RevMan [40].
In the case that we are unable to identify a reason for heterogeneity, we will offer a narrative description of the results of the individual included studies.
Equity‐related assessment
We will assess and comment on the geographical location of the included studies. Assessing study locations (e.g. in Asia versus not in Asia) is essential because the prevalence of specific driver mutations in lung cancer differs markedly between regions, influencing targeted treatment strategies. We have not planned any other equity‐based analysis.
Sensitivity analysis
We plan to carry out the following sensitivity analyses (if applicable) to test the robustness of the results of the review.
In the case of missing data that are thought to introduce considerable bias, we will exclude the affected studies to evaluate the impact of them on the overall results.
If we can include three or more studies, we will exclude studies with a high risk of bias.
We will perform a random‐effect analysis, for which we will apply the DerSimonian and Laird method [47], to evaluate whether this generates different results than those generated by our fixed‐effect analysis.
Certainty of the evidence assessment
We will describe the certainty of the evidence for the following outcomes in a summary of findings table.
Overall survival at 3, 6, and 12 months
Progression‐free survival at 3, 6, and 12 months
Grade 3 to 5 adverse events
Patient‐reported health‐related quality of life at 12 weeks
As suggested in chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions [48], we will use the GRADE approach for assessing the certainty or quality of a body of evidence, grading at the levels of 'high', 'moderate', 'low' and 'very low' [49]. This will be independently performed by two review authors (AN and OA), and if necessary, a third author (WvG) will be consulted to reach consensus.
To draw proper conclusions, we will use the five GRADE considerations to assess the certainty of the body of evidence.
Risk of bias
Consistency of effect
Imprecision
Indirectness
Publication bias
For each domain, we will check if there is a reason for downgrading the evidence. Serious reasons will lead to downgrading the certainty by one level and very serious reasons will downgrade the certainty by two levels. We will use chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions as a guideline for this [48].
Consumer involvement
Consumer involvement was ensured through the Dutch national patient organisation for lung cancer patients, Longkanker Nederland, represented by their ambassador Lidia Barberio. They express strong support for the priority and necessity of this review.
Criteria for considering studies for this review
Types of studies
We will include phase III randomised controlled trials (RCTs) studying the effects of G12C‐inhibitors in a second‐line setting and beyond, versus chemotherapy, in adults with advanced or metastatic lung adenocarcinoma.
We will apply no restriction on language or publication status. We will also include unpublished online data and meeting abstracts, but only when sufficient data are available and a proper risk of bias assessment can be made.
Cross‐over RCTs will not be eligible in this review because of possible bias by 'carry‐over' effects. Cluster‐randomised RCTs will not be eligible because, in these trials, the unit of randomisation is typically a group, rather than individual participants. This contrasts with the focus of this review, which aims to assess outcomes for individual participants. We will also exclude non‐randomised studies of interventions (NRSIs) and trials that have used inappropriate strategies to allocate interventions (i.e. quasi‐randomised studies) because these designs carry a higher risk of bias and do not ensure true random assignment of participants to intervention groups.
Types of participants
We will include participants (> 18 years old) with advanced NSCLC with a KRAS G12C mutation who have already been treated with checkpoint inhibitors, platinum doublet chemotherapy, or a combination of both, and who receive either a G12C‐inhibitor or chemotherapy in a second‐line setting or beyond.
Advanced NSCLC is defined as pathologically confirmed, either stage III disease not suitable for chemoradiation and thus requiring systemic treatment or stage IV according to the ninth edition of the TNM (tumor, node, metastasis) classification of lung cancer [33], or corresponding stages from previous editions.
In the case of a study with only a subset of eligible participants, we will contact the study's authors to retrieve the raw data of these participants. If this is not possible, or if the data cannot be disaggregated, we will exclude the study.
Types of interventions
The planned comparison for this review is a G12C inhibitor versus chemotherapy (any, not specifically specified beforehand) in a second‐line setting and beyond.
We will include any type of G12C inhibitor (regardless of dosage/duration/frequency). Sotorasib and adagrasib are G12C inhibitors that are already available; in the near future, we expect that more trial data with other G12C inhibitors will become available.
Outcome measures
Critical outcomes
Overall survival (OS) at 3, 6 and 12 months, defined as the time from start of treatment until death from any cause.
Progression‐free survival (PFS) at 3, 6 and 12 months, defined as the time from start of treatment until disease progression.
Important outcomes
Survival rate at specific time points: 3, 6, 12 and 24 months
Overall response rate
Adverse events (AE), defined as grade 3 to 5 toxicities according to the Common Terminology Criteria for Adverse Events (CTCAE) [34]. We will include all reported grade 3 to 5 AEs occurring at any time during the study period, including follow‐up, up to at least 28 days after the final study dose.
Patient‐reported health‐related quality of life (HRQoL) by validated scales: EORTC QLQ‐C30 and EORTC QLQ‐LC13 (questionnaires from the European Organisation For Research and Treatment of Cancer [35]) at 12 weeks. We will specifically assess global health, physical functioning, and dyspnoea, ranked in this order.
We searched the Core Outcome Measures in Effectiveness Trials (COMET) initiative database to see whether there was a core outcome set relevant to this patient population, but this was not found [36]. However, these outcomes are in line with trials already published in this population.
Search methods for identification of studies
We will search for all possible comparisons formed by the interventions of interest according to the methods in the Cochrane Handbook for Systematic Reviews of Interventions [37].
Electronic searches
We will conduct searches to identify RCTs in the following electronic databases.
MEDLINE (accessed via PubMed) from 1951 to date
Embase (accessed via Elsevier) from 1947 to date
Cochrane Central Register of Controlled Trials (CENTRAL), the Cochrane Library, current issue
We will search these databases according to the search strategies proposed in Supplementary material 1, which were designed by the Information Specialists of the Cochrane Lung Cancer Group and by author YH.
We will search the databases using appropriate controlled vocabulary (e.g. MeSH, Emtree) and free‐text terms. We will perform the MEDLINE searches according to the Cochrane Highly Sensitive Search Strategy, as described in chapter 4 of the Cochrane Handbook for Systematic Reviews of Interventions [38].
Searching other resources
We will search the reference lists of the included trials by hand to identify any additional eligible studies.
We will conduct searches (from 1947 to date) in the US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov) and the EU Clinical Trials Register at www.clinicaltrialsregister.eu to identify any ongoing trials.
We will also check meeting libraries of the following conferences (from 2023 to date) for eligible studies.
American Society of Clinical Oncology (ASCO)
European Society for Medical Oncology (ESMO)
World Conference on Lung Cancer (WCLC)
We will also examine any postpublication amendments, relevant retraction statements/notices and errata of studies for information, as described in section 4.4.6 of the Cochrane Handbook for Systematic Reviews of Interventions [38]. We will do this for studies that we consider eligible for inclusion in the review. For future updates of the review, we will do this for all studies.
Data collection and analysis
Selection of studies
Two review authors (AN and OA) will independently screen titles and abstracts retrieved from our search strategies for possible inclusion in the proposed review, according to the instructions in chapter 4 of the Cochrane Handbook for Systematic Reviews of Interventions [38]. We will retrieve full‐text articles of eligible or potentially eligible studies, which AN and OA will then independently screen. Any disagreement in this process will be resolved through discussion, and if necessary, we will consult a third author (WvG) to reach consensus. The results of our search will be managed in Rayyan [39]. If translation is needed, we will ask a native speaker or use AI‐based translation software.
Review authors who are involved in potential included studies will not be involved in study selection, data extraction or risk of bias assessment for their own study or be involved in the GRADE assessment that uses outcome data from their own study.
Data extraction and management
With the use of an electronic data collection form that we will pilot‐test ahead of the extraction, two authors (AN, OA) will independently document study characteristics and outcomes from the included studies for this review. We will cross‐check the data and, if necessary, a third author (WvG) will be consulted to reach consensus. We will then transfer data to the Cochrane Review Manager (RevMan) software [40].
We will extract the following data from each included study.
General information: authors, journal of publication, year of publication
Methods: study design, study setting, inclusion and exclusion criteria, duration of trial (start and end dates) and duration of follow‐up, number of trial centres and their locations
Participants: inclusion and exclusion criteria, number of participants, gender, age, performance status; and if available: prior treatment(s), treatment line
Interventions: type of G12C‐inhibitor, agent used as comparator; with no limitation on duration or number of cycles
Outcomes: results for primary and secondary endpoints, with time points
Miscellaneous: funding sources, notable conflicts of interest of investigators
Risk of bias assessment in included studies
Two review authors (AN, OA) will independently evaluate the risk of bias of each individual study included in the review, according to chapters 7 and 8 of the Cochrane Handbook for Systematic Reviews of Interventions [41, 42], using the RoB 2 guidelines [43]. We will consult a third author (WvG) if necessary.
We will assess the following bias domains.
Bias arising from the randomisation process
Bias due to deviations from intended interventions
Bias due to missing outcome data
Bias in measurement of the outcome
Bias in selection of the reported result
In the domain ‘bias due to deviations from intended interventions’, we will qualify the effect of adhering to the intervention as specified in the trial protocol (‘per‐protocol effect’).
We will assess the different domains as either ‘low risk of bias’, ‘some concerns’ or ‘high risk of bias’ at the study level and for each outcome. We will use signalling questions/tool algorithms. We will reach the overall risk of bias for each outcome within a study by considering all domains relevant to the outcome (i.e. both study‐level entries, such as allocation sequence concealment, and outcome‐specific entries, such as blinding), and we will summarise our judgement in a risk of bias table. In the ‘characteristics of included studies’ table, we will provide a brief statement per included study regarding bias.
Measures of treatment effect
For time‐to‐event data (OS, PFS and time‐specific survival rate), we will use hazard ratios (HRs) to measure treatment effect. We will report HRs with a 95% confidence interval. We will also report median OS, and median PFS if applicable.
If HRs are not reported, we will contact the study authors to try to retrieve the original individual participant data and use these data to calculate the HR, with the help of methods suggested in chapter 6 (section 6.8.2) of the Cochrane Handbook for Systematic Reviews of Interventions [44].
For continuous outcomes (HRQoL), we will use the mean difference (MD) when similar outcome scales are used. When different outcome scales are used, we will use the standardised mean difference (SMD).
For dichotomous outcomes (AE), we will use the risk ratio (RR) and a 95% confidence interval.
Unit of analysis issues
For studies with more than one intervention arm, we plan to only include the relevant study arms and analyse these groups separately. For studies with multiple relevant intervention groups versus the same control group, we will combine these intervention groups to perform a single pair‐wise comparison.
In the case of repeated observations for participants (e.g. OS at different times), we would define several time‐based outcomes, and plan separate analyses for these outcomes. We do not expect to include trials using a cross‐over or cluster‐RCT design.
Dealing with missing data
In the case of missing data, we will contact the trial investigators to obtain this data. If the missing data cannot be obtained, we will analyse only the available data and use chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions as guidance regarding how to deal with the data [45]. When missing data are thought to introduce serious bias (we will use chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions as guidance for this [42]), we will explore the impact of including such studies in the overall assessment of results through a sensitivity analysis. In the Discussion section, we plan to address the possible impact of missing data on the results of the review.
Reporting bias assessment
Review authors AN and OA will both, independently, investigate the risk of reporting bias/bias due to missing results for each included study. If needed, we will consult a third author (WvG) to reach consensus.
To minimise this risk, we will compare the published study data with the original study protocol to identify possible unreported data. If applicable, we will make attempts to obtain these unpublished data.
If more than 10 studies are included in the meta‐analysis, we will use funnel plots to explore the possibility of small study bias and non‐reporting bias.
Synthesis methods
If a sufficient number of studies can be identified that are eligible for the review, we will pool their results in a meta‐analysis, according to recommendations in chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions [45].
We will use the fixed‐effect model/method for meta‐analysis. We chose this method with the assumption that the included participants represent the same population (i.e. all participants have progressive disease after the standard first line of treatment) and undergo the same type of intervention. In other words, we expect that the included studies will be conducted similarly in the same patient groups.
For continuous outcomes and time‐to‐event outcomes, we will use the inverse‐variance method. For dichotomous outcomes, the Mantel‐Haenszel method will be used. However, in the case of rare events, we will use the Peto odds ratio method [46].
If the included studies are too dissimilar, we will not proceed with the meta‐analysis, but will offer a narrative synthesis of the results instead. If, for any other reason, a meta‐analysis cannot be performed, we will summarise the results using tables or describe study outcomes in text for each individual included study.
Investigation of heterogeneity and subgroup analysis
For the assessment of heterogeneity, we will follow the recommendations in chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions [45].
We will check for heterogeneity by visually inspecting funnel plots for asymmetry, and by using the Chi2 test. Within this test, we will use a significance level of 0.10 (rather than 0.05) to determine statistical significance, accounting for the possibility of a small number of included studies. We will use the I2 statistic to assess the impact of heterogeneity on the meta‐analysis.
Using the recommendations in chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions, we will apply the following thresholds for the interpretation of the I2 statistic [45].
0% to 40%: might not be important
30% to 60%: may represent moderate heterogeneity
50% to 90%: may represent substantial heterogeneity
75% to 100%: considerable heterogeneity
If we find substantial heterogeneity, we will then investigate whether clinical or methodological heterogeneity could potentially be the cause for this.
The following factors/variables are considered to be potential causes of heterogeneity, and will therefore be explored by subgroup analyses if we find significant heterogeneity.
G12C inhibitor used in second‐line treatment versus third‐line treatment or beyond
Known brain metastases at baseline versus no known brain metastases at baseline
We will compare subgroups with the use of the formal test for subgroup differences in RevMan [40].
In the case that we are unable to identify a reason for heterogeneity, we will offer a narrative description of the results of the individual included studies.
Equity‐related assessment
We will assess and comment on the geographical location of the included studies. Assessing study locations (e.g. in Asia versus not in Asia) is essential because the prevalence of specific driver mutations in lung cancer differs markedly between regions, influencing targeted treatment strategies. We have not planned any other equity‐based analysis.
Sensitivity analysis
We plan to carry out the following sensitivity analyses (if applicable) to test the robustness of the results of the review.
In the case of missing data that are thought to introduce considerable bias, we will exclude the affected studies to evaluate the impact of them on the overall results.
If we can include three or more studies, we will exclude studies with a high risk of bias.
We will perform a random‐effect analysis, for which we will apply the DerSimonian and Laird method [47], to evaluate whether this generates different results than those generated by our fixed‐effect analysis.
Certainty of the evidence assessment
We will describe the certainty of the evidence for the following outcomes in a summary of findings table.
Overall survival at 3, 6, and 12 months
Progression‐free survival at 3, 6, and 12 months
Grade 3 to 5 adverse events
Patient‐reported health‐related quality of life at 12 weeks
As suggested in chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions [48], we will use the GRADE approach for assessing the certainty or quality of a body of evidence, grading at the levels of 'high', 'moderate', 'low' and 'very low' [49]. This will be independently performed by two review authors (AN and OA), and if necessary, a third author (WvG) will be consulted to reach consensus.
To draw proper conclusions, we will use the five GRADE considerations to assess the certainty of the body of evidence.
Risk of bias
Consistency of effect
Imprecision
Indirectness
Publication bias
For each domain, we will check if there is a reason for downgrading the evidence. Serious reasons will lead to downgrading the certainty by one level and very serious reasons will downgrade the certainty by two levels. We will use chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions as a guideline for this [48].
Consumer involvement
Consumer involvement was ensured through the Dutch national patient organisation for lung cancer patients, Longkanker Nederland, represented by their ambassador Lidia Barberio. They express strong support for the priority and necessity of this review.
Supporting Information
Supporting Information
Supplementary materials are available with the online version of this article: 10.1002/14651858.CD016054.
Supplementary materials are published alongside the article and contain additional data and information that support or enhance the article. Supplementary materials may not be subject to the same editorial scrutiny as the content of the article and Cochrane has not copyedited, typeset or proofread these materials. The material in these sections has been supplied by the author(s) for publication under a Licence for Publication and the author(s) are solely responsible for the material. Cochrane accordingly gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, such material.
Supplementary material 1 Search strategies
Supplementary materials are available with the online version of this article: 10.1002/14651858.CD016054.
Supplementary materials are published alongside the article and contain additional data and information that support or enhance the article. Supplementary materials may not be subject to the same editorial scrutiny as the content of the article and Cochrane has not copyedited, typeset or proofread these materials. The material in these sections has been supplied by the author(s) for publication under a Licence for Publication and the author(s) are solely responsible for the material. Cochrane accordingly gives no representations or warranties of any kind in relation to, and accepts no liability for any reliance on or use of, such material.
Supplementary material 1 Search strategies
출처: PubMed Central (JATS). 라이선스는 원 publisher 정책을 따릅니다 — 인용 시 원문을 표기해 주세요.
🏷️ 같은 키워드 · 무료전문 — 이 논문 MeSH/keyword 기반
- A Phase I Study of Hydroxychloroquine and Suba-Itraconazole in Men with Biochemical Relapse of Prostate Cancer (HITMAN-PC): Dose Escalation Results.
- Self-management of male urinary symptoms: qualitative findings from a primary care trial.
- Clinical and Liquid Biomarkers of 20-Year Prostate Cancer Risk in Men Aged 45 to 70 Years.
- Diagnostic accuracy of Ga-PSMA PET/CT versus multiparametric MRI for preoperative pelvic invasion in the patients with prostate cancer.
- Clinical Presentation and Outcomes of Patients Undergoing Surgery for Thyroid Cancer.
- Association of patient health education with the postoperative health related quality of life in low- intermediate recurrence risk differentiated thyroid cancer patients.