Circulating tumour DNA in patients with colorectal cancer undergoing curative-intent LIver MEtastases Surgery in France: a prospective, multicentre GERCOR G-118 CLIMES PRODIGE 77 cohort study protocol

Abstract

Introduction

Curative-intent resection of colorectal cancer (CRC) liver metastasis (CRLM), combined with perioperative chemotherapy, is the standard of treatment for selected patients. However, accurately predicting which patients will benefit from this strategy remains challenging. Circulating tumour DNA (ctDNA) has emerged as a promising non-invasive biomarker for detecting minimal residual disease and predicting prognosis. This study aims to evaluate the prognostic value of ctDNA in patients with CRC undergoing surgery for CRLM.

Methods and analysis

CRC patients undergoing LIver curative-intent MEtastasis Surgery is a prospective multicentre, observational cohort study designed to enrol 232 patients with upfront or potentially resectable CRLM. All patients will receive standard-of-care treatment. Serial blood samples will be collected at multiple time points: before chemotherapy (baseline, if applicable), before surgery, after surgery and during follow-up. The primary objective is to assess the association between preoperative ctDNA status and disease-free survival. Secondary objectives include evaluating ctDNA dynamics over time, exploring associations with clinical and pathological features and identifying prognostic factors for recurrence and survival. ctDNA will be analysed using targeted next-generation sequencing and digital droplet PCR. Outcomes will be assessed using Kaplan-Meier survival analysis, Cox proportional hazards models and multivariable regression modelling.

Ethics and dissemination

This protocol was approved by the Comités de Protection des Personnes Ouest-I Ethics Committee (N°2022-A02593-40) on 31 January 2023. Study findings will be disseminated through peer-reviewed publications and relevant national and international conference presentations.

Trial registration number

This study was registered in ClinicalTrials.gov (NCT05627681).

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Serial preoperative and postoperative blood sampling will allow longitudinal assessment of circulating tumour DNA (ctDNA) dynamics.

• Standardised protocols for sample collection and ctDNA analysis aim to reduce technical variability.

• As an observational study, residual confounding cannot be fully excluded.

• Long-term outcomes beyond the follow-up period may not be fully captured.

Introduction

Colorectal cancer (CRC) is the third most common cancer worldwide and the second leading cause of cancer-related deaths, with its burden expected to increase in the coming years.¹ Metastatic dissemination, which occurs in up to 50% of cases, represents the primary cause of CRC-related deaths, with the liver being the most frequent metastatic site.² For patients with upfront or potentially resectable colorectal cancer liver metastasis (CRLM), curative-intent resection combined with perioperative chemotherapy is considered the standard option of care. While this strategy has shown to improve disease-free survival (DFS), its benefit for overall survival (OS) has not been clearly demonstrated.³,⁶

The decision to proceed with a curative strategy relies on the assessment of oncosurgical resectability, defined as the ability to achieve a complete macroscopic resection with sufficient postoperative liver function in patients with controlled disease (partial response or stable disease) after neoadjuvant chemotherapy, while preserving quality of life.⁷ However, both the radiological assessment of disease control and its prognostic value remain debated.^{8 9} Furthermore, resectability criteria remain heterogeneous among surgeons and different centres, as international consensus is still lacking.^{10 11}

Considering the morbidity associated with CRLM surgery, the development of robust prognostic tools is essential to refine patient selection. Several clinical risk scores, based on the number and size of metastases, RAS mutational status, carcinoembryonic antigen (CEA) level and the timing of metastatic disease (synchronous or metachronous), have been proposed. However, their predictive accuracy in clinical practice remains limited.^{12 13}

Circulating tumour DNA (ctDNA) has emerged as a promising prognostic factor in patients with CRC. In stage II disease, ctDNA detection and higher molecular tumour burden have been associated with early recurrence and poor OS in a recent prospective trial.¹⁴ In CRLM, several small cohort studies have shown that detectable preoperative and postoperative ctDNA correlates with shorter recurrence-free (RFS) and lower OS rates in patients undergoing curative liver resection.¹⁵,²¹ In a recent study, Kawashima et al retrospectively analysed preoperative and postoperative ctDNA in 53 patients undergoing curative resection of CRLM at a single centre in Japan. Among these patients, 39 experienced recurrence, with significantly poorer RFS (HR 6.99 (3.19–15.33), p<0.0001) and OS (HR 5.19 (2.00–13.43), p=0.0005) associated with postoperatively detectable ctDNA. In multivariate analysis, detectable preoperative ctDNA was also found to be an independent risk factor for recurrence.²¹

Other studies have reported similar associations between ctDNA and recurrence rates in CRLM, ranging from 79% to 100% in the case of ctDNA positivity.¹⁶,¹⁸ Interestingly, Wang et al found that preoperative, postneoadjuvant chemotherapy ctDNA positivity was more predictive of recurrence than postoperative ctDNA.¹⁶ Beyond a binary positive or negative status, several studies have reported associations between higher ctDNA levels, its persistence over time or increasing ctDNA burden and progressively worse outcomes, suggesting a dose-response relationship.^{16 22} Collectively, these findings underscore ctDNA as a strong indicator of minimal residual disease and an early predictor of relapse.

Despite these promising findings, no large cohort studies have systematically correlated ctDNA status and its dynamic changes with other validated prognostic (clinical, radiological and pathological) factors in metastatic CRC. This prospective multicentre cohort study in CRC patients undergoing LIver curative-intent MEtastasis Surgery (CLIMES) aims to evaluate the prognostic value of ctDNA in combination with other established prognostic factors, before and after curative-intent surgery in this patient population.

Methods and analysis

Study designs and setting

CLIMES is a prospective, observational cohort study conducted across multiple academic hospitals and high-volume cancer centres in France, within the framework of the gastrointestinal oncology PRODIGE (Partenariat de Recherche en Oncologie DIGEstive; Digestive Oncology Research Partnership) cooperative group, in collaboration with FFCD (Fédération Francophone de Cancérologie Digestive), GERCOR (Groupe Coopérateur Multidisciplinaire en Oncologie) and UNICANCER Gastrointestinal (UCGI) groups. Each site will screen patients diagnosed with CRC and potentially or upfront resectable liver metastases, according to the inclusion and exclusion criteria. Investigators will be encouraged to include patients consecutively to minimise selection bias. The first participant was recruited on 5 February 2024. The overall trial end date is 30 June 2027. All patients will receive standard-of-care treatment consisting of perioperative chemotherapy followed by curative-intent resection of liver metastases, as determined by local multidisciplinary teams. No experimental treatments are administered as part of the study, and all clinical discussions remain at the discretion of the treating physicians. Outcomes will be analysed according to biomarker-defined subgroups. The primary analysis will compare DFS between patients with detectable ctDNA (ctDNA-positive group) and those with undetectable ctDNA (ctDNA-negative group) before surgery to determine the prognostic value of preoperative ctDNA status. This manuscript has been reported in accordance with the Standard Protocol Items Recommendations for Interventional Trials (SPIRIT).²³

Exposures

Exposure is defined as preoperative ctDNA status (positive vs negative). Participants will be classified as ctDNA-positive or ctDNA-negative at each time point based on ctDNA detection using targeted next-generation sequencing (NGS) and, when indicated, digital droplet PCR (ddPCR). A patient is considered ctDNA-positive at a given time point if any tumour-specific mutation or methylated biomarker is detectable above the assay’s analytical threshold in plasma ctDNA; otherwise, they are ctDNA-negative.

Objectives

The primary objective of the CLIMES trial is to assess the prognostic value of preoperative ctDNA in predicting the risk of recurrence in patients undergoing curative-intent treatment of CRLM. This prospective cohort study aims to evaluate the correlation between the detection of ctDNA and DFS. Secondary objectives include evaluating the dynamics of ctDNA levels during and after treatment (eg, clearance or persistence), assessing the prognostic significance of these changes and exploring associations between ctDNA status and clinical, pathological and biological features. Additionally, the study aims to identify independent prognostic factors for disease recurrence and survival using multivariable modelling.

Eligibility criteria

Inclusion criteria

Patients eligible for inclusion in this trial must meet all the following criteria:

• Age ≥18 years.

• Eastern Cooperative Oncology Group performance status (ECOG PS) ≤1.

• Histologically confirmed adenocarcinoma of the colon or upper rectum that has been resected or is eligible for resection within 6 months.

• CRLM deemed resectable or potentially resectable by a multidisciplinary team meeting (based on CT scans of the chest, abdomen and pelvis or MRI if CT is not possible).

• No evidence of extrahepatic metastases (infracentimetric, non-specific lung lesions <1 cm in diameter and ≤3 in number are allowed).

• Surgical intervention is feasible after short term (<9 cycles or 4 months±1 month) of preoperative chemotherapy for metastatic disease.

• Written informed consent was obtained according to international guidelines and local laws.

• National healthcare coverage available (Protection Universelle Maladie included).

Exclusion criteria

Patients eligible for this trial must not meet any of the following criteria:

• Low or middle rectal cancer.

• Definitively unresectable CLRM, defined by a multidisciplinary team as metastases that cannot be completely resected with negative margins while preserving an adequate functional liver remnant, even after consideration of contemporary surgical and interventional strategies (such as portal vein embolisation, liver venous deprivation, staged resections or combined local ablative techniques).

• History of another primary cancer within the last 5 years, except for non-melanomatous skin cancer and carcinoma in situ of the cervix.

• History of previous hepatic or extrahepatic metastatic disease.

• Prior resection or ablation of CRC metastasis (eg, pulmonary, hepatic or pulmonary radiofrequency ablation).

• More than two surgical procedures required for complete resection of the primary tumour and/or liver metastases.

• Deficient mismatch repair/microsatellite instability tumour treated with immunotherapy.

• Impossibility to collect blood samples before and after the resection.

• Pregnancy or breastfeeding.

• Follow-up deemed unfeasible due to geographical, social or psychological conditions.

• Medical or psychiatric condition or occupational responsibilities that may interfere with protocol adherence.

• Under guardianship, curatorship or other legal protection.

Assessments

Preoperative

A preoperative visit that must take place within 28 days before surgery and the following data must be collected unless otherwise specified:

• Vital signs: including weight and blood pressure.

• ECOG PS.

• Laboratory assessments:

  • Biochemistry: albumin, lactate dehydrogenase, bilirubin, alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase and CEA.

  • Haematology: complete blood count and platelets (including absolute neutrophil count and lymphocytes).

• Tumour assessment following the Response Evaluation Criteria in Solid Tumors (RECIST) V.1.1 guidelines (recommended), including centralised review, date of examination, number and maximal diameter of CRLM, vascular (hepatic and Glissonian pedicles) contact, location (unilobar vs bilobar), disappearance of targeted and/or non-targeted lesions and response to chemotherapy if applicable (complete response, partial response, stable disease and progressive disease).

Postoperative (5–10 days after surgery)

Reports on the surgeries, anatomopathology and hospitalisation details will be collected centrally. Postsurgery reports must include the following data unless otherwise stated:

• Surgical procedure form consent,

• Date and type of surgery,

• Date of discharge,

• Blood transfusion during surgery,

• Length of hospital stay after surgery,

• Laboratory assessment: CEA levels.

Follow-up at 4 months (±1 month) after the last surgery

At 4 months (±1 month) after the last surgery, the following data must be collected:

• Vital status:

  • If alive: recurrence status, date and site of recurrence, date of last tumour assessment according to RECIST V.1.1.

  • If deceased: date and cause of death (disease-related or other).

• Postoperative complications: all complications occurring within 90 days of surgery (Clavien-Dindo classification recommended).

• Laboratory assessments: CEA levels.

Follow-up at 12 months after the last surgery

Twelve months after the last surgery, the following data must be collected:

• Vital status:

  • If alive: recurrence status, date and site of recurrence, date of last tumour assessment according to RECIST V.1.1.

  • If deceased: date and cause of death (disease-related or other),

• Laboratory assessments: CEA levels.

Chemotherapy data collection (no specific visit scheduled)

Three types of chemotherapy treatments are allowed in this study:

1. No preoperative treatment—upfront CRLM resection.

2. Neoadjuvant—up to nine cycles.

3. Interval between two surgeries—up to four cycles.

4. Adjuvant—after the last surgery—six or 12 cycles if neoadjuvant or no adjuvant treatment was received, respectively.

For all patients receiving chemotherapy, the following data will be collected throughout the study without scheduled visits:

• Date of the first and last chemotherapy cycle.

• Total number of cycles received.

• Name of drug administered (oxaliplatin, 5-fluorouracil, CPT-11, bevacizumab, aflibercept, panitumumab and cetuximab).

• Adverse effects.

Extended follow-up up to 2 years after the last surgery

During routine follow-up and up to 2 years after the last surgery, the following data will be recorded in the e-case report form (eCRF) in tabular format:

• Vital status:

  • If alive: recurrence status, date and site of recurrence, date of last tumour assessment according to RECIST V.1.1.

  • If deceased: date and cause of death (disease-related or other).

• Survival follow-up and tumour status: response and progression data.

Blood sampling (BS) for ctDNA analysis

Blood will be collected at predefined study time points for study-specific ctDNA analysis, which is not part of standard clinical care. Samples will be collected alongside routine blood tests whenever possible.

Blood sampling (BS) time points:

1. Preoperative chemotherapy (BS0): optional baseline sample, collected before the first cycle of chemotherapy (whenever possible), at the same time as standard blood tests. This is limited to patients treated at inclusion centres.

2. Preoperative (BSPREOP 1 or 2): collected up to 7 days prior to surgery, at the same time as standard preoperative blood tests. For patients undergoing reverse surgery or two-staged hepatectomy, BSPREOP will be collected before each surgery: BSPREOP1—before resection of CRLM and BSPREOP2—before resection of the primary CRC.

3. Early postoperative (BSPOSTOP 1 or 2): collected 5–10 days after surgery or at hospital discharge if earlier, at the same time as standard postoperative blood tests.

4. Late postoperative follow-up (BSFU): at the end of postoperative treatment, ideally at least 2 weeks (14 days) after the last postoperative chemotherapy cycle. If no adjuvant chemotherapy is administered, the sample will be collected systematically at 4 months (±1 month) after surgery.

Sample volume and processing:

1. ctDNA samples: 3×8.5 mL Streck tubes per time point.

2. Samples are shipped at room temperature within 24 hours to the central laboratory for processing.

3. All labelling, storage and shipment follow prespecified standard operating procedures to ensure traceability and consistency.

Protocol compliance and deviations

Compliance with the study protocol, particularly regarding timing of visits and specimen collection, is expected. In the case of minor deviations such as unexpected preoperative chemotherapy exceeding nine cycles, identification of low- or mid-rectal cancer during restaging or diagnosis of intercurrent extra-hepatic disease during preoperative chemotherapy, patients will remain included in the study, and sample collection will continue. In the event of disease progression during preoperative chemotherapy, even if it contraindicates surgical resection of the primary tumour or liver metastases, a final sample (BS disease progression (BSPD)) will be collected at the time of progression, and no further sampling will be performed. Any deviations from the study plan must be reported to the sponsor, who will determine whether an amendment to the protocol is necessary. Reasons for early discontinuation from the study include withdrawal of consent, screening failure or loss to follow-up. Patients may withdraw from the study at any time and for any reason. In the event of withdrawal, previously collected data and samples will be retained unless the patient explicitly requests otherwise. Patient clinical care will not be affected by study withdrawal.

Protocol adherence and data completeness

To promote adherence to study-related procedures, particularly the timing of blood sample collection and follow-up assessments, the study will implement standardised site initiation training, regular monitoring visits and protocol-specific reminders. To ensure efficient and consistent collection of perioperative blood samples, inclusion centres are limited to sites where surgical resection is performed. This allows preoperative and postoperative samples to be collected on-site. Additionally, the baseline sample (BS0; before chemotherapy) will be collected only for patients who receive chemotherapy at the inclusion centre, thereby avoiding decentralisation of sample collection and preserving sample integrity.

Relevant concomitant care permitted or prohibited during the trial

There are no restrictions on concomitant care during the study. All diagnostic procedures, systemic therapies and supportive treatments deemed necessary by the treating physician are permitted and should follow institutional standards and clinical guidelines. No investigational treatments are administered as part of this observational study. Participation in other interventional trials involving systemic therapies is allowed during the study period.

Provisions for post-trial care

After the study period, patients will continue to receive standard follow-up and treatment according to local clinical practice. In accordance with French regulations, all participants are covered by the sponsor’s insurance policy, which provides compensation for any harm directly related to study-specific procedures.

Outcomes

The primary endpoint is DFS, defined as the time from the date of complete liver resection (final surgery if two-stage strategy (reverse or two-stage hepatectomy)) to the first documented relapse or death from any cause. Patients without an event will be censored at the date of last contact. DFS will be assessed in relation to ctDNA status before surgery. Additionally, the following secondary endpoints will be analysed: OS, defined as the time from study inclusion to death from any cause. Patients who are alive at the time of analysis will be censored at the last known date of being alive, either during the study treatment period or during follow-up; event-free survival (EFS), defined as time between inclusion and first event (progression relapse or death from any cause); rate of ctDNA positivity before surgery, defined as proportion of patients with detectable ctDNA at the preoperative time point (following chemotherapy); ctDNA dynamics, defined as changes in ctDNA status and levels across treatment milestones, based on serial sampling and associations with clinical/pathological factors, defined as correlations between ctDNA status and features, such as RAS mutational status, number and size of metastasis, CEA levels and tumour burden.

Participant timeline

CLIMES CIRCULATE-4 specific assessments schedule is provided in figure 1.

Figure 1. CLIMES CIRCULATE-4 specific assessments schedule. *Only for patients for whom two-stage surgical strategy (reverse or two-stage hepatectomy) was chosen at inclusion.BS, blood sample; BS₀, baseline; BS_FU, follow-up (4 months ±1 month from last surgery); BS_POSTOP1/2, 5 to 10 days post-operative; BS_PREOP1/2, 7 days pre-operative; CLIMES, CRC patients undergoing LIver curative-intent MEtastasis Surgery.

Data collection and management

Biological sampling and biomarker analyses

ctDNA will be analysed at the Centre de Recherche des Cordeliers using targeted NGS with the AmpliSeq Colon and Lung Cancer V2 Panel. When required, ddPCR for methylated biomarkers (WIF1 and NPY) will be performed. Formalin-fixed paraffin-embedded (FFPE) tumour and non-tumour tissues will undergo somatic mutation analysis using validated panels for recurrent CRC mutations. Additional biomarker analyses (angiogenic/inflammatory markers) will be performed on stored plasma samples. All biological sample collection procedures are standardised:

• ctDNA samples: three 8.5 mL Streck tubes per time point, shipped at room temperature within 24 hours.

• Plasma for biomarker analysis: two 6 mL EDTA tubes per time point, centrifuged and stored at −80°C at local centres.

• Tumour/non-tumour tissue: FFPE samples retrieved from surgery centres and centralised at Saint-Antoine Hospital (Paris).

Sample labelling, storage and shipment follow prespecified standard operating procedures to ensure consistency and traceability. Sample tracking will be recorded in both the eCRF and transmittal sheets.

Clinical data and outcome assessment

All baseline and outcome data will be collected using standardised eCRFs. Investigators and trained staff will be responsible for accurate and timely data entry into the secure, web-based data management platform. Perioperative data and postoperative complications (eg, according to the Clavien-Dindo classification) will also be collected and centrally reviewed to ensure data quality and consistency.

Participant retention and follow-up

Participants will be followed for at least 1 year after their last blood sample or until recurrence or death. To promote completeness of follow-up, designated clinical research staff at each study site will monitor participants, with reminders for BS and follow-up visits integrated into local patient management systems. Outcome data, including survival status and recurrence, will be collected even for participants who discontinue standard treatment or deviate from follow-up schedules, unless consent is withdrawn.

Data management

All study data will be entered into a secure, web-based eCRF system using anonymised patient identification codes. Data entry will be performed by site investigators or delegated personnel under the supervision of the monitor. Data will be checked regularly for internal consistency and completeness, with queries issued and resolved through the eCRF platform. A formal data review meeting will be held prior to database lock. The final dataset will be exported into SAS and R formats for statistical analysis by the Methodology and Quality of Life Unit in Oncology (INSERM 1098; Besançon), under the responsibility of a designated statistician. All data will be archived for a minimum of 25 years, in accordance with French regulatory requirements.

Confidentiality

All personal information collected in this study will be handled in accordance with data protection regulations (eg, General Data Protection Regulation). Each participant will be assigned a unique identification code for use in the eCRF. No names or identifiable information will be entered into study databases or transmitted to the sponsor. All data will be stored securely both at study sites and at the sponsor’s facilities, with access restricted to authorised site personnel and auditors. Signed informed consent forms and source documents will be retained at each site in locked files and will be available for audit or regulatory inspection.

Collection, laboratory evaluation and storage of biological specimens

Blood samples for ctDNA (Streck tubes) and plasma biomarkers (EDTA tubes) will be collected at baseline, prior to and after surgery, and during follow-up. Samples will be shipped at ambient temperature (18–25°C) within 24 hours via an approved carrier to the Biological Resource Platform of the CHU de Besançon for central storage. Blood specimens collected in Streck tubes remain stable for up to 7 days under these conditions. Plasma mutation analysis will be performed using targeted NGS and ddPCR. Tumour and non-tumour tissues (FFPE blocks) will be collected during surgery and centralised at Saint-Antoine Hospital (Paris). DNA extraction and mutation analysis will be conducted using validated assays (eg, Qiagen FFPE DNA Kit). Only deidentified samples will be used in ancillary studies. Supplementary ctDNA analysis methods may be explored in predefined subgroups in collaboration with academic or industrial partners.

Sample size

The sample size calculation is based on detecting a difference in DFS between the two groups using a two-sided log-rank test with a 5% type I error rate (α=0.05) and 80% power (β=0.20). The null hypothesis assumes equal DFS between the groups (HR 1), while the alternative hypothesis assumes an HR of 0.56, corresponding to a 44% reduction in the risk of recurrence or death in the ctDNA-negative group compared with the ctDNA-positive group. This effect size corresponds to an absolute increase in 1-year DFS from 40% in the ctDNA-positive group to 60% in the ctDNA-negative group, assuming exponential survival distributions. Based on these assumptions and considering that approximately 20% of patients are expected to have detectable ctDNA, a total of 208 patients will be required to observe 160 events (defined as recurrence or death). With an expected dropout rate of 10%, the final planned sample size is 232 patients. This sample size also allows for multivariate modelling of prognostic factors following the Concato rule, which recommends at least 20 events per variable.²⁴ With an expected 160 events, up to eight variables may be included in the Cox proportional hazards model. These variables may include ctDNA status, number and size of metastases, preoperative CEA level, RAS metastatic status and other clinically relevant factors.

Statistical analysis

Primary and secondary outcomes

The primary analysis will be performed once all enrolled patients have completed at least 1 year of follow-up after liver/colorectal resection or completion of postoperative chemotherapy, whichever occurs later. For the primary objective, DFS will be estimated using the Kaplan-Meier method and compared between groups (ctDNA-positive vs ctDNA-negative) with a two-sided log-rank test (α=0.05). A Cox proportional hazards model will be used to estimate HR and 95% CI for the association between ctDNA status and DFS. Complementary analyses using restricted cubic splines will be applied to explore ctDNA as a continuous variable. Secondary objectives will include analyses of OS, EFS, ctDNA dynamics and correlations with clinical/pathological characteristics. These outcomes will be assessed using descriptive statistics, Kaplan-Meier survival curves and univariate/multivariate Cox proportional hazards regression models, as appropriate. All p values for secondary endpoints will be two-sided and interpreted in an exploratory manner, without adjustment for multiple testing. The full statistical analysis plan will be finalised before database lock. All statistical analyses will be performed using SAS V.9.4 and R software V.2.15.2.

Interim analysis

No formal interim analyses for efficacy or safety are planned, as this is a non-interventional, observational cohort study. No interim hypothesis testing will be conducted, and no statistical stopping rules have been defined. The trial steering committee will review study conduct and data quality, with reviews performed after every 50 enrolled patients or at least every 6 months, whichever occurs first. These evaluations will focus on perioperative data completeness, sample collection rates, cohort distributions and overall study feasibility. The sponsor reserves the right to temporarily suspend or terminate the study at any time for reasons including, but not limited to unexpected or unacceptable AEs, unsatisfactory patient enrolment, significant protocol non-compliance and incomplete or non-evaluable data. Any decision to discontinue or amend the study will be communicated to investigators and relevant authorities as required.

Additional and exploratory analyses

Multivariate Cox proportional hazards regression models will be used to identify independent predictors of recurrence or survival, including ctDNA status. Variable selection will be guided by univariable screening and assessment of collinearity among candidate predictors. Non-linear associations will be explored using fractional polynomials and restricted cubic spline functions, as appropriate. Subgroup analyses will evaluate the prognostic impact of ctDNA dynamics (clearance/persistence) and its association with clinical, biological and pathological characteristics. These analyses will also assess the prognostic value of ctDNA across predefined clinical strata, including the timing of metastasis, RAS metastatic status and surgical strategy. Additional exploratory analyses will examine the associations between ctDNA dynamics and angiogenic or inflammatory biomarkers, with variable selection based on univariate screening and collinearity assessments. Non-linear associations will be explored using fractional polynomials and restricted cubic splines. All subgroup and exploratory analyses will be considered hypothesis-generating and will not influence clinical management within this study.

Analysis of population and missing data

The primary analysis population will include all eligible patients with complete curative-intent liver resection and available preoperative ctDNA data. Missing data are expected to be minimal. Patients with missing primary outcome data will be excluded from the primary analysis. Sensitivity analyses will be conducted to assess the impact of missing data, including worst-case assumptions. No imputation is planned for the primary endpoint, although approaches such as multiple imputation may be considered in sensitivity analyses if necessary.

Oversight and monitoring

Trial Steering Committee and Data Monitoring Committee (DMC)

The study is coordinated by the sponsor, which oversees day-to-day trial operations, regulatory compliance and data management across participating centres. The Trial Steering Committee provides independent supervision and advises the principal investigator and sponsor on all aspects of the study to ensure adherence to good clinical practice (GCP). The committee will meet after every 50 patients are enrolled or at least every 6 months to review data quality, protocol adherence and cohort composition. It may also advise on the continuation of specific subgroups based on scientific or operational considerations. It will assume a Data Monitoring Committee (DMC)-like role for any study procedure-related adverse events (AEs). The final study report will be prepared by the sponsor in collaboration with the coordinating investigator. No DMC or Data Safety Monitoring Board has been established, as CLIMES is a non-interventional, observational cohort study with no investigational treatments.

AE reporting and harms

Although this is a non-interventional cohort study, AEs related to study-specific procedures, such as BS and surgery, will be actively monitored and reported. All AEs and serious AEs (SAEs) will be collected from the date of informed consent until 90 days after the last surgery. Events will be graded using the National Cancer Institute Common Terminology Criteria for Adverse Events V.5.0. Postoperative complications will also be classified according to the Clavien-Dindo system.²⁵ Only SAEs considered related to study-specific procedures (BS or surgery) must be reported to the sponsor, in accordance with French routine care vigilance requirements. AEs and SAEs will be documented in the eCRF and followed until resolution or stabilisation. Causality will be assessed by the investigator using clinical judgement. Deaths not clearly attributable to disease progression must be reported as SAEs. All events will be evaluated and reported according to the European Union regulatory definitions and national vigilance procedures.

Auditing

Audits or inspections may be conducted by the sponsor, independent auditors or regulatory authorities to ensure compliance with GCP, the study protocol and applicable regulations. These may occur during or after the trial and may include review of facilities, documents and source data. Investigators must ensure that required personnel and documents are available during audits. The sponsor will coordinate all audits not conducted independently.

Protocol amendments

Substantial amendments, such as changes to eligibility criteria, outcomes, statistical analyses or study procedures, will be submitted by the sponsor to the relevant ethics committee and regulatory authorities for approval, in accordance with applicable French and European regulations. Once approved, all participating investigators and study sites will be promptly informed of the amendment and provided with updated documents. Clinical trial registries (eg, ClinicalTrials.gov) will be updated where applicable. If an amendment affects participant information or consent, patients will be reconsented using an updated informed consent form.

Patient and public involvement

In this study, patients are involved primarily at the stage of data collection through provision of informed consent. Participants will be fully informed of the study objectives and procedures prior to enrolment.

Ethics and dissemination

The study will be conducted in compliance with the ethical principles outlined in the Declaration of Helsinki. This protocol was approved by the Comités de Protection des Personnes Ouest-I Ethics Committee (N°2022-A02593-40) on 31 January 2023. Written informed consent will be obtained by the referring oncologists or surgeons at the participating centres. All investigators responsible for obtaining consent will be trained in GCP and will be familiar with the study protocol. During surgical procedures, tumorous and non-tumorous tissue samples (paired primary/metastatic lesions and adjacent normal tissue) may be collected for future translational research. All patients will be asked to provide consent for the collection, storage and future use of biological samples at the same time as the main study informed consent. Patients who do not consent to tissue collection may still participate in the main study. Patients who consent to tissue collection may choose whether their samples are stored for future research or destroyed after the study.

Data sharing and access

The full study protocol will be publicly available on publication. Requests for access to deidentified participant-level data and statistical code will be reviewed by the sponsor on a case-by-case basis by sponsor. Access may be granted for legitimate academic research purposes, subject to applicable data protection laws, ethical approvals and signing of a data-sharing agreement. A full list of participating study sites is available from the sponsor on request.

Publication policy

The results of the study will be submitted for publication in peer-reviewed journals and presented at relevant national and international conferences, irrespective of the study outcomes. Authorship will be determined according to the International Committee of Medical Journal Editors criteria, based on substantial contributions to study conception, design, data acquisition, analysis, interpretation and manuscript preparation. Any formal publications or presentations of data collected from this study will be considered joint publications involving investigators and designated representatives of the sponsor. The sponsor, in collaboration with the coordinating investigator, will oversee the preparation of manuscripts and presentations to ensure accuracy, consistency with study data and compliance with applicable regulatory or contractual obligations. All study data are the property of the sponsor. Investigators must obtain written approval from the sponsor before communicating or publishing any study-related data, including abstracts and manuscripts.

Trial status

This protocol corresponds to V.2.1 dated 19 March 2025. Patient inclusion began on 9 January 2024, and the anticipated date of recruitment completion is January 2026.