Optimal Control of Liver Metastases From Colorectal Cancer With Cetuximab and Hepatic Artery Infusion of Chemotherapy

Primary end-point: incidence of complete macroscopic resections of liver metastases (R0+R1).

Secondary end-points:

• the rate of histologic complete responses,
• the individual rates of R0 and R1 resections,
• the rate and site(s) of relapse in the resected patients throughout the 3-year span that follows hepatectomy,,
• the relapse-free survival curve and median in the resected patients,
• the progression-free and the overall survival in the patients receiving at least 4 full courses of HAI therapy and in all the patients (intent to treat),
• the objective response rate,
• the rate of adverse events,
• the dose intensities over 3, 6 and 9 courses,
• the per-operative and post-operative complications associated to liver surgery.

The study also includes a pharmacokinetic analysis, a translational research and a rest/activity monitoring investigation.

Open, label, European, non randomized, multicenter, phase II study of intravenous cetuximab (ERBITUX®) and hepatic artery infusion of three-drug chemotherapy (irinotecan, 5-fluorouracil and oxaliplatin) using conventional or chronomodulated delivery (according to institution experience) in patients with liver metastases from colorectal cancer.

Patients undergo partial hepatectomy after 6 ± 3 courses of therapy whenever possible. The minimum of 3 and the maximum of 9 courses before liver surgery depend upon results from iterative onco-surgical evaluations. A minimum of 6 and up to 9 additional courses of therapy will be administered after surgery, depending upon results of liver surgery, pathology report and patient's status. Overall, the patients will receive 9 to 18 courses of protocol therapy.

The interval between the last course of cetuximab-HAI chemotherapy and surgery will be 2 to 4 weeks. Post operative treatment will be initiated 2 to 4 weeks after liver surgery.

TRANSLATIONAL RESEARCH:

1. Pharmacokinetics:

   For a subset of 16 patients (8 on conventional administration and 8 on chronotherapeutic delivery), plasma pharmacokinetics of irinotecan, 5-FU and oxaliplatin and main metabolites will be evaluated after the first course.

2. Rest-Activity monitoring:

   Rest-activity will be monitored with a wrist-worn actigraph (Ambulatory Monitoring, USA) for 1 week prior to treatment onset and during the 2 weeks following treatment onset (3 weeks). This evaluation will be repeated before, during and after the 4th treatment course and before during and after 7th treatment course. Participation of the centers to this investigation will be left optional.

   Time series will be analyzed before, during and after chemotherapy course, with the time courses of the following parameters:

   • Autocorrelation coefficient at 24 h (r24)
   • Dichotomy index I\<O
   • Wavelet-based model function at baseline, and deviation from this model function during and after treatment course delivery.
3. Predictive molecular factors:

In primary tumor and/or in metastases obtained any time prior to inclusion and in resected metastases and non tumoral liver:

• EGFR immunohistochemistry and gene expression or amplification and polymorphism.
• K-ras mutations.
• Clock genes polymorphism or mRNA or protein expression.
• Gene polymorphisms or expression for main pharmacology determinants of irinotecan, 5- FU and oxaliplatin efficacy.
• Inflammatory and immune cell subsets infiltration. Whenever possible, a biopsy of a liver metastasis will also be obtained before treatment onset for documenting these translational endpoints.

In serum, upon inclusion and after 3 courses:

• Determination of serum levels of TGFa, EGF, VEGF, IL-6, IL-8, amphiregulin and epiregulin.

In blood cells upon inclusion:

• Constitutive polymorphisms for ADCC (FCII and FCIII), circadian clock and EGFR.

STATISTICAL METHODS AND SAMPLE SIZE:

The main endpoint will be the incidence of macroscopically complete resections of liver metastases (R0+R1). The confidence intervals (CI) for the response rate will be based on the exact binomial distribution.

For the secondary endpoints, rate of histologic complete responses, rate of R0 resections, rate of R1 resections, rate of relapses, rate of objective responses, the results will be estimated with CI based on the exact binomial distribution. Overall survival time (OS) and progression free survival (PFS) will be analyzed by Kaplan Meier curve. Multivariate analysis will be performed using the Cox proportional hazard model. This analysis will include the following factors: center, performance status, gender, liver disease characteristics, leucocyte count and alkaline phosphatases upon inclusion and treatment delivery schedule, Relapse incidence (RI) is defined as the probability of having had a relapse before time t. Death without experiencing a relapse is a competing event. The method of analysis will be therefore the estimation of Cumulative Incidence curve in a competing risk setting. Multivariate analysis will be performed by the Fine \& Gray model. All patients achieving complete remission will be analyzed.

The goal is to increase the rate of complete resections (R0+R1) by 15%, i.e. from 15% (p0=0.15) to 30% (p1=0.30). According to the exact single-stage phase II design method, the trial for accepting that p1 is to be preferred to p0 (with a=5% and b=20%) requires 48 patients assessable for response. To obtain 48 patients assessable for response, 60 patients will be included in the trial.

A major additional effect expected further from this combined treatment modality is the increase of complete histologic responses, a secondary endpoint in this trial, from 5% (with systemic chemotherapy) to 20% with combined cetuximab and 3-drug HAI. This improvement should further reduce the risk of relapse in the liver or outside the liver by 15 %, i.e. from 80% to 65%.

Therefore, the total number of patients will be 60 patients, including 48 assessable and 12 estimated as unassessable for technical reasons (i.e. HAI catheter dysfunction within the first 2 months).