From September 2004 to February 2006, patients who fulfilled the following eligibility criteria were enrolled in this phase II study. The eligibility criteria for this study were 1) histologically confirmed metastatic colorectal adenocarcinoma without central nervous system (CNS) metastasis, 2) prior palliative chemotherapy with fluoropyrimidine (5-FU, capecitabine), irinotecan and oxaliplatin, 3) documented progression of radiological evidence during chemotherapy with these agents, 4) at least one unidimensionally measurable lesion according to Response Evaluation Criteria In Solid Tumors (RECIST), 5) performance score (PS) of 0-2 according to the Eastern Cooperative Oncology Group (ECOG) scale, and 6) adequate bone marrow function (hemoglobin ≥10.0 g/dL; leukocyte ≥4,000 cells/µL; platelet ≥100,000 cells/µL), adequate liver function (serum bilirubin level ≤2.0 mg/dL; serum transaminase level ≤3 times the upper limit of the normal range) and renal function (serum creatinine level ≤1.4 mg/dL).

Immunohistochemical evidence of the EGFR expression was measured semiquantitatively (>0 on a scale of 0, 1+, 2+, or 3+) in a single reference laboratory. These measurements were performed at the Central Laboratory (Prince of Wales Hospital, Sha Tin, Hong Kong) and graded using a commercially available kit (EGFRpharmDx; Dako Corporation, Carpinteria, CA, U.S.A.) on paraffin-embedded tumor specimens according to the manufacturer's instructions.

This study was designed as a phase II trial. All patients provided written informed consent, and the institutional review boards (IRB) of Asan Medical Center approved the study protocol. In this study, the estimation of the sample size was as below. The Simon two-stage phase II design provided 90% power and a 0.05 level of significance overall to distinguish between the null and alternative hypotheses, where the null hypothesis (H0) is that the true overall response rate is <5%, and the alternative hypothesis (H1) is that the true overall response rate is >25%. The trial would be terminated at nine patients if zero patients have responded or be studied at 30 patients if the trial goes on to the second stage.

Cetuximab (400 mg/m²) was administered intravenously (IV) on day 1 over 2 hr, followed by weekly 1-hr infusions of 250 mg/m². A histamine-receptor antagonist was given as premedication before at least the first infusion. FOLFIRI consisted of irinotecan (150 mg/m²) IV over 90 min, and leucovorin (100 mg/m²) IV over 2 hr, immediately followed by 5-FU (400 mg/m²) IV bolus on day 1, and followed by 5-FU (2,400 mg/m²) by continuous IV over 46 hr. Patients were scheduled to receive biweekly FOLFIRI with weekly cetuximab. If a patient experiences grade 3 of skin toxicity, cetuximab therapy may be deferred for up to two consecutive weeks without changing the dose level. If the toxicity resolves to grade 2 or less by the following treatment period, the treatment may resume. With the second or third occurrences of grade 3 skin toxicity, cetuximab therapy may be deferred again for up to two consecutive weeks with concomitant dose reductions to 200 or 150 mg/m², respectively. Patients had to discontinue cetuximab if more than two consecutive infusions were withheld or a fourth occurrence of grade 3 skin toxicity occurred despite an appropriate dose reduction. Cetuximab was not withheld for FOLFIRI-related toxicity. The dose of FOLFIRI was reduced by 25% if grade 3 or higher non-hematologic toxicities (such as stomatitis or diarrhea) were observed during the treatment period or grade 2 hematologic toxicity (leukocyte 3,000-4,000 cells/µL or platelet 75,000-100,000 cells/µL) was observed immediately prior to administration of the FOLFIRI regimen. If there was grade 3 or higher leukopenia or thrombocytopenia, the FOLFIRI regimen was delayed until recovery.

The primary end point was the rate of confirmed radiologic tumor response, according to RECIST after every three cycles of chemotherapy. Radiologic evaluation consisted of chest radiography and abdominopelvis CT scans. Secondary end-points were the evaluation of time to disease progression (TTP), time to treatment failure (TTF), overall survival (OS) and the safety profile in the intent-to treat analysis. TTP was defined as the time from treatment start to either progression of cancer or death from any cause. TTF was defined as the time from treatment start to either discontinuation of treatment or death from any cause. OS was defined as the time from treatment start to death from any cause. Survival curves were estimated using the Kaplan-Meier method. Safety was assessed in terms of toxicity and evaluated as grades 1 to 4 based on the Common Terminology Criteria for Adverse Events (CTCAE), version 3.0.

To assess the relative significance of potential prognostic factors, we performed univariate and multivariate analyses using the log-rank test and the Cox proportional hazards models, respectively. A p value <0.05 was considered statistically significant, and all analyses were performed using SPSS 12.0 for Windows.

From September 2004 to February 2006, a total of 31 patients met the eligibility criteria; their baseline characteristics are listed in Table 1. Of these patients, 25 (80.6%) underwent surgical resection of their primary tumor and 14 (45.2%) had received more than 2 regimens of palliative chemotherapy. The median number of cycles of cetuximab plus FOLFIRI administered was four (range: 1-23).

The overall response rate (i.e. complete responses [CR]+partial responses [PR] rates) was 25.8% (95% CI, 10.4-41.2%). The median duration of response was 5.4 months (95% CI, 2.1-8.7 months). The disease control rate (i.e. CR+PR+stable disease [SD]) was 58.0% patients (95% CI, 40.6-75.4%) (Table 2).

Of the 31 patients, 11 (33.3%) remained alive at a median follow-up of 13.2 months. The median TTP was 2.9 months (95% CI, 1.4-4.4 months) and the median TTF was 2.1 months. Treatment failure was caused by disease progression (87.0 %), financial burden (6.5%), and inability to tolerate treatment (6.5%). The median OS was 10.9 months (95% CI, 3.8-18.0 months), and the 1-yr OS rate was 47.6% (Fig. 1).

Among the 15 patients whose tumor tissue was available to test for EGFR expression, 13 (86.7%) had tumor cell expression ranging from 1+ to 3+. The presence or degree of EGFR expression did not correlate significantly with clinical response rate (p=0.32) (Table 3).

The 31 patients received 212 cycles of chemotherapy. Safety evaluation showed that the most common hematologic toxicity was neutropenia (54.8%), followed by thrombocytopenia (3.2%). Grade 3 or higher neutropenia developed in 11 (35.5%) patients, but there were no incidents of neutropenic fever or treatment-related mortality. An acne-like skin rash was observed in 25 (80.6%) patients, with grade 3 toxicity in 6 (19.4%). After the sixth administration of cetuximab (median two, range 1-6), almost all patients developed a skin rash. Other common non-hematologic toxicities were mucositis (32.3%), asthenia (22.6%), diarrhea (12.9%), and paronychial cracking (12.9%) (Table 4).

There was a correlation between the presence and severity of the acne-like skin toxicity and response rate and survival. As shown in Table 5, there were superior response rates (p=0.02) and survival rates (p<0.01) with higher grades of skin toxicity.

Univariate analysis of the relationship between survival outcome and clinicopathologic factors showed that the absence of skin rash was significantly associated with TTP, whereas poor performance status and the absence of skin rash were significant negative prognostic factors for OS. Multivariate analysis also identified the absence of skin rash as an independent factor indicative of poor prognosis for TTP, and the poor performance status and the absence of skin rash were independent prognostic factors negatively affecting the overall survival (Table 6).