Introduction

The incidence of cholangiocarcinoma is estimated to be 2,000-3,000 cases per year in the United States. Significant progress has been made at the cellular and molecular level of carcinogenesis of this disease but an effective chemotherapy treatment remains elusive. ₁ Complete resection with negative margins remains the single most important prognostic factor for long-term survival. ₂ For patients with no chance of surgical cure, the choices of chemotherapy are very limited. Historically, fluorouracil (5-FU) is the foundation of combination chemotherapy regimens for unresectable and metastatic cholangiocarcinoma. The response rate of fluorouracil as a single agent is about 10%. ₃ In combination regimens, Taieb et al. reported the use of two days of 5-FU combined with cisplatin on day 2 and showed an objective response rate of 34% and weight gain in 45% of the participants. ₄ However, the median progression-free survival of 6.5 months and the median overall survival of 9.5 months do not represent a substantial improvement. An ECF (epirubicin, cisplatin, 5-FU) regimen showed only a 10% partial response rate among patients with cholangiocarcinoma. ₅

Capecitabine is an oral pro-drug of 5-FU. Conversion to the active drug is dependent on thymidine phosphorylase, which is expressed at a higher level in tumor cells than in normal tissue. Results from a randomized trial indicate that single agent capecitabine is at least as effective as 5-FU in metastatic colorectal cancer. ₆ As a single agent, capecitabine has modest activity in hepatobiliary tumors. ₇ Knox et al. reported a median overall survival of 14 months using a combination of Gemcitabine and capecitabine in patients with biliary cancers. This indicates that oral capecitabine could replace infusional 5-FU for the treatment of biliary cancers. ₈

Several clinical trials have shown that irinotecan has activity against biliary cancers. Of the five patients that had a partial response in a phase I trial testing the combination of irinotecan and docetaxel in advanced solid tumors, one had cholangiocarcinoma. ₉ Two of 11 patients who had a partial response in another phase I trial using a combination of irinotecan, oxaliplatin and 5-FU in advanced tumors had cholangiocarcinoma. ₁₀ A case report indicated that single agent irinotecan given at a low weekly dose could produce a long-lasting response in metastatic cholangiocarcinoma. ₁₁

COX-2 enzyme is constitutively expressed in normal hepatocytes but not in normal bile duct epithelium. ₁₂ Studies performed in cell lines indicate that bile acids will induce COX-2 expression in cholangiocarcinoma cells. Therefore, COX-2 over-expression is a potentially important factor in promoting continued proliferation and progression of the tumor. ₁₃ The mechanisms by which COX-2 over-expression modulates other growth regulating factors in cholangiocarcinoma cells are not fully understood, ₁₄ but COX-2 over-expression in cholangiocarcinoma inhibits Fas-mediated apoptosis. ₁₅ Celecoxib has also been shown to induce apoptosis and inhibit cell proliferation in the human cholangiocarcinoma cell line QBC939. ₁₆ In two other cholangiocarcinoma cell lines, C611B ChC and LY294002, Celecoxib blocks phosphorylation of Akt and significantly reduces viability of the cells. ₁₇

We began offering patients with metastatic biliary cancers off-label treatment with a combination of irinotecan, capecitabine, and celecoxib. We report here our observations about this combination.

Material And Methods

Patient Population

Treatment plan and toxicity evaluation

Response Criteria

Statistical methods

Results

Twelve patients were treated with this regimen. The patient characteristics are shown in table 1. The median age of the patients was 56 years (range, 31 to 62 years). Of the 12 patients, 9 were females and 3 were males. 10 patients had cholangiocarcinoma and 2 had gallbladder cancer. All patients were chemotherapy naïve when starting the treatment. All 12 patients had metastatic disease on presentation. The most common site of metastases was liver (10/12) followed by local lymph nodes (5/12).

A total of 117 treatment cycles were delivered, with a range of 1-21 per patient. Based on the tumor marker CA19-9, 7 of the patients had a PR (7/12), including 3 patients who had normalization of CA19-9 (table 1). In seven patients who had sequential CT scan evaluations, the partial response rate was 43% (3/7) and two more patients had stable disease. One example of CT scan response is shown in figure 1. No CR was observed by CT scan criteria. The response duration or progression-free survival, from the time of initiation of chemotherapy to the time of documented tumor progression, was 13 months (Figure 2). Median survival from the date of starting chemotherapy was 17 months (Figure 3). Four patients had irinotecan dosage increased upon disease progression but no further responses were observed.

Toxicity

Discussion

Lin et al reported that concurrent use of Celecoxib with capecitabine could attenuate the toxicity and increase the tumor response. ₁₈ This report and the knowledge of COX-2 over expression in biliary cancers prompted the use of the above described treatment. In 2004, COX-2 inhibitors were linked to an increased cardiovascular risk ₁₉ causing difficulties in obtaining insurance company approval for Celecoxib renewal. The dosage used in our regimen (400 mg daily) is clearly linked to an increased risk of cardiovascular events compared to placebo (hazard ratio of 3.0) in the colorectal adenoma prevention trial. ₂₀ However, because of the grim prognosis of biliary cancer, this should not be an issue when used as a chemotherapeutic agent. Indeed, in our study, patients died of their cancer, and none had cardiovascular events related to the use of Celecoxib. Despite recent data from the human cholangiocarcinoma cell line (CCCLP1) supports the suppression of the COX-2 pathway as a therapeutic target, ₂₁ it is highly unlikely that a randomized trial will ever be carried out to test COX-2 inhibition in biliary cancers.

Irinotecan in combination with capecitabine at low dose is effective and deserves further evaluation. The low toxicity profile is related to the low dose of both irinotecan and capecitabine. The dose intensity of irinotecan in our regimen is 40 mg/m ² /week which is less than 50% of the most commonly used irinotecan dose in the treatment of metastatic colorectal cancer (FOLFIRI with irinotecan at 180 mg/m ² every two weeks or dose intensity of 90 mg/m ² /week). It raises the question of whether a lower dose of irinotecan, with a much-improved tolerability, may also be effective for other tumors. Furthermore, we cannot rule out the possibility that Celecoxib attenuates the toxicity of chemotherapy.

The Capecitabine, irinotecan, and celecoxib regimen, with its low toxicity and long duration of disease control, is ideal for palliation. The risk for cardiovascular events and the increasing difficulty in obtaining insurance approval for Celecoxib could prevent further development of this regimen.

Figure 1

Figure 1a: CT response of one patient.

Figure 2

Figure 1b: CT response of one patient.

Before treatment showing large gallbladder mass with liver metastases.

B. 9 weeks after initiating treatment showing more than 50% reduction in the tumors.

Figure 3

Figure 2: Progression-free survival after treatment for cholangiocarcinoma

Estimate of median time to progression from start of treatment is 13 months

Figure 4

Figure 3: Survival after treatment for cholangiocarcinoma

Estimate of median survival from start of treatment is 17 months

Figure 5

Table 1: Patients Characteristics

Figure 6

Table 2: Toxicity

Correspondence to

Fa-Chyi Lee, MD UNM CRTC 900 Camino de Salud Albuquerque NM 87131 e-mail: flee@salud.unm.edu