Phase II study of irinotecan (CPT-11) as salvage therapy for advanced nasopharyngeal carcinoma

Abstract

BACKGROUND

This Phase II study was designed to evaluate the efficacy and safety of irinotecan in patients with advanced nasopharyngeal carcinoma (NPC).

METHODS

Patients with disseminated, undifferentiated NPC that progressed during or within 3 months of platinum-based and/or taxane-based regimen were eligible. Irinotecan at a dose of 100 mg/m² was administered on Days 1, 8, and 15 every 28 days, up to a maximum of 6 cycles, until disease progression or the appearance of intolerable toxicity.

RESULTS

Twenty-eight patients were evaluable for toxicity and response. Patient characteristics were as follows: The median age was 46.5 years (range, 40.3–71.6 years), the median number of prior lines of chemotherapy was 2 (range, 1–9), the majority of patients (89%) had good Eastern Cooperative Oncology Group performance status (0–1), and the majority of patients (82.1%) had ≥ 2 sites of distant metastases. A total of 79 cycles of irinotecan with a median of 3 cycles per patient were administered. Toxicity > Grade 3 included neutropenia in 5 patients (17%), anemia in 5 patients (17%), and diarrhea in 4 patients (14%). The best response outcomes were 4 patients (14%) who achieved partial responses and 1 patient (4%) who achieved stable disease. Global quality-of-life scores were stable during treatment. Using the Kaplan–Meier method, the median progression-free survival was 3.9 months, and the median overall survival was 11.4 months. The partial responders had a durable response (range, 5.7–12.2 months).

CONCLUSIONS

Results from this trial suggest that irinotecan is an active salvage agent with modest toxicity in patients with advanced NPC who are refractory to platinum/taxane-based chemotherapy. Studies combining irinotecan with other active agents in the first-line setting are warranted. Cancer 2005. © 2004 American Cancer Society.

The Chinese, especially those from southern China, have the highest incidence of nasopharyngeal carcinoma (NPC).1 Because the predominant ethnic group in Singapore is Chinese (76% of the residential population), mainly comprised of descendants of immigrants from southern China, NPC also is endemic in Singapore, although at a lower incidence. The age-standardized rates for NPC among Singaporean Chinese are 16.7 per 100,000 males and 5.5 per 100,000 females.2–5 NPC is known for its chemoresponsiveness, with consistent overall response rates of 65–75% to platinum-based regimens.6–11 However, this response is not durable, with a time to progression of ≈ 7 months and a median survival of ≈ 11 months. Obviously, there is a need to look for more active novel agents or combinations that can improve on the results of current platinum-based combinations. Furthermore, an analysis of our patient data base revealed that NPC remains chemosensitive after failing first-line treatment, with response rates > 30% to second-line and even third-line salvage regimens.12

Our recent clinical research efforts have shown that paclitaxel alone or in combination with carboplatin and gemcitabine are active agents against advanced NPC and were recent additions to the armamentarium of salvage agents.13–15 It also has been found that irinotecan, which is a topoisomerase I inhibitor, induces apoptosis in NPC cell lines in vitro.16 With its potential noncross-resistance with established chemotherapy in patients with advanced NPC, this sets the stage for the investigation of irinotecan as a salvage agent in NPC. Phase I trials of weekly irinotecan in Japan, with a study population of Asian origin akin to our own patients, established a maximum tolerated dose of 100 mg/m², with myelosuppression and diarrhea as dose-limiting toxicities.17 This weekly dosing schedule, which we adopted for the current study, allowed for higher dose intensity relative to irinotecan administration every 3–4 weeks. Although various Phase II trials have demonstrated the efficacy of irinotecan in solid tumors, such as small cell and nonsmall cell lung carcinoma, colorectal carcinoma, gastric carcinoma, ovarian carcinoma, cervical carcinoma, and head and neck carcinomas (squamous cell carcinoma variety), its activity in nasopharyngeal carcinoma has not been assessed previously.18, 19 The objective of the current study was to evaluate the efficacy and safety of irinotecan used on a weekly basis as single-agent salvage in patients with advanced NPC who had disease progression during or after platinum-based and/or paclitaxel-based regimens.

MATERIALS AND METHODS

Eligibility Criteria and Pretreatment Evaluation

Eligibility criteria included a histologically confirmed diagnosis of NPC, an Eastern Cooperative Oncology Group performance status of 0–2, age 18–70 years, a life expectancy ≥ 3 months, no prior use of irinotecan, no prior radiotherapy to site(s) of measurable disease, and at least 1 bidimensionally measurable (≥ 2 cm) site of disease. A wash-out period of at least 4 weeks from the last dose of prior chemotherapy was required before the administration of the first dose of irinotecan. Other inclusion criteria included adequate hematologic, renal, and liver function using standard laboratory measurements; the absence of brain/leptomeningeal metastases; no history of other malignancy, except treated basal cell and squamous cell skin carcinomas; no active, uncontrolled infection; and no concomitant pneumonitis, large pleural effusion, inflammatory bowel disease, preexisting cardiac disease, or uncontrolled diabetes. Written informed consent was obtained from all patients who met the eligibility criteria prior to enrolment in this study. Our center's ethics committee approved the study protocol. The primary endpoints were the analysis of tumor response and toxicity. Secondary endpoints were quality of life (QoL) during study treatment, progression-free survival, and overall survival. The pharmacokinetic and pharmacogenetic aspects of this study are reported separately.

Treatment Regimen and Dose Modification

Irinotecan was supplied by Aventis Pharma as 5-mL vials containing 100 mg of the drug. Irinotecan was administered on a weekly basis for 3 consecutive weeks followed by a 1-week rest. This 4-week treatment period constituted 1 treatment cycle. The weekly dose was fixed at 100 mg/m² and was diluted in 250 mL of normal saline. It was administered as an intravenous infusion over 90 minutes. Full doses of the drug were given in patients with hemoglobin ≥ 8.0 mg/dL, absolute neutrophil count (ANC) ≥ 1000/mm³, and platelets ≥ 100,000/mm³. In patients with ANC < 1000/mm³ and/or platelets < 100,000/mm³, the dose was withheld. The dose was reduced by 20% in patients with ANC ≤ 1000/mm³ and/or platelets ≤ 100,000/mm³ for ≥ 2 weeks before recovery. For patients with any Grade 3 or 4 nonhematologic toxicity, the dose also was delayed until toxicity resolved to Grade ≤ 2. The patient was evaluated weekly if the dose was delayed, and treatment could be delayed for up to 42 days, beyond which, the patient's participation in the study was discontinued.

Pretreatment, Response, and Toxicity Evaluation

Pretreatment evaluation consisted of baseline studies, including medical history, physical examination, blood chemistries, urine analysis, urine pregnancy test, and electrocardiographic assessment within 7 days of irinotecan administration. Computed tomography (CT) scans or magnetic resonance images of all evaluable and measurable disease sites, in addition to physical examinations, were obtained at baseline assessments (within 21 days of starting trial treatment) and at follow-up assessments (after every 2 cycles of treatment) until Cycle 6 and every 2 months thereafter, upon maximal tumor response or disease stabilization, until disease progression was documented. The treatment response was evaluated according to World Health Organization (WHO) criteria20 for the assessment of chemotherapy efficacy. A complete response was defined as the complete disappearance of all clinical evidence of tumor. A partial response was defined as a reduction ≥ 50% in the sum of the products of the greatest perpendicular dimensions of all measurable lesions for at least 4 weeks. Stable disease was defined as a decrease < 50% or an increase < 25% in well outlined lesions for at least 4 weeks. Progressive disease was defined as an increase > 25% in the cross-sectional area of ≥ 1 lesion or the occurrence of new lesions. Toxicity was evaluated using the WHO toxicity grading scale. Those patients with stable disease or responsive disease received further treatment until disease progression or up to a maximum planned total of six cycles of chemotherapy.

QoL Assessment

Patient QoL during treatment was assessed using the European Organization for Research and Treatment of Cancer QoL questionnaire (QLQ-C30; version 3.0) at study entry, before every cycle, and at every follow-up.21 The global functioning scale was analyzed to investigate the overall level of patient QoL. The scores were transformed to a 0–100 scale, and the “half-rule” was used to impute missing values.22 A higher score denoted better QoL.

Statistical Analysis

Using a Simon 2-stage minimax design,23 a lower activity level (p0) of 0.05, and a target activity level (p1) of 0.20, the planned accrual of 27 patients was sufficient to test this hypothesis (type I error, 0.05; type II error, 0.20). At the end of the first stage, at least 1 response had to be found in 13 patients before continued accrual, and efficacy would be claimed if ≥ 4 responses were observed in the 27 patients. The time to disease progression was defined as the time from the date of initiating study treatment to the date of documented disease progression, whereas overall survival was calculated from the date of enrolment into the study to the date of death. All analyses were performed using SAS software (version 8; SAS, Inc., Cary, NC). Survival was estimated using the Kaplan–Meier method. QoL data were analyzed using mixed-model methodology, which adjusted for bias due to missing values that depended on observed QoL scores.24

RESULTS

Patient Characteristics

From September, 2001, to November, 2003, 30 patients with advanced NPC were enrolled into the study. Two patients were deemed ineligible according to our internal compliance audit unit shortly after enrolment, because trial treatment started within 4 weeks of their prior chemotherapy treatment. These two patients were withdrawn from the study. The characteristics of the remaining 28 patients who met eligibility criteria are listed in Table 1. The majority of patients were Chinese and male, and all patients had WHO Type III NPC. Eighteen patients had distant metastases without evidence of local recurrence, and the remaining 10 patients had evidence of both locoregional disease and distant metastases. The majority of patients (82%) had multiple sites of distant metastases. Eighteen patients had received prior radiotherapy to the locoregional primary site of disease, but only 1 patient had received previous radiotherapy to bony metastases within a bone marrow site. A median of 2 lines of chemotherapy (range, 1–9 lines of chemotherapy) was administered prior to entry to the study. Prior chemotherapy agents included 5-fluorouracil, taxane, gemcitabine, and platinum-based drugs in 18 patients (64.3%), 20 patients (71.4%), 25 patients (89.3%), and 28 patients (100%), respectively.

Table 1. Characteristics of the Enrolled Patients

Characteristic	No. of patients (%)
AJCC: American Joint Committee on Cancer; ECOG: Eastern Cooperative Oncology Group; SD: standard deviation; BM: bone marrow. a Higher numbers denote worse performance status.
Total no. of patients	28 (100.0)
Gender
Male	24 (85.7)
Female	4 (14.3)
Race
Chinese	26 (92.9)
Malay	1 (3.5)
Indian	1 (3.5)
Age (yrs)
Mean ± SD	50.2 ± 7.5
Range	40.3–71.6
Pathology
Undifferentiated carcinoma	28 (100.0)
AJCC stage (6th edition)
Stage IV C	28 (100.0)
Locally advanced disease with distant metastases	10 (35.7)
Distant metastases only without local disease	18 (64.3)
Metastatic disease at initial diagnosis	5 (17.9)
ECOG performance statusa
0	5 (17.9)
1	20 (71.4)
2	3 (10.7)
Prior treatment
Mean ± SD no. of lines of chemotherapy	2.0 ± 1.8
Radiotherapy to locoregional disease	18 (64.3)
Radiotherapy to bony metastases involving BM sites	1 (3.6)
Sites of metastases
Lung	19 (67.9)
Liver	20 (71.4)
Bone	12 (42.9)
>1 Site of distant metastases	23 (82.1)

Response

For the intent-to-treat analysis, all 28 patients were evaluated for efficacy. Early discontinuation of study treatment occurred before first evaluation in two patients. One patient developed severe, intractable diarrhea (Grade 4) and was withdrawn from the study after the first cycle of chemotherapy, and the other patient withdrew for personal reasons. Results of the best overall objective response are summarized in Table 2. According to the intent-to-treat analysis, the best overall response rate was 14% (4 of 28 patients). Objective responses were as follows: Four patients achieved a partial response after 2 cycles, 1 patient achieved stable disease after 4 cycles, and 23 patients had progressive disease. Because, according to the protocol, we had 26 observations after the 2 withdrawals, the per-protocol analysis response of 4 of 26 patients satisfied our study definition of efficacy for irinotecan in patients with NPC based on a Simon minimax design.

Table 2. Best Overall Objective Response

Response	No. of patients (%)
No. of patients evaluable	28 (100.0)
Complete response	0 (0.0)
Partial response	4 (14.3)
Stable disease	1 (3.6)
Progressive disease	23 (82.1)
Disease control (partial response and stable disease)	5 (17.9)

Survival Outcome

The median follow-up was 7.6 months (range, 0.3–21.5 months). The median time to disease progression was 3.9 months (range, 0.3–12.2 months). The majority of patients (68%) had disease progression after completing 2 cycles of irinotecan. The median overall survival was 11.4 months (range, 0.3–21.5 months). The progression-free survival and overall survival Kaplan–Meier plots are depicted in Figures 1 and 2, respectively. The 4 patients who responded to irinotecan had a durable response ranging from 5.6 months to 12.2 months.

Figure 1. Kaplan–Meier plot for progression-free survival.

Figure 2. Kaplan–Meier plot for overall survival.

Toxicity

The median number of cycles of chemotherapy was 3 cycles (range, 1–6 cycles), and a total 235 doses of irinotecan were administered. Table 3 summarizes the hematologic and nonhematologic toxicities. There was 1 patient with Grade 5 neutropenia who died from neutropenic sepsis after the first cycle of irinotecan. This patient was pretreated heavily with three prior lines of chemotherapy. The most common toxicities experienced were diarrhea (82%), anemia (64%), and neutropenia (57%). Transient Grade 1–2 transamnitis was observed in 13 patients but did not cause symptoms and resolved with treatment delays of 1–2 weeks. Grade 3 toxicities included neutropenia in 4 patients (14%), anemia in 5 patients (17%), and diarrhea in 3 patients (11%). There was only 1 patient who experienced Grade 4 toxicity, which was late-onset diarrhea that necessitated early termination of study treatment after 1 cycle of irinotecan.

Table 3. Toxicity Profile: Highest Toxicity Grade per Patient

Toxicity	No. of patients (%): WHO grade (n = 28 patients)
	1	2	3	4	5
WHO: World Health Organization.
Hematologic
Neutropenia	8 (28.6)	5 (17.9)	4 (14.3)	0 (0.0)	1 (3.6)
Anemia	6 (21.4)	8 (28.6)	5 (17.9)	0 (0.0)	0 (0.0)
Thrombocytopenia	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Hepatic
Alkaline phosphatase	7 (25.0)	5 (17.9)	2 (7.1)	0 (0.0)	0 (0.0)
Alanine transaminase	5 (17.9)	3 (10.7)	0 (0.0)	0 (0.0)	0 (0.0)
Aspartate transaminase	8 (28.6)	5 (17.9)	0 (0.0)	0 (0.0)	0 (0.0)
Bilirubin	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Gastrointestinal
Diarrhea	5 (17.9)	15 (53.6)	2 (7.1)	1 (3.6)	0 (0.0)

QoL

Because the patients who received more cycles of treatment tended to have a better baseline QoL, those patients with good QoL gradually dominated the sample population and conferred a statistical bias in the global QoL score-naïve analysis. If this bias is not taken into account, then it would appear that the QoL score increased by 2.18 points per cycle (P = 0.071). By classification of the study population into three groups, according to attrition per two cycles of treatment, and using a mixed-model approach described by Fairclough24 to adjust for this bias, the global QoL scores were stable over the duration of treatment compared with baseline scores. This is illustrated in Figure 3. There was a negligible decline of 0.78 points per cycle, which was not significant (P = 0.440). This shows the minimal impact of irinotecan on the QoL of the patients during treatment.

Figure 3. Graph showing European Organization for Research and Treatment of Cancer QLQ-C30 global quality-of-life (QoL) scores over progressive cycles of irinotecan. The patients were classified according to attrition per two cycles of irinotecan to remove drop-out bias in the analysis.

DISCUSSION

The results of the current study confirm the single-agent activity of irinotecan in heavily pretreated patients with advanced NPC. The response rate of 14% was similar to the results reported from Phase II studies of single-agent irinotecan administered to pretreated patients with other advanced solid malignancies, such as gastric carcinoma, nonsmall cell lung carcinoma, and colorectal carcinoma,25–27 which showed response rates of 12.5%, 16%, 14–22%, respectively. The main toxicities of leukopenia and diarrhea encountered in our study were similar to a Japanese Phase II study in patients with previously untreated nonsmall cell lung carcinoma that used the same weekly dosing schedule of irinotecan administration.28 The incidence of neutropenia and diarrhea in the current study was lower than reported in the Japanese study results of Grade 3 or 4 neutropenia (25%), and diarrhea (21%). This may be due to random chance and may not be a true significant difference as a result of the small sample size. The current QoL assessment revealed that irinotecan administered in this weekly dosing schedule did not affect the QoL of patients significantly during treatment. This is most important in the palliative setting. The median progression-free survival of 3.9 months and the median overall survival of 11.4 months observed in the current study were favorable considering the fact that most of our patients were late in the course of refractory and progressive, metastatic disease. It is noteworthy that the durable responses seen in those who achieved partial remission ranged from 5.6 months to 12.2 months. Although we acknowledge the presence of selection bias in this Phase II trial, these survival outcomes are encouraging when viewed with the knowledge that a progression-free survival of ≈ 7 months and an overall survival of ≈ 11 months were achieved with first-line cisplatin-based regimens in patients with advanced NPC.

In conclusion, the use of irinotecan in combination with other active agents for first-line therapy in patients with disseminated NPC should be given due consideration in view of the limited efficacy of current active regimens. Single-agent irinotecan is active as a salvage agent in patients with advanced NPC who are refractory to or who have been treated previously with platinum/taxane-based regimens with modest toxicities.