Chronomodulated Irinotecan, Oxaliplatin, and Leucovorin-Modulated 5-Fluorouracil as Ambulatory Salvage Therapy in Patients with Irinotecan- and Oxaliplatin-Resistant Metastatic Colorectal Cancer

Abstract

Purpose. To evaluate the activity and tolerability of salvage chronomodulated chemotherapy combining irinotecan (I), 5-fluorouracil/leucovorin (5-FU/LV), and oxaliplatin (O) (chronoIFLO) in patients with metastatic colorectal cancer (MCRC) and prior progression on four drugs.

Patients and Methods. Seventy-seven nonhospitalized MCRC patients received chronoIFLO every 3 weeks, with day 1: I (180 mg/m² over 6 hours, with peak infusion rate at 05:00) and days 2–5: 5-FU/LV (700/300 mg/m² per day over 12 hours, with peak flow rate at 04:00), and O (20 mg/m² per day over 12 hours, with peak flow rate at 16:00). Toxicity and response were assessed every 3 weeks and every 2 months, respectively.

Results. Three or more prior chemotherapy lines were given to 75% of the patients. Two or more organs had metastatic disease in 65% of the patients. A median number of six courses of chronoIFLO was given. The main grade 3–4 toxicities were diarrhea (39% of the patients, 9% of the courses) and neutropenia (30% of the patients and 7% of the courses). Grade 3 peripheral sensory neuropathy occurred in 14% of the patients. Two patients achieved a partial response and 61 had stable disease, resulting in disease control for 82% of the patients. The median time to progression (TTP) was 5.5 months (95% confidence interval, 3.7–6.0). The median overall survival time was 14.2 months (9.8–17.3). Baseline performance status, serum carcinoembryonic antigen (CEA) level, and CEA doubling time were independent prognostic factors of TTP.

Conclusions. ChronoIFLO safely and durably halted tumor progression in most extensively pretreated MCRC patients.

Introduction

The discovery of new anticancer drugs and more effective infusion drug delivery schedules have led to a near doubling of the median survival time of patients with metastatic colorectal cancer (MCRC) over the past decade [1–10]. Novel treatment strategies have combined these effective chemotherapy regimens with resection of metastases, resulting in an unprecedented long-term survival and even cures despite an initially large tumor burden [11, 12].

However, this malignant disease still recurs in the majority of patients who benefit from initially effective chemotherapy [1–12]. Thus, effective salvage treatments that improve disease control and patient survival also play an important role in the management of MCRC patients [8, 13, 14]. Optimal patient management now needs to offer a therapeutic strategy involving patient-adjusted selection and sequencing of available treatment options. The goals here are not only to achieve the best antitumor activity and to resect metastases whenever possible, but also to spare toxicities to normal host tissues so as to be able to offer prolonged effective therapy whenever it is needed.

Toward this goal, our team has adapted chemotherapy delivery according to circadian rhythms (chronotherapy) [1–3, 11, 12, 15]. Chronotherapy of human CRC is based on the demonstration of circadian time dependencies in the pharmacology, toxicology, and therapeutic efficacy of 5-fluorouracil/leucovorin (5-FU/LV), oxaliplatin, and irinotecan in experimental models and/or in cancer patients [1, 2, 15–22]. The control of cell proliferation by the circadian timing system represents an essential mechanism that determines optimal drug delivery timing. For instance, in human bone marrow, skin, and oral and rectal mucosae, DNA synthesis decreases by 50% or more between midnight and 4 a.m., compared with daytime [15, 23, 24]. As a result, irinotecan and 5-FU, two drugs that are primarily toxic for S-phase cells, produce less damage to normal tissues when delivered at night. The detoxification of oxaliplatin is partly mediated by reduced glutathione [25], a tripeptide exhibiting the highest tissue and plasma levels near the middle of the day in humans [26, 27], thereby supporting oxaliplatin administration in the early afternoon [17].

The clinical relevance of chronotherapy was demonstrated in two randomized trials [1, 2]. In a multicenter trial involving 186 patients with previously untreated MCRC, the chronomodulated administration of oxaliplatin and 5-FU/LV was compared with a flat infusion of the same drugs at the same doses and over the same 5-day duration: chronotherapy resulted in a fivefold lower incidence of severe mucositis and an incidence of peripheral sensory neuropathy that was half that seen with nonchronomodulated administration, while the response rate was significantly higher, 51% versus 29% [2].

Because of their different mechanisms of action and nonoverlapping toxicity profiles, studies have explored the combination of irinotecan and oxaliplatin in MCRC [28]. A first-line two-drug combination with irinotecan and oxaliplatin resulted in objective response rates of 35%–43.5% and median times to progression of 6.5–7.1 months in MCRC patients [7, 29]. Adding 5-FU/LV to irinotecan/oxaliplatin as first-line chemotherapy resulted in major objective response rates of 58%–71% and median times to progression of 10.4–13 months in two studies [30, 31]. However, the severe toxicities encountered with these regimens limit their use. In tumor-bearing mice, though, the combination of irinotecan and oxaliplatin was both synergistic and nontoxic when irinotecan and oxaliplatin were administered at their respective least toxic circadian times, that is, near the end of the rest span for irinotecan and near the middle of the activity span for oxaliplatin [20]. Based on the circadian time structure of mouse and human species, these times, respectively, correspond to 5 a.m., that is, in the second half of sleep, and 4 p.m., that is, near the middle of wakefulness in humans.

In this study, we evaluated the tolerability and activity of achronomodulated ambulatory chemotherapy regimencombining irinotecan, 5-FU, LV, and oxaliplatin (chronoIFLO) as salvage therapy in patients resistant to previous oxaliplatin-, 5-FU/LV-, and irinotecan-containing regimens.

Patients and Methods

Eligibility Criteria

Patients with MCRC progressive on oxaliplatin plus 5-FU/LV and on irinotecan administered either as a single agent or combined with 5-FU/LV were admitted to this study. Inclusion criteria consisted of histologically proven, unresectable, locally advanced or metastatic colon or rectal cancer. Eligible patients had measurable lesions, an age of at least 18 years, a World Health Organization (WHO) performance status (PS) score of 0–2, a life expectancy of at least 3 months, and no prior chemotherapy over the previous 4 weeks. No eligible patient had concomitant active infections, severe heart dysfunction, or any malignant tumor other than cured in situ carcinoma of the cervix or nonmelanoma skin carcinoma. Required blood parameters consisted of hemoglobin ≥9 g/dl, platelet count ≥100 × 10⁹/l, neutrophil count ≥1,500 × 10⁶/l, serum creatinine concentration ≤1.25 × the upper limit of normal (ULN), serum total and conjugated bilirubin ≤1.5 × ULN, serum transaminases ≤2.5 × ULN, and serum alkaline phosphatases ≤5 × ULN. Persistent WHO grade 3–4 toxicity from prior chemotherapy regimens was an exclusion criterion.

Clinical resistance to a specific chemotherapy regimen was defined as the occurrence of radiologically documented disease progression while receiving this regimen or within the 3 months following withdrawal. Progression was based on the appearance of new lesions or an increase >25% in the sum of the products of the perpendicular greatest dimensions of measurable disease (WHO). The study was approved by the internal review board of Paul Brousse Hospital.

Pretreatment Evaluation

Pretreatment evaluation consisted of a complete clinical, radiological, and biological workup including baseline thoracic, abdominal, and pelvic computed tomography (CT) scan, liver ultrasound, and serum carcinoembryonic antigen (CEA) and cancer antigen (CA) 19-9 determinations. CEA doubling time was considered as an estimate of tumor growth rate for patients presenting with baseline CEA >5 ng/ml and with at least two baseline CEA determinations 2 weeks apart.

Chemotherapy Regimen and Dose Modifications

A dual-chamber central venous access port was implanted in all patients to deliver chronoIFLO through a programmable-in-time, multichannel, ambulatory pump (Melodie^®; Aguettant, France). Irinotecan (180 mg/m² per day) was infused on day 1, from 02:00 to 08:00 with a nocturnal peak at 05:00. On days 2–5, oxaliplatin (20 mg/m² per day) was delivered from 9:45 to 21:45 with a diurnal peak at 16:00, and 5-FU (700 mg/m² per day) and LV (300 mg/m² per day) were both administered from 21:45 to 9:45, with a nocturnal peak at 04:00 (Fig. 1). The initial daily dose of oxaliplatin was reduced by 5 mg/m² per day in patients with grade 2 peripheral sensory neuropathy from previous oxaliplatin administration according to National Cancer Institute Common Terminology Criteria for Adverse Events v3.0. Courses were repeated every 21 days. The dose intensity (DI) of each drug was calculated over the first 3 (DI-3) and 6 (DI-6) treatment courses as previously described [32]. A scopolamine patch (1 mg) was applied from day 1 to day 3 to prevent irinotecan-induced cholinergic syndrome. Oral antiemetic prophylaxis with anti–serotonin-3 receptors was administered to all patients during the 5 days of chemotherapy. No primary prophylactic G-CSF support was allowed. No steroid medication was administered, because it could interfere with physiologic entrainment of circadian rhythms.

In the absence of WHO grade 2 or greater toxicity after the first course, daily doses of irinotecan, 5-FU, and oxaliplatin were increased by 20 mg/m², 50 mg/m², and 5 mg/m², respectively, at the second course. Irinotecan and oxaliplatin doses were reduced by 30 mg/m² per day and 5 mg/m² per day, respectively, in cases of grade 3–4 diarrhea or hematologic toxicity. Daily 5-FU doses were reduced by 50 mg/m² in cases of grade 3–4 diarrhea. Grade 3 peripheral sensory neuropathy prompted oxaliplatin dose reduction by 5 mg/m² per day. Persistent grade 3–4 toxicities called for treatment delay for a maximum of 2 weeks until clinical and/or hematological recovery, otherwise chronoIFLO was discontinued.

Patient Evaluation

Toxicity was assessed according to WHO criteria before the beginning of each course. Oxaliplatin-induced peripheral sensory neuropathy was assessed according to a scale previously developed by our group [3]. Serum CEA and CA 19-9 levels were determined before each treatment course. A radiological evaluation with thoracic and abdominopelvic CT scan eventually complemented with abdominal ultrasound was performed after every 3 chronoIFLO courses. Tumor response was assessed radiologically by investigators using WHO criteria as per previous experience [1–3, 33]. Two independent radiologists reviewed all imaging assessments, and their results were considered as a quality control of tumor response. No surgery of metastases was allowed during the study period. Patients were withdrawn from the study for severe toxicity, clinical alteration, or documented progressive disease.

Statistical Considerations

All data were analyzed on an intent-to-treat basis using Stata 8 software [34]. Time to progression (TTP) and overall survival (OS) curves were generated using the Kaplan-Meier method. TTP and OS time were calculated from the first day of chronoIFLO administration to the respective date of disease progression or death (or last follow-up). Ten patient characteristics upon inclusion were considered as potential prognostic factors of TTP and survival: gender, age at inclusion (<50, 50–65, >65 years), PS score (0 or 1 versus 2), number of previous chemotherapy lines (2 or 3 versus ≥4), number of organs with metastases (1, 2, or ≥3), baseline serum CEA and CA 19-9 levels (normal [N] versus 1–10× N versus >10× N), CEA doubling time (continuous variable), and presence of liver or lung metastases. Factors displaying differences with a p-value ≤ .15 on univariate analysis were selected for multivariate analysis using the Cox proportional hazards regression model. All statistical tests were two-tailed and a p-value of .05 or less was considered to be significant. The cutoff point for survival and safety data was June 2005.

Results

Patients

From September 1999 to June 2004, 77 consecutive eligible patients were included in the study (Table 1). The median age was 59 years (range, 33–82 years), with 12 patients aged 70 years or more (16%). Sixty-two patients (80%) had a PS score of 0 or 1 at baseline. Fifty patients (65%) had more than one metastatic site, with liver and lungs being involved in 80% and 56% of patients, respectively. Fifty-eight patients (75%) had previously received three chemotherapy lines or more. The prior median cumulated doses of oxaliplatin and irinotecan were 1,100 mg/m² (range, 580–4,040 mg/m²) and 1,440 mg/m² (540–6,250 mg/m²), respectively. The median time elapsed between the beginning of the first chemotherapy line for metastatic disease and chronoIFLO onset was 24 months.

Drug Doses and Toxicities

A total of 530 courses of chronoIFLO was administered, with a median of 6 courses per patient (range, 2–25). Median dose intensities (25th–75th percentiles) remained similar over the first three and six courses for irinotecan (60 mg/m² per week [50–60]), 5-FU (933 mg/m² per week [800–933]), and oxaliplatin (27 mg/m² per week [20–27]).

All courses were assessed for toxicity. Seven patients (9%) experienced major grade 3–4 toxicities requiring hospitalization for persistent grade 3–4 diarrhea (four patients) or for febrile grade 4 neutropenia (three patients, including one with a septic shock). No treatment-related deaths occurred. Six patients (8%), all with a baseline PS score of 2, were withdrawn from the study after two courses of chronoIFLO that produced persistent grade 3 asthenia. The most frequent grade 3–4 toxicities were diarrhea (39% of the patients, including 23% after a single course) and neutropenia (30% of the patients) (Table 2). Grade 3 or 4 mucositis and thrombocytopenia were observed in <10% of the patients. Forty-one patients (53%) experienced cumulative grade 2 or 3 peripheral sensory neuropathy. However, 37 patients had grade 1 or 2 peripheral sensory neuropathy prior to study entry as a result of previous extensive administration of oxaliplatin. PS score remained unchanged in 20 patients (26%) and was one grading unit less for 39 patients (51%) upon progression. Baseline body weight varied by <10% for 70 patients (91%).

Response, TTP, and Survival

Two patients (3%) achieved a partial response (PR), where as 61 patients (79%) had stable disease (SD). Disease continued to progress in 14 patients (18%). The median TTP was 5.5 months (95% confidence interval [CI], 3.7– 6.0 months) (Fig. 2).

After a median follow-up of 38 months (range, 12–69 months), 73 patients died from disease progression (95%). Two patients were alive with metastases. Two patients were alive without any detectable disease, following repeated resections of metastases and postoperative chronochemotherapy. The median overall survival time of the 77 patients was 14.2 months (95% CI, 9.8–17.3 months). The overall survival rates at 6 and 12 months were 79% (68%–87%) and 53% (41%–64%), respectively (Fig. 3).

Further Treatments

Fifty patients (65%) received further chronochemotherapy after withdrawal from chronoIFLO. The regimens consisted of oral capecitabine or i.v. 5-FU combined with oxaliplatin and/or irinotecan for 31 patients, cetuximab and irinotecan for nine patients, and hepatic artery infusions of irinotecan, 5-FU, and oxaliplatin for 10 patients with unresectable liver-only metastases. Six patients had a surgical debulking of their main liver and/or lung metastases.

Prognostic Factors of TTP and Survival

Among the 10 patient characteristics studied, baseline PS score and serum CEA level, serum CEA doubling time, number of metastatic organs, and liver metastases were prognostic factors for TTP on univariate analysis. In the multivariate analysis, baseline PS score, serum CEA level, and CEA doubling time were found to be independent prognostic factors for TTP, with chronoIFLO being more effective in patients with a PS score <2, normal serum CEA level, and CEA doubling time >30 days. Baseline PS score and the number of organs involved with metastases were the only independent prognostic factors of survival in the multivariate analysis (Table 3).

Discussion

ChronoIFLO halted tumor progression for a median duration of 5.5 months in heavily pretreated, oxaliplatin-, irinotecan-, and 5-FU/LV–resistant MCRC patients. The association of all four drugs [7, 30] and the chronomodulation of their delivery most likely accounted for this tumor control in such a patient population. In a phase II study involving 35 MCRC patients, chronomodulated 5-FU/LV and oxaliplatin were combined with a 1-hour morning infusion of irinotecan [35]. That regimen produced an objective response rate of 22.9% and disease stabilization in 42.9% of the patients. However, only 42% of the patients had previously received irinotecan, 5-FU/LV, and oxaliplatin; a stabilization rate of 53.3% and no objective responses were obtained in this patient subset [35]. In our study, all the patients had previously progressed on 5-FU/LV, irinotecan, and oxaliplatin. Despite this, chronoIFLO controlled metastatic disease in 82% of the patients, possibly because of a better antitumor activity of chronomodulated irinotecan. In a previous pilot study, 32 heavily pretreated MCRC patients received chronomodulated 5-FU/LV and oxaliplatin combined with either standard or chronomodulated irinotecan. Disease control (objective response or stable disease) was achieved in 28.5% and 77.8% of the patients, respectively [36]. A circadian rhythm in the transformation of irinotecan into its active metabolite, SN-38, has been demonstrated in mice [19]. Greater SN-38 production was also found in cancer patients receiving chronomodulated irinotecan with a peak at 5 a.m., compared with those receiving a 30-minute infusion between 10 a.m. and 6 p.m. [22]. A multicenter trial of chronoIFLO by the European Organization for Research and Treatment of Cancer Chronotherapy Group is currently determining the relationships between irinotecan timing and the activity and the tolerability of the four-drug regimen as first- or second-line chemotherapy for MCRC.

The median overall survival time of 14.2 months provides further evidence for the benefit of third and further chemotherapy lines in MCRC patients. In a similar patient population of heavily pretreated MCRC, the combination of irinotecan and cetuximab achieved a median survival time of 8.6 months [8]. PS score and the number of metastatic organs were the only independent prognostic factors of survival in our study, in line with other reports [1–9, 37]. The possible selection of better-prognosis patients in this trial cannot be ruled out, because 80% of the patients had a PS score of 0 or 1 and the median time between the start of first-line chemotherapy for metastatic disease and study entry was 24 months. However, poor prognostic factors, such as a rapidly growing neoplasia (median CEA doubling time of 45 days or less), two or more metastatic organs, and progressive disease after three or more chemotherapy lines, also characterize the majority of our study patients. The good baseline PS score could have resulted from the extensive prior administration of chronomodulated chemotherapy regimens in 83% of the patients. This assumption is further supported by the median cumulative doses of oxaliplatin and irinotecan administered prior to study entry (1,100 mg/m² and 1,400 mg/m², respectively), which largely exceeded the doses usually tolerated in current practice with nonchronomodulated regimens.

The overall tolerability of chronoIFLO was acceptable. Despite the fact that our study population was heavily pretreated, the incidence of severe toxicity was similar for diarrhea and two- to threefold less for emesis and neutropenia than with the conventional delivery of these four drugs as first-line chemotherapy [30, 31]. Furthermore, in our study, dose escalation was allowed near maximum-tolerated doses, so that severe toxicity was generally encountered once. Severe asthenia occurred in only 8% of the patients (all with a baseline PS score of 2). No or only minor PS and body weight changes were documented for 77% of the patients. All treatment courses were delivered to outpatients. Taken together, these data added to disease control in 82% of the patients support a clinical benefit of chronoIFLO for most patients.

Having a good PS score, a normal baseline CEA level, and a serum CEA doubling time >30 days was predictive of the best tumor control with chronoIFLO. Of those patients with a PS score of 2, 40% were taken off study for severe toxicities and the remaining ones displayed only minor clinical benefit. Conversely, patients older than 65 years benefited similarly as younger patients from chronoIFLO, without experiencing more adverse events. Older patients in trials usually represent the fit elderly only. In our study also, chronoIFLO was also safe in older patients with a PS score <2. The activity and acceptable tolerability of chronoIFLO could relate to persistent circadian physiology in MCRC patients with a good PS score, as revealed by strong 24-hour rhythms in rest/activity and plasma cortisol in these patients [33, 38, 39]. Conversely, ablated circadian physiology constituted a poor prognostic factor for patients with metastatic cancer from colorectal as well as breast origin [33, 38–40]. Similarly, tumors grew faster in rodent models with circadian physiology that had been ablated through clock gene mutation or environmental alterations [41–43].

Conclusion

Chronomodulated ambulatory irinotecan, 5-FU/LV, and oxaliplatin offered effective and prolonged tumor control, with adequate tolerability and possibly longer survival in heavily pretreated CRC patients with disease resistant to irinotecan, 5-FU/LV, and oxaliplatin. The patients who benefited most were those who displayed a good PS score and had a CEA doubling time estimate exceeding 1 month. ChronoIFLO deserves further evaluation in MCRC, possibly in combination with recently developed active biological agents.

Acknowledgments

This work was supported in part by the Association Internationale pour la Recherche sur le Temps Biologique et la Chronothérapie (ARTBC internationale), Hôpital Paul Brousse, Villejuif, France.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

References