Whole-body hyperthermia (41.8°C) combined with bimonthly oxaliplatin, high-dose leucovorin and 5-fluorouracil 48-hour continuous infusion in pretreated metastatic colorectal cancer:a phase II study

Original article

Annals of Oncology 13: 1197–1204, 2002 DOI: 10.1093/annonc/mdf216

Whole-body hyperthermia (41.8°C) combined with bimonthly oxaliplatin, high-dose leucovorin and 5-fluorouracil 48-hour continuous infusion in pretreated metastatic colorectal cancer: a phase II study S. Hegewisch-Becker1*, Y. Gruber1, A. Corovic1, U. Pichlmeier3, D. Atanackovic1, A. Nierhaus2 & D. K. Hossfeld1 1

Department of Oncology/Hematology, 2Department of Anesthesiology and 3Institute of Mathematics and Computer Science in Medicine, University Hospital Hamburg-Eppendorf, Hamburg, Germany Received 27 August 2001; revised 17 December 2001; accepted 9 January 2002

Background: Second- and third-line treatments remain a challenge in advanced colorectal cancer. Studies of bimonthly regimens of high-dose leucovorin (LV) and 5-fluorouracil (5-FU) by continuous infusion combined with oxaliplatin (L-OHP) have shown encouraging response rates in patients not responding to a bimonthly LV/5-FU regimen. Hyperthermic enhancement of L-OHP efficiency by increased DNA adduct formation has been demonstrated in vitro. This study was designed to address feasibility, toxicity and efficacy issues of whole-body hyperthermia (WBH) as an adjunct to L-OHP/ LV/5-FU in pretreated patients after progression to first- and second-line treatments with LV/5-FU by continuous infusion and irinotecan. Patients and methods: Forty-four patients with advanced colorectal cancer, who had progressed during or within 3 months after completion of chemotherapy with LV/5-FU 24-h infusion (LV/5-FU24h) (eight patients) or irinotecan combined with or after LV/5-FU24h (36 patients), were treated with L-OHP 85 mg/m2, 2-h intravenous (i.v.) infusion, followed by LV 200 mg/m2, 1-h i.v. infusion, and 5-FU 3 g/m2, 48-h continuous infusion. Every second cycle of the biweekly regimen was combined with WBH, thus allowing a comparison of toxicity with and without WBH in the same patient. Whole-body hyperthermia was administered by a humidified radiant heat device. The target temperature of 41.8°C was maintained for 60 min. L-OHP (2-h infusion) was started at a core body temperature of 39°C. Results: All patients could be evaluated for toxicity, and 41 patients were evaluable for response. A total of 273 L-OHP-containing regimens were administered, 130 with and 143 without WBH. Hyperthermic treatment combined with L-OHP/LV/5-FU showed no unexpected toxicities. WHO grade 3 toxicities were rare and evenly balanced between cycles given with or without WBH. One early death occurred due to sepsis and tumor lysis. The overall response rate was 20%, with two complete and six partial responses. Twenty-three patients (56%) had stable disease and nine patients (22%) progressive disease. With a median observation time of 70 weeks, the median time to progression was 21 weeks [95% confidence interval (CI) 17–25 weeks] and the median survival was 50 weeks (95% CI 39–61 weeks) from the start of therapy. Conclusions: This trial suggests some advantage of combining L-OHP/LV/5-FU with WBH. Results compare favorably with the activity of similar regimens without WBH in less extensively pretreated patients. These data support further evaluation and warrant phase III studies. Key words: advanced colorectal cancer, 5-fluorouracil, leucovorin, oxaliplatin, whole-body hyperthermia

*Correspondence to: Prof. S. Hegewisch-Becker, Department of Oncology/Hematology, Clinic of Internal Medicine, University Hospital Hamburg-Eppendorf, Martinistrasse 52, D-20246 Hamburg, Germany. Tel: +49-40-42803-3971; Fax: +49-40-42803-8054; e-mail: [email protected] © 2002 European Society for Medical Oncology

1198

Introduction Leucovorin/5-fluorouracil (LV/5-FU) applied either weekly or every 4 weeks is still considered standard in the first-line treatment of patients with advanced colorectal cancer. However, due to the availability of several new drugs, especially irinotecan and oxaliplatin (L-OHP), the arsenal for the treatment of colorectal cancer has grown considerably. Irinotecan proved to be superior as second-line chemotherapy in patients with 5-FU-resistant disease in terms of progression-free survival and overall survival (OS), when compared with either infusional 5-FU or best supportive care [1]. In first-line chemotherapy, the addition of LV/5-FU to irinotecan significantly improved response rate, median time to progression (TTP) and OS [2]. Because irinotecan is approved in the United States and Europe for first-line therapy in advanced colorectal cancer, many patients now receive this drug early in the course of treatment, either combined with or after LV/5-FU. Secondand third-line therapy of advanced colorectal cancer in such pretreated patients remains a challenge. A number of clinical trials have demonstrated synergy between L-OHP and LV/5-FU in patients with tumor progression during or after the same LV/5-FU treatment. Various weekly or bimonthly regimens of high-dose LV/5-FU by continuous infusion combined with L-OHP have been studied in patients who had progressed on an LV/5-FU regimen as firstor second-line treatment. The initial trial using the FOLFOX2 regimen led to a response rate of 46% [3]. This response rate could not be reproduced in subsequent trials using schedules of weekly LV/5-FU plus L-OHP or different variations of the biweekly FOLFOX regimen (FOLFOX3/4/6) [4–9]. In these trials response rates between 11% and 31%, and a tumor growth control rate [complete response (CR), partial response (PR) and stable disease (SD)] of 42–55% could be obtained. Preliminary data suggest that response rates in patients pretreated with irinotecan and LV/5-FU may be even worse [10]. Combining heat with antineoplastic drugs has produced suggestive evidence of antitumor synergism. For example, it has been shown that hyperthermia enhances the cytotoxic effect of several drugs, especially of alkylating agents and of platin analogs [11–17]. Furthermore, hyperthermia can overcome acquired drug resistance [13, 18]. Hyperthermic enhancement of the action of platinum compounds has been shown in vitro. The mechanisms involved include increased DNA adduct formation, decreased adduct removal, a rise in cellular drug accumulation and altered DNA repair. Thermal enhancement, defined as the ratio of the dose causing 99% cell death at 37°C compared with 41°C, was found to be 7.3 for cisplatin and 2.0 for L-OHP, with only a slight (cisplatin) or no (L-OHP) further increase at 43°C [12]. Other effects such as interdependent changes of tumor blood flow, pO2 and pH, cytokine induction and induction of heat shock proteins (HSP), modulation of the immune system, and enhanced lymphocyte recruitment to tissues may be contributing factors [19–26].

A phase I study was initiated to investigate pharmacokinetics and toxicities of whole-body hyperthermia (WBH) (41.8°C) alone, of WBH plus carboplatin (CBDCA) and of CBDCA alone. This study confirmed that the combination improved the therapeutic index without increasing CBDCAinduced myelosuppression [27], with the latter phenomenon probably being related to cytokine induction. Since this study, a number of phase I–II studies have produced evidence that WBH may be beneficial in combination with chemotherapy [28–31]. The WBH program was started at the University Hospital in Hamburg in 1997. Since then >400 treatments have been performed. Considering the known hyperthermic enhancement of L-OHP [12], it seemed worthwhile to study the potential of WBH in combination with L-OHP in intensively pretreated patients with advanced colorectal cancer. The design of this clinical trial allowed a comparison of the toxicities of chemotherapy alone versus WBH plus chemotherapy within the same patient.

Patients and methods Patient selection Eligible patients in the age range 18–70 years had to have a diagnosis of histologically proven adenocarcinoma of the colon or rectum, metastatic disease not amendable by surgery or radiation therapy, a WHO performance status of 0–2, a life expectancy of >3 months, and bidimensionally measurable metastases evaluated by computed tomography (CT) scan. In order to be included, patients had to have progressive disease (PD) as defined by radiological imaging according to WHO guidelines (>25% increase of assessable disease or the appearance of new neoplastic lesions during treatment or within 3 months after completion of previous chemotherapy). Preceding treatment consisted of LV/5-FU 24-h infusion (LV/ 5-FU24h), irinotecan in combination with LV/5-FU24h or irinotecan alone after infusional 5-FU-resistant colorectal cancer. Other eligibility criteria included an adequate bone marrow function (white blood cells >3/nl, absolute granulocyte count >1.5/nl and platelet count >100/nl), adequate liver function (total bilirubin <2 mg/dl), normal coagulation profile, adequate renal function (creatinine <1.5 mg/dl or a creatinine clearance of >30 ml/min and blood urea nitrogen <30 mg/dl) and normal electrolytes. Ineligibility criteria included prior L-OHP-containing therapy, relapse after completing 5-FU-based adjuvant therapy, unresolved bowel obstruction or diarrhea, an allergy to lidocaine or evidence of central nervous system metastasis, a severely compromised respiratory status or documented coronary artery disease, a history of angina pectoris, congestive heart failure or serious dysrhythmias. Table 1 gives the demographic profiles and includes previous therapy of all patients studied. The ethics committee of our institution approved the protocol. All patients signed a written informed consent before study entry. Pretreatment evaluation included a complete history and physical examination, a chest X-ray, an abdominal CT and thoracic CT scan if patients had lung metastases. Furthermore, complete pulmonary function tests, ECG, a full chemistry and hematological survey including carcinoembryonic antigen and carbohydrate antigen 19-9 assays were conducted.

1199 Treatment regimen

Table 1. Patients’ characteristics Characteristic

No. of patients

Percent

Patients treated

44

–

Median

57

–

Range

26–65

–

Age (years)

Sex Male

32

73

Female

12

27

Primary tumor Colon

27

61

Rectum

17

39

Site of metastases Liver

41

93

Liver only

22

50

Lung

13

29

7

16

Peritoneal carcinomatosis

5

11

Bone

3

7

Other

5

11

1

24

54

2

15

34

≥3

5

11

0

29

45

1

12

27

2

3

7

Lymph nodes

Involved sites

WHO performance status

Pre-existing symptoms No

20

45

Yes

23

52

1

2

Within normal range

25

57

Elevated

19

43

Unknown Alkaline phosphatases

Elevated markers CEA CA 19-9 Prior initial surgery Prior surgery for metastatic disease Prior adjuvant chemotherapy Prior concurrent chemotherapy and radiation

38

86

35

79

32

73

7

16

17

37

6

14

Type of prior chemotherapy regimens for metastatic disease LV/5-FU24h LV/5-FU24h alone LV/5-FU24h combined with irinotecan LV/5-FU24h followed by irinotecan mono

44

100

8

18

28

64

8

18

Number of prior chemotherapy regimens for metastatic disease 1

21

2

17

48 38

3

6

14

CEA, carcinoembryonic antigen; CA 19-9, carbohydrate antigen 19-9; 5-FU, 5-fluorouracil; LV, leucovorin.

The regimen consisted of a 2-h infusion of L-OHP 85 mg/m2 followed by a 1-h infusion of LV 200 mg/m2, and a 48-h continuous infusion of 5-FU 3 g/m2. Every second cycle of the biweekly regimen was combined with WBH, thus allowing for a comparison of toxicity with and without WBH in the same patient. A maximum of 12 cycles was administered. All patients received ondansetron or granisetron for emetic prophylaxis. Dexamethasone was not allowed as antiemetic medication during WBH cycles since we did not want to interfere with a possible activation of the immune system.

WBH treatment procedure and timing of chemotherapy Whole-body hyperthermia was administered by a humidified radiant heat device (RHS-7500; Enthermics Medical Systems, Inc., Menomonee Falls, WI, USA), exposing the patient to a low-density radiant heat while preventing evaporative heat loss. The radiant heat system for delivering WBH has been previously described in detail [32, 33] the only difference being that our system uses an array of thermocabeling instead of a technology using circulating hot water in a cylinder constructed from copper tubing. A hyperthermia treatment session was defined as raising the patient’s core temperature to 41.8°C. A typical WBH treatment session lasted ∼3.5–4 h, including a mean of 100 min to reach target temperature, 60 min at 41.8°C and a 1-h cooling period. A 2-h infusion of L-OHP was started when a core body temperature of 39°C had been reached and continued until the end of the target temperature. After L-OHP, folinic acid was infused for 1 h followed by 5-FU for 48 h. When the target temperature was achieved, the patient was removed from the radiant heat chamber and wrapped in a blanket and a plastic sheet that served as a vapor barrier to prevent heat loss and maintain a stable plateau phase. Esophageal, rectal, skin and ambient air temperatures were monitored continuously using Mallinckrodt temperature probes (Mon-aTherm TM Skin® and Mon-a-Therm Thermistor 400®; Mallinckrodt Medical, Athlone, Ireland) and Enthermics thermometry software. Before each treatment the probes were calibrated against a thermometer standard (accuracy ± 0.02°C). Heart rate, cardiac rhythm, respiratory rate, oxygen saturation and blood pressure were monitored continuously. Urinary output and electrolytes were checked to ensure fluid and electrolyte homeostasis during and after the procedure. All WBH treatments were performed under general anesthesia. A preceding randomized phase II study had been performed at our institution to identify the most suitable form of anesthesia in this group of patients. Compared with a sedation regimen, general anesthesia proved to be slightly superior regarding oxygen saturation and was less stressful, as shown by a significantly lower increase in the release of adrenaline and noradrenaline. Also, in our experience general anesthesia provided more flexibility regarding the inclusion of patients with less-than-optimal cardiac and pulmonary function. Total intravenous (i.v.) anesthesia was provided with target-controlled propofol infusion, remifentanil and rocuronium. Dopamine, lidocaine and NaCl were administered by continuous infusion to prevent renal malfunction, cardiac dysrhythmias and hypotonic conditions. After induction, the patients were intubated and ventilated with an O2/air mixture (FiO2 >0.4) until they had returned to a core temperature of 39.5°C after the plateau period; i.v. colloid and crystalloid fluids were provided to compensate for fluid loss, and noradrenaline was added when necessary to keep blood pressure within an acceptable range [27].

1200 Safety analysis All patients who had received at least one cycle of L-OHP were evaluated for side-effects. Toxicity was assessed according to WHO criteria [34] except for neurological toxicity and fatigue, which were graded according to an L-OHP-specific scale [35] and to National Cancer Institute common toxicity criteria (NCI-CTC), respectively.

Response analysis To be eligible for tumor-response assessment, patients had to have measurable disease, and received at least four treatment cycles, including two WBH treatments, except in cases of early clinical disease progression. Patients were evaluated for response based on standard criteria for objective regression of measurable lesions [34]. Computed tomography scan or magnetic resonance imaging assessed the antitumor activity every four cycles or earlier when there was clinical deterioration. Complete response is defined as disappearance of all evidence of tumor previously assessed without the development of new malignant lesions lasting for at least 4 weeks. Partial response refers to a >50% decrease in tumor size lasting for at least 4 weeks without the appearance of new lesions. Stable disease is considered to be a <50% regression of measurable disease and a <25% progression of measurable disease lasting for at least 8 weeks, whereas progressive disease is a >25% increase in the size of lesions present at the start of therapy or appearance of new metastatic lesions. Time-to-progression and OS were calculated from the beginning of treatment containing L-OHP until disease progression or death, respectively.

Statistical planning and analysis Sample size was determined according to the optimal two-stage design for phase II studies proposed by Simon [36]. The study was designed for two strata, thus allowing for a separate analysis of the two subgroups of patients pretreated with irinotecan alone or combined with LV/5-FU24h, and those pretreated with LV/5-FU24h only. In each subgroup the null hypothesis tested was a true response probability of <10% (α = 10%). The stratum-specific alternative hypothesis was a true response rate of at least 30%. If the specified alternative hypothesis was true, the probability of rejecting further studies was limited to β = 5%. Based on these criteria, the design required a sample size of n = 33 patients in each subgroup. The first stage consisted of n = 26 patients in each group. If the number of responses was less than four, the trial could be terminated in the specific subgroup. Otherwise, data acquisition had to be continued. If the number of responses after the second stage exceeded five, the group-specific null hypothesis could be rejected. Nominal and ordinal parameters were analyzed via absolute and relative frequencies. For response rates, exact two-sided 95% Clopper– Pearson confidence intervals are presented. Kaplan–Meier methods were applied for ‘time to failure’ variables. Median progression-free survival, and median OS times are presented with their 95% confidence intervals (CI). In addition, 1-year survival rates were calculated. In addition, toxicity was analyzed on a per-patient basis by calculating the cumulative toxicity, which was defined as the worst toxicity grade observed during therapy.

Results From January 1999 to January 2001, 44 patients were enrolled. Table 1 outlines the patients’ characteristics. Thirtytwo patients (73%) had previous definitive surgery; in the adjuvant setting 17 patients (39%) had received prior 5-FU-

containing chemotherapy and six patients (17%) had received radiation therapy. A total of 20 patients (45%) were symptomatic at baseline. The most common symptom was pain related mainly to the hepatic and bone metastases. All patients had been treated previously for metastatic disease. Twentyfive patients (57%) had proof of disease progression while on therapy. The remaining 19 patients had SD but progressed within 3 months of completion of therapy. Of the 41 patients evaluated for response, 16 had received 5-FU-containing chemotherapy in the adjuvant setting. For metastatic disease, seven patients were pretreated with LV/5-FU24h alone, and 34 patients were pretreated with LV/5-FU24h and irinotecan, given either sequentially (eight patients) or simultaneously (25 patients). Nineteen patients had received one and 22 patients had received two previous regimens. Due to slow recruitment of patients pretreated with LV/ 5-FU24h alone, this stratum had to be closed after the inclusion of eight patients. Subsequently, antitumor activity was assessed for the group as a whole and separately for the patients pretreated with LV/5-FU24h and irinotecan. All 44 patients were included in the safety analysis. Fortyone patients met all eligibility criteria and were included in the efficacy analysis.

Extent of exposure and safety A total of 273 cycles were administered, 130 with and 143 without WBH. The mean dose intensity per cycle was 82 mg/m2 for L-OHP (96.5% of the planned dose) and 3 g/m2 (100% of the planned dose) for 5-FU. The average number of treatment courses was six (range two to twelve). Overall, patients tolerated the therapy fairly well. Grade III toxicity was rare and evenly balanced between cycles given with or without WBH. Treatment-related prolongation of therapyfree intervals occurred in 8% of cycles with and 10% of cycles without WBH, with an apparent relationship to the number of preceding cycles. Common toxicities per patient and per cycle are reported in Tables 2 and 3. L-OHP-induced neurosensory toxicities were the most frequent side-effects. Sixty-eight percent of patients reported some symptoms of neuropathy, which were usually mild using the L-OHP-specific scale [35]. Only one patient developed severe sensory neuropathy with functional impairment due to a loss of sensitivity in the fingertips and soles of the feet towards the end of the therapy. Interestingly, almost all patients reported neurotoxicity to be less pronounced in cycles combined with WBH as compared with those given without WBH. Other neurological events included one episode each of a general and focal convulsive attack occurring during WBH treatment. The most common acute adverse events, easily controlled with loperamide and anti-emetics, were grade 1 and 2 diarrhea and nausea/vomiting. Grade 3 diarrhea was rare, affecting only two patients. Severe (grade 3) neutropenia and thrombocytopenia were uncommon (two patients). One patient died of sepsis and pneumonia

1201 Table 2. Toxicity per patient (maximum grade) evaluated for 273 cycles given to 44 patients Side-effects

Grade 0 [n (%)]

1 [n (%)]

2 [n (%)]

3 [n (%)]

4 [n (%)]

2 (5)

0 (0)

a

Hematological Leukopenia

35 (80)

2 (5)

5 (11)

Thrombocytopenia

36 (82)

2 (5)

4 (9)

2 (5)

0 (0)

Anemia

35 (80)

3 (7)

4 (9)

2 (5)

0 (0)

Nausea–vomiting

10 (23)

21 (48)

11 (25)

2 (5)

0 (0)

Mucositis

38 (86)

4 (9)

1 (2)

1 (2)

0 (0)

Diarrhea

30 (68)

7 (16)

4 (9)

3 (7)

0 (0)

10 (23)

22 (50)

11 (25)

1 (2)

0 (0)

Gastrointestinala

Peripheral neurotoxicityb a

WHO criteria. Oxaliplatin-specific scale [35].

b

Table 3. Toxicity per cycle (maximum grade) evaluated for 273 cycles, 130 with and 143 without whole-body hyperthermia (WBH) [values are n (%)] Side-effects

Grade 0

Grade 1

Grade 2 No WBH

WBH

Grade 3

WBH

No WBH

WBH

No WBH

WBH

No WBH

Leukopenia

119 (92)

134 (94)

1 (1)

0 (0)

8 (6)

7 (5)

2 (2)

2 (1)

Thrombocytopenia

125 (96)

135 (94)

1 (1)

2 (1)

2 (2)

2 (1)

2 (2)

4 (3)

Anemia

114 (88)

133 (93)

8 (6)

5 (3)

7 (5)

3 (2)

1 (1)

2 (1)

Hematologicala

Gastrointestinala Nausea–vomiting

87 (67)

99 (69)

26 (20)

30 (21)

14 (11)

12 (8)

3 (2)

2 (1)

Mucositis

129 (99)

139 (97)

0 (0)

2 (1)

0 (0)

1 (1)

1 (1)

1 (1)

Diarrhea

106 (82)

129 (90)

17 (13)

9 (6)

5 (3)

3 (2)

2 (2)

2 (1)

51 (39)

70 (49)

53 (41)

50 (35)

24 (18)

21 (15)

2 (2)

2 (1)

Peripheral neurotoxicityb a

WHO criteria. Oxaliplatin-specific scale [35].

b

following WBH. In this patient the WBH treatment was well tolerated. Three days after WBH the patient developed pneumonia and demonstrated a significant increase in lactate dehydrogenase; thus a tumor lysis syndrome was suspected. The patient died 4 days later. Toxicity related to WBH included 17 episodes of mucosal herpes infection, readily responsive to aciclovir, with a history of mucosal herpes infection being identified as a risk factor in all these patients. We observed only two such episodes in cycles applied without WBH. Pressure sores (WHO grade 1 and 2) developed in three patients, in two cases at the heels and in one case near the tip of the finger, where the measurement of oxygen saturation had been performed using a finger clip. A fatigue syndrome grade 3 and 4 according to NCI-CTC criteria was more often seen in cycles applied with WBH than without (20% and 6% versus 5% and 3%, respectively). We observed five episodes of transient cardiac arrhythmias and

electrocardiographic signs of myocardial ischemia in five patients (WHO grade 3). All episodes occurred during the first WBH treatment at temperatures >41°C and resolved quickly after lowering the patient’s core temperature. To exclude underlying coronary artery disease, further diagnostic (myocardial perfusion imaging under stress) was performed in four patients with negative results. Treatment was continued in these patients, and further WBH showed no cardiac problems. One patient rejected further testing and had to be excluded from the study. Two patients could not receive a second WBH treatment, one because of acute heart failure (WHO grade 3) in the treatment-free interval following the first WBH treatment and the other due to problems following esophageal intubation before the second WBH treatment. Thus, in accordance with response analysis criteria these three patients were not eligible for response.

1202

Antitumor activity The overall response rate (CR and PR) for the 41 patients evaluable for efficacy was 20% (95% CI 9% to 35%) with two CRs and six PRs. In addition, 23 patients (56%) had SD and nine (22%) had PD, thus tumor growth control was achieved in 76% of cases. With a median observation time of 70 weeks (range 22–125 weeks), the median TTP was 21 weeks (95% CI 17–25 weeks). The median OS since the start of protocol treatment was 50 weeks (95% CI 39–61 weeks). The two CRs lasted for 32 and 62+ weeks. One early death occurred due to sepsis and tumor lysis. The crude Kaplan–Meier-estimated probability of 1-year survival was 46%. No substantial differences in response rates were observed between the subgroups of patients who developed PD while receiving therapy and those who progressed within 3 months after completing prior chemotherapy.

Discussion In concordance with earlier reports, the study confirmed that the use of radiant heat provides a relatively non-toxic mode for the application of WBH [27, 28, 30, 31]. No WHO grade 4 toxicities were observed. Whole-body hyperthermia is undoubtedly associated with cardiac stress. Heart rate and cardiac output are known to increase significantly with rising core temperature [32]. Cardiac arrhythmias may occur despite strict exclusion criteria regarding a history of coronary artery disease or dysrhythmias. This underlines the necessity for continuous cardiac monitoring during WBH treatment. An increased incidence of oral herpes simplex in conjunction with WBH has been described previously [27, 28, 31]. Although this has been associated with the administration of ondansetron [27], the nature of this observation clearly requires further investigation. The increased incidence of fatigue observed after WBH cycles is most likely a consequence of the combination of general anesthesia, heat exposure and chemotherapy. Interestingly, there was no evidence of a hyperthermiainduced enhancement of known drug-related hematological and gastrointestinal toxicities, despite the fact that in vitro data have demonstrated a clear thermal enhancement of cytotoxicity for L-OHP [12]. In this context, it is worth mentioning that the addition of WBH to CBDCA has been shown to reduce myelosuppression in an earlier trial [27]. It was speculated that this observation was possibly due to a WBHinduced, but short-lived increase in the production of different cytokines including granulocyte colony stimulating factor and granulocyte–macrophage colony stimulating factor (GMCSF) [20, 21]. The majority of patients had been pretreated with LV/5-FU by continuous infusion plus irinotecan. Results presented here suggest that L-OHP/LV/5-FU combined with WBH is active in this group of patients. A tumor growth control (CR, PR or

SD) in 76% of patients and a TTP of 21 weeks were obtained. These data compare favorably with the activity of similar regimens in less extensively pretreated patients treated with biweekly [8, 9, 37] or weekly schedules [5–7]. The efficacy of L-OHP/LV/5-FU in patients pretreated not only with LV/ 5-FU but also with irinotecan has not been studied in larger trials. Preliminary data indicate an even smaller clinical benefit in this group of patients [10]. Yet, a meaningful comparative analysis of the trials mentioned above is impaired by inconsistent inclusion criteria. Some trials allowed inclusion of patients who had disease progression while on different LV and 5-FU regimens, a treatment-free interval for metastatic disease of <6 months, or a relapse within 6 months after completing 5-FU-based adjuvant treatment [7, 9, 37]. Others included only fully refractory patients [6, 8, 10]. One can only speculate on the contribution of WBH to conventional chemotherapy. A hyperthermic enhancement of L-OHP may play a role. This hypothesis is supported by in vitro data demonstrating a clear relationship between response and length of exposure and concentration for L-OHP in different tumor cell lines [38]. Furthermore, several publications suggest that L-OHP dose escalation provides a significant advantage in terms of response rate and progression-free survival in patients with advanced colorectal cancer [4, 39–41]. In these trials, WHO grade 3–4 neutropenia was observed in up to 39% of patients. This was not attributed to high L-OHP dose intensity but rather to 5-FU [39]. In our trial, we did not observe a single episode of WHO grade 4 neutropenia or thrombocytopenia. Hyperthermia is known to induce a complex immune response, which is in part mediated by cytokine release, including interleukin (IL)-1β, IL-6, IL-8, IL-10 and tumor necrosis factor-α peripherally [20, 42], and at the level of the bone marrow, IL-3 and GM-CSF [21]. In our patients, we found evidence for a prolonged T-cell activation suggested by an increase of CD69 and the soluble IL-2 receptor, both known as T-cell activation markers and an increase of the number of cytotoxic T cells following WBH treatment [26]. The early activation of heat shock genes encoding the preferential synthesis of hsp is part of the ‘thermal stress response’ [43]. Members of different HSP families (e. g. hsp70, hsp90related glycoprotein gp96) most probably play a role in the stimulation of the immune response, and it has been shown that the HSP chaperon the antigenic repertoire of the cells from which they are purified [44, 45]. Furthermore, the heatinducible hsp70 is expressed on the cell surface of certain tumor cells and acts as a recognition signal for natural killer cells [22, 46]. Enhanced L-selectin-dependent adhesion of lymphocytes to vascular endothelium with the consequence of increased lymphocyte recruitment to tissues has been shown following fever-range hyperthermia [24, 25]. Whether and to what extent these factors contribute to the results of this study is not yet clear. As mentioned already, the design of our study does not allow us to answer the question of whether WBH has a beneficial therapeutic effect. Therefore a phase III study has

1203 been initiated to compare the chemotherapeutic regimen as presented in this paper with and without WBH. We will continue to apply WBH to every second treatment cycle only, since we believe that due to the need for general anesthesia a biweekly exposure to WBH would not be feasible.

Acknowledgements This study was supported by The Erich and Gertrud Roggenbuck Foundation, Hamburg, Germany.

References 1. Vanhoefer U, Harstrick A, Achterrath W et al. Irinotecan in the treatment of colorectal cancer: clinical overview. J Clin Oncol 2001; 19: 1501–1518. 2. Douillard JY, Cunningham D, Roth AD et al. Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: a multicentre randomised trial. Lancet 2000; 355: 1041–1047. 3. de Gramont A, Vignoud J, Tournigand C et al. Oxaliplatin with highdose leucovorin and 5-fluorouracil 48-hour continuous infusion in pretreated metastatic colorectal cancer. Eur J Cancer 1997; 33: 214– 219. 4. Maindrault-Goebel F, Louvet C, André T et al. Oxaliplatin added to the simplified bimonthly leucovorin and 5-fluorouracil regimen as second-line therapy for metastatic colorectal cancer (FOLFOX6). Eur J Cancer 1999; 35: 1338–1342. 5. Buechele T, Balleisen L, Beck M et al. Weekly high-dose (HD) 5-fluorouracil (5-FU) and folinic acid (FA) with addition of oxaliplatin (LOHP) after documented progression under high-dose infusional 5-FU/FA in patients (pts) with advanced colorectal cancer (CRC). Ann Hematol 1998; 77 (Suppl): 176. 6. Van Cutsem E, Szanto J, Roth A et al. Evaluation of the addition of oxaliplatin (OXA) to the same Mayo or German 5-FU regimen in advanced refractory colorectal cancer (ARCRC). Proc Am Soc Clin Oncol 1999, 18: 234a (Abstr 900). 7. Jannis J, Papakostas P, Samelis D et al. Second-line chemotherapy with weekly oxaliplatin and high-dose 5-fluorouracil with folinic acid in metastatic colorectal carcinoma: A Hellenic Cooperative Oncology Group phase II feasibility study. Ann Oncol 2000; 11: 163–167. 8. André T, Louvet C, Raymond E et al. Bimonthly high-dose leucovorin, 5-fluorouracil infusion and oxaliplatin (FOLFOX3) for metastatic colorectal cancer resistant to the same leucovorin and 5fluorouracil regimen. Ann Oncol 1998; 9: 1251–1253. 9. Brienza S, Bensmaïne MA, Souli P et al. Oxaliplatin added to 5fluorouracil-based therapy (5-FU ± FA) in the treatment of 5-FU pretreated patients with advanced colorectal carcinoma (ACRC): Results from the European compassionate-use program. Ann Oncol 1999; 10: 1311–1316. 10. Abón G, Munoz A, Rubio I et al. Third-line oxaliplatin with leucovorin (LV) and continuous infusion 5-fluorouracil (5-FU) in metastatic colorectal cancer. Ann Oncol 2000; 11 (Suppl 4): 48. 11. Wiedemann GJ, Siemens HJ, Mentzel M et al. Effects of temperature on the therapeutic efficacy and pharmacokinetics of ifosfamide. Cancer Res 1993; 53: 4268–4272.

12. Rietbroeck RC, van de Vaart PJM, Havemann J et al. Hyperthermia enhances the cytotoxicity of platinum-DNA adduct formation of lobaplatin and oxaliplatin in cultured SW 1573 cells. J Cancer Res Clin Oncol 1997; 123: 6–12. 13. Hettinga JV, Lemstra W, Meijer C et al. Hyperthermic potentiation of cisplatin toxicity in a small cell lung carcinoma cell line and a cisplatin-resistant subline. Int J Hyperthermia 1994; 10: 795–805. 14. Wilke AV, Jenkins C, Milligan AJ et al. Effect of hyperthermia on normal tissue toxicity and on adriamycin pharmacokinetics in dogs. Cancer Res 1991; 51: 1680–1683. 15. Ohno D, Siddik ZH, Baba H et al. Effect of carboplatin combined with whole body hyperthermia on normal tissue and tumor in rats. Cancer Res 1991; 51: 2994–3000. 16. Cohen JD, Robins HI, Javid M. Sensitization of C6 glioma to carboplatin toxicity by hyperthermia and thymidine. J Neurooncol 1990; 9: 1–8. 17. Wallner KE, DeGregorio MW, Li GC. Hyperthermic potentiation of cis-diamminedichloroplatinum (II) cytotoxicity in Chinese hamster ovary cells. Cancer Res 1986; 46: 6242–6245. 18. Juvekar AS, Chitnis MP. Circumvention of drug resistance of P388/R cells by the combination of adriamycin and mitoxantrone with hyperthermia (42°C). Neoplasma 1991; 38: 207–211. 19. Waterman FM, Tupchong L, Nerlinger R et al. Blood flow in human tumors during local hyperthermia. Int J Radiat Oncol Biol Phys 1991; 20: 1255–1262. 20. Robins HI, Kutz M, Wiedemann et al. Cytokine introduction by 41.8°C whole-body hyperthermia. Cancer Lett 1995; 97: 195–201. 21. Katschinski DM, Wiedemann GJ, Longo W et al. Whole body hyperthermia cytokine induction: a review, and unifying hypothesis for myeloprotection in the setting of cytotoxic therapy. Cytokine Growth Factor Rev 1999; 10: 93–97. 22. Multhoff G, Botzler C, Wiesnet M et al. A stress-inducible 72-kDa heat-shock protein (hsp72) is expressed on the surface of human tumor cells, but not on normal cells. Int J Cancer 1995; 61: 272–279. 23. Schafhausen P, Panse J, Nierhaus A et al. Whole body hyperthermia mobilizes natural killer cells (NK-cells) in the peripheral blood of cancer patients. Blood 1998; 92: 3247a. 24. Burd R, Dziedzic TS, Xu Y et al. Tumor cell apoptosis, lymphocyte recruitment and tumor vascular changes are induced by low temperature, long duration whole body hyperthermia. J Cell Physiol 1998; 177: 137–147. 25. Wang CW, Goldman LM, Schleider DM et al. Fever range hyperthermia enhances L-selectin-dependent adhesion of lymphocytes to vascular endothelium. J Immunol 1998; 160: 961–969. 26. Atanackovic D, Hegewisch-Becker S, Faltz C, Hossfeld DK. Whole body hyperthermia induces expression of the activation antigen CD69 on peripheral lymphocytes of cancer patients. Blood 2000; 96: 619a. 27. Robins HI, Cohen JD, Schmitt CL et al. Phase I clinical trial of carboplatin and 41.8°C whole-body hyperthermia in cancer patients. J Clin Oncol 1993; 11: 1787–1794. 28. Wiedemann GJ, Robins HI, Gutsche S et al. Ifosfamide, carboplatin and etoposide (ICE) combined with 41.8°C whole-body hyperthermia in patients with refractory sarcoma: A phase II study. Eur J Cancer 1996; 32A: 888–892. 29. Wiedemann GJ, Katschinski DM, Westerman AM et al. A Systemic Hyperthermia Oncology Working Group Trial: ifosfamide (IFO), carboplatin (CBDCA) and etoposide (VP-16) combined with aquatherm induced 41.8°C whole body hyperthermia (WBH) for refractory sarcoma. Proc Am Soc Clin Oncol 2000; 19: 562a (Abstr 2116).

1204 30. Robins HI, Rushing D, Kutz M et al. Phase I clinical trial of melphalan and 41.8°C whole-body hyperthermia in cancer patients. J Clin Oncol 1997; 15: 158–164. 31. Westermann AM, Grosen EA, Katschinski DM et al. A pilot study of whole body hyperthermia and carboplatin in platinum-resistant ovarian cancer. Eur J Cancer 2001; 37: 1111–1117. 32. Robins HI, Dennis WH, Neville AJ et al. A nontoxic system for 41.8°C whole-body hyperthermia: results of a phase I study using a radiant heat device. Cancer Res 1985; 45: 3937–3944. 33. Robins HI, Woods JP, Schmitt CL, Cohen JD. A new technological approach to radiant heat whole body hyperthermia. Cancer Lett 1994; 79: 137–145. 34. Miller AB, Hoogstraten B, Staquet M et al. Reporting results of cancer treatment. Cancer 1981; 47: 207–214. 35. Caussanel JP, Levi F, Brienza S et al. Phase I trial of 5-day continuous venous infusion of oxaliplatin at circadian rhythm-modulated rate compared with constant rate. J Natl Cancer Inst 1990; 82: 1046–1050. 36. Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin Trials 1989; 10: 1–10. 37. André T, Bensmaine MA, Louvet C et al. Multicenter phase II study of bimonthly high-dose leucovorin, fluorouracil infusion, and oxaliplatin for metastatic colorectal cancer resistant to the same leucovorin and fluorouracil regimen. J Clin Oncol 1999; 17: 3560–3568. 38. Raymond E, Lawrence R, Izbicka E et al. Activity of oxaliplatin against human tumor colony-forming units. Clin Cancer Res 1998; 4: 1021–1029.

39. Maindrault-Goebel F, de Gramont A, Louvet C et al. Evaluation of oxaliplatin dose intensity in bimonthly leucovorin and 48-hour 5fluorouracil continuous infusion regimens (FOLFOX) in pretreated metastatic colorectal cancer. Ann Oncol 2000; 11: 1477–1483. 40. Gamelin E, Meyer V, Delva R et al. Oxaliplatin and 5-fluorouracil synergismus in pretreated advanced colorectal patients. Proc Am Soc Clin Oncol 1997; 16: 310a (Abstr 1102.5). 41. Berteault-Cvitkovic F, Jami A, Ithzaki M et al. Biweekly intensified ambulatory chronomodulated chemotherapy with oxaliplatin, 5fluorouracil and leucovorin in patients with metastatic colorectal cancer. J Clin Oncol 1996; 14: 2950–2958. 42. d’Oleire F, Robins HI, Cohen JD et al. Cytokine induction in humans by 41.8°C whole body hyperthermia. J Natl Cancer Inst 1993; 85: 833–834. 43. Morimoto RA, Sarge KD, Abravaya K. Transcriptional regulation of heat shock genes. A paradigm for inducible genomic response. J Biol Chem 1992; 267: 21087–21990. 44. Fuller KJ, Issels RD, Slosman DO et al. Cancer and the heat shock response. Eur J Cancer 1994; 30A: 1884–1891. 45. Udono H, Srivastava PK. Heat shock protein 70-associated peptides elicit specific cancer immunity. J Exp Med 1993; 178: 1391–1396. 46. Multhoff G, Botzler C, Jennen L et al. Heat shock protein 72 on tumor cells. A recognition structure for NK cells. J Immunol 1997; 158: 4341–4350.