In vitro cell culture data and preclinical models suggest that tamoxifen modulates tumor cell sensitivity to a wide range of therapeutic agents. In the current study, the authors examined whether high-dose tamoxifen (HDT) improved the overall and complete response in patients with metastatic melanoma who were treated with concurrent biochemotherapy.
Forty-nine patients were treated with a biochemotherapy regimen of dacarbazine, vinblastine, cisplatin, decrescendo interleukin-2, interferon-α-2b, and tamoxifen. The study had a 2-step design, beginning with a tamoxifen dose escalation from 40 mg to 320 mg (17 subjects) to evaluate safety and tolerability, followed by Phase II accrual of 32 patients to HDT (320 mg) to assess clinical efficacy. Efficacy was compared with a similar modified biochemotherapy regimen with low-dose tamoxifen (LDT). Pharmacokinetic studies were performed to determine in vivo tamoxifen levels.
Tamoxifen dose escalation was completed without any reported dose-limiting toxicity. The overall response rate in the HDT group was 50% (95% confidence interval, 33.2%–66.8%), with a complete response rate of 6% and a median survival of 9.5 months. The overall response rate was not improved and the complete response and survival appeared inferior compared with that of patients recently treated with concurrent biochemotherapy and LDT. Serum tamoxifen levels were found to correlate with the dose administered, with a mean of 0.9 μM at the 40-mg dose to 4.6 μM at the 320-mg dose. Ultrafiltered protein-free sera demonstrated low (< 0.01 μM) concentrations of tamoxifen.
It is estimated that one of every 75 Americans alive today will be diagnosed with malignant melanoma during their lifetime and this incidence is increasing rapidly.1 Unfortunately, the prognosis for patients with advanced melanoma (American Joint Committee on Cancer [AJCC] Stage IV) remains poor and to our knowledge has not changed during the past 30 years, with a median survival of 6–9 months and 5-year survival rate of < 5%.2–4 These statistics reflect the rapid dissemination of melanoma by lymphatic and hematogenous routes, and the relative failure of single-agent or multiagent chemotherapeutic regimens.5–9 Biotherapy using interleukin-2 (IL-2) or interferon-α (IFNα) reportedly results in response rates of 15–20%,10–12 with durable remission reported in 3–5% of patients.13, 14
Another approach is the simultaneous use of chemotherapeutic and biotherapeutic agents: biochemotherapy. Over the last decade, various biochemotherapy regimens (multidrug chemotherapy, IL-2, and/or IFNα) have produced encouraging results in multiple single-institution Phase II trials, with response rates of 50–60%, complete remission (CR) rates of 10–20%, and a median overall survival of approximately 12 months.15–29 One single-institution randomized trial comparing sequential biochemotherapy with chemotherapy alone recently was presented and demonstrated an improved response rate and time to progression (TTP) and borderline improved overall survival for the biochemotherapy arm.30 Two randomized, multicenter Phase III clinical trials in the U.S. and Europe comparing biochemotherapy with combination chemotherapy alone (U.S.) and with or without IL-2 (Europe) are currently ongoing and to our knowledge will be presented in the next several years. However, even if these trials confirm improved response and survival with biochemotherapy, only 10% of patients will achieve durable long-term disease remissions. Therefore, the development of improved biochemotherapy regimens with greater rates of CR and durable response is critical.
Based on in vitro cell culture data and preclinical animal models, tamoxifen appears to modulate tumor cell sensitivity to a wide range of therapeutic agents including vinblastine, cisplatin, and interferon, which are all components of the modified biochemotherapy regimen.31–37 Phase II studies with low-dose tamoxifen (LDT) as part of the Dartmouth chemotherapy regimen (carmustine, dacarbazine [DTIC], cisplatin, and tamoxifen) indicated encouraging tamoxifen synergy.38–40 However, two large Phase III randomized trials failed to confirm this effect. One study evaluated the efficacy of the Dartmouth regimen with and without a higher dose of tamoxifen (160 mg)41 whereas the other study compared DTIC alone with the Dartmouth regimen with LDT (20 mg).42 Previous Phase II studies of biochemotherapy combined with LDT (20–40 mg) to date have failed to show evidence of improved response compared with standard biochemotherapy regimens.43, 44 One explanation for the negative clinical trials to date may be that the doses of tamoxifen employed in these regimens (20–160 mg) have been too low to achieve serum levels matching those required for the steep dose response curve of in vitro activity. Studies by McClay et al. suggested that, although synergy with cisplatin may result from low concentrations of tamoxifen in the < 0.1 μmol/L range, which is easily achievable at the 20-mg daily tamoxifen dose level, cisplatin-resistant melanoma cell lines require at least a 10-fold increase in tamoxifen concentration, to ≥ 1 μmol/L, to produce synergy. The attainment of such concentrations in vivo requires tamoxifen doses of at least 160 mg/day.33, 34 Similar in vitro studies suggest that even higher concentrations, ≥ 5 μmol/L, may be required to promote tamoxifen-induced reversal of multidrug resistance.35, 36
Therefore, in an effort to maximize the potential synergy of tamoxifen with biochemotherapy and improve the efficacy of our previously reported concurrent biochemotherapy regimen,44 we escalated the dose of tamoxifen from 20 mg to 320 mg. The 320-mg target dose was chosen based on published data indicating that this dose level would be required to achieve serum concentrations in vivo corresponding to in vitro chemosensitization.45–47
MATERIALS AND METHODS
Eligibility for study was based on the following criteria: a histologically confirmed diagnosis of malignant melanoma with measurable AJCC Stage IV disease, age of at least 18 years but < 65 years, Eastern Cooperative Oncology Group (ECOG) performance status of 0–2, acceptable end-organ function (total bilirubin < 2.0 mg/dL, serum creatinine < 1.6 mg/dL, a leukocyte count > 3000/μL, and a platelet count > 100,000/μL), and adequate cardiac function and pulmonary reserve. Patients with ischemic heart disease, uncontrolled hypertension, congestive heart failure, arrhythmia requiring therapy, or a left ventricular ejection fraction (LVEF) of < 40% were ineligible, as were those patients with a recent (< 3 months) history of deep vein thrombosis (DVT) or pulmonary embolus. Patients with a smoking history or symptomatic pulmonary disease were eligible only if pulmonary function tests (1-second forced expiratory volume [FEV1], and ratio of 1-second forced expiratory volume to vital capacity [FEV1/FVC]) were > 75% of the predicted value. Patients receiving previous biochemotherapy or previous chemotherapy plus independent biologic therapy were not eligible. All prior therapies had to have been completed at least 3 weeks before the initiation of biochemotherapy. Patients with < 5 untreated, asymptomatic brain metastases were eligible if all lesions measured < 10 mm in greatest dimension. Patients with larger brain metastases were required to have undergone definitive therapy with surgery and/or stereotactic radiation, followed by a 4-week period without the appearance of new lesions. The concomitant use of corticosteroids for any reason was prohibited. The participating institutional review board approved the study protocol and written informed consent was obtained from all patients.
The pretreatment evaluation, which was to be completed ≤ 2 weeks before the initiation of biochemotherapy, included a complete history and physical examination and routine laboratory tests, including complete blood cell count with differential and a 12-channel biochemistry panel. A negative serum pregnancy test was required for women of childbearing potential. Baseline staging of metastatic disease was assessed using computed tomography scans of the chest, abdomen, and pelvis, and magnetic resonance imaging of the brain. Each patient's cardiac function was assessed by an electrocardiography and either two-dimensional echocardiogram or nuclear cardiac scanning. Further cardiac and pulmonary testing was performed on selected patients. A subclavian, dual-lumen in-dwelling catheter was inserted in all patients before the initiation of biochemotherapy.
Patients were admitted to the inpatient oncology unit every 21 days for a 5-day regimen of DTIC, vinblastine, cisplatin, IL-2, IFN-α-2b, and tamoxifen, administered as shown in Table 1. DTIC was administered over a 1-hour period. Each dose of cisplatin was preceded by intravenous (IV) hydration with 1 L of normal saline delivered over 1 hour. Vinblastine was delivered during cisplatin prehydration. IL-2 was administered in decrescendo fashion for a total dose of 36 million IU/m2.
Table 1. 21-Day Regimen of Concurrent Biochemotherapy
Patients were assigned to increasing daily tamoxifen doses (40 mg, 80 mg, 160 mg, 240 mg, and 320 mg) in cohorts of 3 patients. High-dose tamoxifen (HDT) administration began 5 days before and continued each day during the 5-day inpatient biochemotherapy cycle. This daily dose decreased to a maintenance dose of 20 mg on Days 6–16. The administration of granulocyte–colony-stimulating factor(G-CSF) began on Day 6 of each cycle and was given subcutaneously (usually self-administered by the patients at home) each day for 7–10 days at a dose of 5 μg/kg. G-CSF administration was discontinued when the patient's absolute neutrophil count was > 10,000/mm3.
Dose Escalation Schema
The dose escalation schema is summarized as follows. At least three patients were treated in each dose cohort. If no dose-limiting toxicity (DLT) was observed in any patient, a fourth patient received the next higher dose of tamoxifen. However, if one of three patients in a dose level developed a DLT then another patient was added to the same dose level. If a DLT was observed in this patient (for a total of two of four treated at this dose), then this dose was considered the DLT. If no additional DLT occurred (for a total of one of four treated), the next patient was treated with the next dose level. At least 3 patients must have completed at least 3 weeks of treatment at a given dose level before the next dose level was started. The dose of tamoxifen was increased sequentially as follows: 40 mg (n = 4), 80 mg (n = 4), 160 mg (n = 3), 240 mg (n = 3), and 320 mg (n = 3).
Monitoring and Symptom Management
During each 5-day inpatient treatment, patients were monitored closely and any symptoms of toxicity were managed as described previously.44 Patients were discharged on Day 5 if ambulatory and tolerating oral fluids with antiemetics. After discharge, patients were monitored as described previously.44
Response Criteria and Data Analysis
Response was evaluated for patients who completed at least two cycles of treatment. CR was defined as the disappearance of all clinical evidence of tumor by radiographic studies and physical examination, followed by at minimum a 4-week period without the appearance of new lesions. Partial response (PR) was defined as a > 50% reduction in the sum of the products of the perpendicular dimensions of measurable lesions, without the appearance of new lesions for a minimum of 4 weeks. Stable disease (SD) was defined as either a < 25% increase in the sum of the products of the perpendicular dimensions of all measurable lesions or a 0–49% decrease in the sum of the products of the perpendicular dimensions of all measurable lesions without the appearance of new lesions for a minimum of 8 weeks. Progressive disease (PD) was a ≥ 25% increase in the sum of the products of the perpendicular dimensions of any measurable lesion or the appearance of any new lesion. Development of central nervous system (CNS) metastases was considered to be PD, regardless of the response noted at other disease sites.
The duration of response was measured from Day 1 of treatment to the date of PD, last follow-up date, or death from any cause. Survival was measured from Day 1 of treatment to death or last follow-up.
Determination of Serum Tamoxifen Concentration
Blood samples were obtained from patients on Days 1 and 5 of Cycles 1 and 2. Each sample was allowed to clot and then centrifuged at 4 °C for 5 minutes at 2000 revolutions per minute (rpm). Serum then was transferred into 15-mL, screw-capped polypropylene tubes, centrifuged another 5 minutes under the same conditions, and transferred in 1-mL aliquots into 2-mL O-ringed screwcapped polypropylene tubes. The samples were stored frozen at –70 °C.
To determine the fraction of protein-bound versus protein-free tamoxifen, serum was processed within 1 hour of the blood draw. After the second centrifugation, the serum was transferred in 1-mL quantities into each of 4 ultrafilters (Millipore Centrifree Filter #4104; Millipore Corporation, Bedford, MA), which have a 10,000 molecular weight cutoff. The centrifugation was performed at 4 °C and 2000 rpm for 20 minutes. Approximately 300 μL of filtrate was obtained from each 1 mL of serum filtered. The filtered aliquots were frozen at −70 °C prior to shipping.
Serum concentrations of tamoxifen and its primary human metabolite, N-desmethyltamoxifen, were determined using high-performance liquid chromatography (HPLC) as described previously.48 Briefly, serum samples were thawed and transferred into individual 16 mm × 125 mm borosilicate glass screwcap extraction tubes (Corning, Corning, NY). Volumes were recorded. Each specimen was spiked with an internal standard (200 ng nafoxidine hydrochloride; Sigma Chemical Company, St. Louis, MO) and briefly vortex mixed. All samples were extracted with 9.0 mL of 2% butanol in hexane (volume/volume), vortexed for 1 minute, and centrifuged for 10 minutes at 1000 g. The organic layer then was transferred to a clean 16 mm × 125 mm glass tube and evaporated to dryness at 37 °C under a gentle stream of nitrogen. Extracted samples were frozen at −20 °C until the day of analysis.
Plasma calibration standards were prepared by spiking blank human plasma with standard stock solutions containing 0.067–4.3 μM of tamoxifen and N-desmethyltamoxifen. Six-point standard calibration curves were produced by performing linear regression analysis of the peak height ratios of the test compound to internal standard. All correlation coefficients were > 0.99.
On the day of analysis, all samples and standards were reconstituted in 200 μL of methanol, transferred to an Infrasil quartz cuvette (Fisher Scientific, Pittsburgh, PA), and irradiated for 1 minute with high-intensity ultraviolet light (254 nanometers). The activated samples were injected onto a Whatman (Clifton, NJ) C18 Partisphere 5-μm, 4.6 mm ×, 125 mm reverse-phase HPLC column and eluted isocratically with a mobile phase comprised of 928 mL of methanol, 70 mL of water, and 2 mL of triethylamine per liter. Fluorescence of activated compounds was detected using an Applied Biosystems Model 980 programmable fluorescence detector (Applied Biosystems, Foster City, CA). Peak heights and retention times were recorded.
Comparative analyses of response and toxicity were performed using a historic control group comprised of patients from our previously reported trial of concurrent biochemotherapy with low-dose (20 mg) tamoxifen (LDT).44 The regimen of biochemotherapy administered to historic controls was identical to the regimen reported in the current study, with the exception of the daily tamoxifen dose. Both trials utilized postbiochemotherapy growth factor support with G-CSF.
Response and survival analysis was performed for patients who received treatment with HDT. All patients enrolled at the 320-mg tamoxifen dose level were followed for survival. Only those patients who received at least 2 cycles of therapy at the 320-mg tamoxifen dose level were assessed for response. Overall survival and TTP were estimated using the Kaplan–Meier method. Survival was measured from day one of treatment to date of death or last follow-up. TTP was measured from Day 1 of treatment to date of first evidence of PD. Patients who underwent complete surgical resection of residual disease after biochemotherapy were censored at the date of surgery.
Comparison of response rates and survival in the HDT study group with those of a historic control group comprised of a previously published cohort of patients receiving biochemotherapy with LDT (20 mg)44 also was performed. To reduce assignment and other biases in observed prognostic factors among the two cohorts, a propensity score subclassification method was used.49 A propensity score is the conditional probability of being assigned either the 320-mg or 20-mg tamoxifen dose given observed prognostic factors including age, gender, ECOG performance status, number of disease sites, organ-specific sites of metastasis, prior chemotherapy, prior therapy with allogeneic vaccine, and baseline serum lactate dehydrogenase (LDH) level. For the current analysis, the propensity score was estimated for each subject using a logit model for the treatment group. All the potential prognostic factors mentioned earlier were included in the model. The 80 subjects were classified into 5 strata based on the estimated propensity scores. Within the strata, the prognostic factors were balanced between the two treatment groups.
The association between observed prognostic factors and overall survival was analyzed using a Cox regression model. A stratified Cox regression model,50 with propensity score strata as the stratification factor, was used to compare the hazards of death among the two dose groups. The chi-square test, or Fisher exact test when necessary, was used to examine the relation between the tumor response rate and each prognostic factor within each treatment group. To compare the tumor response rates of the two treatment groups, the odds ratios of tumor response to the two treatments was estimated within each propensity score strata, and an exact test developed by Zelen51 was used to examine the homogeneity of the five odds ratios. If there was no significant difference among the five odds ratios, a common odds ratio was estimated using a Mantel–Haenzsel inference, and whether the common odds ratio was equal to one (e.g., the tumor response odds of the two treatments were equal) was tested.52
Forty-nine patients were enrolled in the study between May 28, 1996 and August 11, 1998 (Table 2), including 34 males (69%) and 15 females (31%). The median age for all patients was 45 years (range, 17–65 years). Nearly all patients (98%) had visceral sites of metastatic disease and the majority (86%) had ≥ 2 (organ) sites of disease. Eighteen patients (37%%) entered the study with an ECOG performance status of 2, and 23 patients (47%) had an elevated LDH value. Of the six patients (12%) who entered the study with a history of CNS metastases, 1 patient had undergone surgical resection followed by whole brain radiation, 2 patients had received stereotactic radiation therapy, and 3 patients were asymptomatic and had not received prior CNS therapy. No study patient received prior therapy with high-dose IFNα or IL-2. The median time from diagnosis of AJCC Stage IV melanoma to the initiation of biochemotherapy was 1.2 months (range, 0.1–24 months). At the time of last follow-up, 8 patients (16%) were alive (range, 18.7–42.8 months); 4 (8%) were alive with disease, and 4 (8%) were clinically free of disease.
The dose of tamoxifen was escalated successfully from 40 mg to 320 mg without any reported DLTs. Seventeen patients participated in the dose escalation study and comprised 3 groups of 3 patients each and 2 groups of 4 patients each. One patient was added at the 40-mg dose level in response to a treatment-related death in this group. The deceased patient failed to complete one cycle of therapy and died on Day 30 of Cycle 1 due to complications of acute renal failure and adult respiratory distress syndrome, believed to be related to IL-2 and cisplatin, not tamoxifen. A fourth patient also was added at the 80-mg dose level due to an acute myocardial infarction experienced by a patient previously treated at this dose level. This event occurred on Day 2 of Cycle 1 and was attributed to the hemodynamic stress of IL-2-based therapy in combination with the patient's preexisting cardiac risk factors; however, tamoxifen may have contributed by inducing a hypercoagulable state. This patient was discontinued from the study and recovered fully.
Response, TTP, and Survival for HDT Patients
Thirty-five patients (including 3 patients from the dose escalation study phase) received biochemotherapy with the 320-mg daily dose of tamoxifen. A total of 162 cycles were administered, with a mean of 4.6 cycles per patient and a range of 1–8 cycles. Thirty-four patients completed at least 2 cycles of therapy and were evaluable for response: 2 patients (6%) achieved a CR whereas 15 patients (44%) achieved a PR, for a combined overall response rate of 50% (95% confidence interval [95% CI], 33.2%–66.8%). Eight patients (24%) had SD whereas 9 patients (26%) experienced PD. The median TTP was 5.3 months and the median survival was 9.5 months (Fig. 1).
Predictors of Response and Survival
Pretreatment factors were evaluated as predictors of response and survival for the 35 patients receiving HDT. These factors included: gender, age (age < 50 years vs. age ≥ 50 years), ECOG performance status (0,1 vs. 2), number of disease sites (1–2 vs. ≤ 3), metastatic disease sites (soft tissue/lymph node/lung vs. other visceral), prior chemotherapy, prior allogeneic vaccine, and LDH (≤ 190 IU/L vs. > 190 IU/L). The association between each prestudy factor and survival was assessed using a Cox regression model. None of the above-mentioned factors were found to be associated significantly with response (CR + PR). The only factor found to be associated significantly with prolonged survival was an ECOG performance status of 0 or 1 versus an ECOG performance status of 2 (P = 0.0212).
A stratified logistic regression model using propensity score strata produced from study subjects and historic controls was used to determine any tamoxifen dose effect (20 mg vs. 320 mg) on overall response (CR + PR) and CR. Study patients and controls shared important prognostic and clinical features (Table 3), with the exception of significantly higher chemotherapy pretreatment among controls. However, none of the aforementioned factors was identified as being a significant predictor of overall response. The estimated common odds ratio of 0.8154 (95% CI, 0.33 –2.00; P = 0.6560) suggests that overall response rates among the HDT and LDT groups were very similar. Only two patients in the HDT group (n = 34) achieved a CR to biochemotherapy, compared with 10 of 44 patients treated with the identical biochemotherapy regimen with LDT (20 mg) (Table 4). Although the LDT group had a higher CR rate than the HDT group, this difference approached, but did not achieve, statistical significance. The estimated common odds ratio for attaining a CR in the HDT group was 0.2065 (95% CI, 0.038–1.107) with an exact P value of 0.0588.
Table 3. Study Patients Versus Historic Controls: Patient Characteristics
The historic controls were derived from: O'Day SJ, Gammon G, Boasberg PD, Martin MA, Kristedja TS, Guo M, et al. Advantages of concurrent biochemotherapy modified by decrescendo interleukin-2, granulocyte colony-stimulating factor, and tamoxifen for patients with metastatic melanoma. J Clin Oncol 1999;17(9):2752–61.
Age ≥ 50 yrs
ECOG performance status
No. of metastatic (organ) sites
Sites of metastasis
Lung ± ST/LN only
Elevated LDH (> 190 IU/L)
Table 4. Response Evaluation: Study Patients (High-Dose Tamoxifen) Versus Historic Controls (Low-Dose Tamoxifen)
Historic control groupa (low-dose tamoxifen) (n = 44)b
The historic controls were derived from: O'Day SJ, Gammon G, Boasberg PD, Martin MA, Kristedja TS, Guo M, et al. Advantages of concurrent biochemotherapy modified by decrescendo interleukin-2, granulocyte colony-stimulating factor, and tamoxifen for patients with metastatic melanoma. J Clin Oncol 1999;17(9):2752–61.
Number evaluable for response.
Overall response rate
As described earlier, two patients in the dose-escalation phase of the study developed toxicity either resulting in death or requiring discontinuation of study participation. One patient died on treatment due to acute renal failure and adult respiratory distress syndrome, and another suffered an acute myocardial infarction. Neither event was believed to be related directly to tamoxifen. After tamoxifen dose escalation, 35 patients were treated at the 320-mg dose level, receiving a total of 162 treatment cycles. Toxicity was graded according to National Cancer Institute common toxicity criteria (Table 5). All patients experienced constitutional or flu-like symptoms during treatment, including chills and rigors, fever, fatigue, myalgia, and headache. No new patterns of constitutional toxicity were observed in patients receiving HDT compared with patients who received biochemotherapy with LDT (20 mg).44
The hematologic toxicity of biochemotherapy with HDT was similar to that of biochemotherapy with LDT.44 Thirty-nine percent of cycles were complicated by Grade 3 or 4 neutropenia with HDT compared with 44% for LDT. The incidence of fever and neutropenia was identical in HDT and LDT (2% of cycles in both). The median absolute neutrophil count nadir was 1328/mm3 and 1392/mm3 for Cycles 1 and 2, respectively, in HDT patients, compared with 962/mm3 and 1214/mm3 in LDT patients. No routine antibiotic therapy was given. Anemia and thrombocytopenia were common and cumulative with repeated cycles of therapy. The degree of anemia was similar in the HDT and LDT groups, whereas thrombocytopenia occurred with slightly higher frequency in the HDT group. Twenty-four HDT patients (69%) required packed red blood cell transfusions through the course of treatment. This compares with the historic figure of 60% in LDT patients. Grade 3 or 4 thrombocytopenia complicated 35% of cycles, compared with 21% for LDT. Eight patients (23%) in the HDT group required a platelet transfusion, with a median of 2 units per transfusion and a range of 1–12 units compared with 40%, of patients in the LDT group requiring a median of 1 unit per transfusion and a range of 1–6 units.
Three patients (9%) at the 320-mg tamoxifen dose level required dopamine infusion for blood pressure support. Two patients (6%) required monitoring in the intensive care unit. Systolic hypotension of < 90 mm Hg was recorded in 57% of patients (20 of 35 patients) during 34% of cycles of biochemotherapy with HDT, compared with 44% of patients during 20% of cycles with LDT. Approximately 18% (n = 6) of patients receiving HDT gained > 10% of body weight, compared with 4% of patients receiving LDT.
Only 12 HDT treatment cycles (7%) were complicated by infection, compared with 13% in the LDT group. The majority of infections were catheter-related; all central line infections were treated successfully with intravenous antibiotics. Removal of the central line catheter was performed in five cases. Clostridium difficile infection occurred in 1 patient (3%) and was treated successfully. There were no reported deaths related to infection.
Grade 3 or 4 gastrointestinal toxicity and renal toxicity was rare and was similar in the HDT and LDT groups: 13% of high-dose cycles versus 12% of low-dose cycles (gastrointestinal toxicity), and 3% in the HDT group versus 2% in the LDT group (renal toxicity). Catheter-related DVT was higher in the HDT patients (9%) compared with the LDT group (2%). In all, 6 patients (17%) were removed from the HDT arm because of related toxicities (3 patients because of catheter-related DVT and 3 patients because of prolonged nausea).
Two patients in the 320-mg tamoxifen dose group died while on study. The first patient had a high volume of metastatic disease and developed tumor lysis syndrome during the first week of therapy. The patient died of a cardiac arrhythmia secondary to metabolic abnormalities. A second patient was clinically stable between Cycles 5 and 6 of treatment and died suddenly of CNS bleeding. The patient had a history of hypertension, but not thrombocytopenia. The patient had no history of melanoma metastases to the brain and CNS; autopsy revealed no CNS melanoma.
In vivo serum levels of tamoxifen and its major metabolite, N-desmethyl-tamoxifen, were assayed by HPLC for all patients receiving daily doses of 40 mg, 80 mg, 160 mg, and 240 mg of tamoxifen, and for the first 16 patients enrolled at the 320-mg tamoxifen level. The results (Fig. 2) show that there was a dose-related increase in serum concentrations with mean tamoxifen and N-desmethyl-tamoxifen levels of 1.8 μM and 2.8 μM, respectively, for the patient group receiving 320 mg of tamoxifen. The concentration of tamoxifen did not appear to alter significantly between Days 1–5 of each biochemotherapy cycle due to the 5-day loading dose prior to each cycle. Fourteen additional samples at the daily dose level of 320 mg were analyzed for tamoxifen protein binding. Results from this analysis (Table 6) demonstrate that both tamoxifen and N-desmethyl-tamoxifen are highly protein-bound with mean levels of free, nonprotein-bound tamoxifen and N-desmethyl-tamoxifen equaling < 0.01 μM.
Table 6. Protein-Depleted Serum Tamoxifen Levels
Tamoxifen dose level of 320 mg
No. of samples
Mean total tamoxifen + metabolite (μM)
Serum tamoxifen levels
Protein-free ultrafiltrate tamoxifen levels
The objectives of the current study were to determine whether tamoxifen could be escalated safely to a high dose sufficient to achieve serum concentrations corresponding to the levels required in vitro for drug synergy, and to examine the HDT group for evidence of improved efficacy. Study data showed that dose escalation of tamoxifen from 40 mg to 320 mg was well tolerated in combination with our modified concurrent biochemotherapy regimen. Toxicity of the HDT biochemotherapy regimen generally was comparable to the LDT regimen. Capillary leak syndrome (hypotension and fluid retention) and thrombocytopenia were somewhat worse with HDT, but this difference appeared to have minimal clinical impact. There was a higher incidence of catheter-related DVT in the HDT group (9% vs. 2%), which was likely related to the hypercoagulable properties of tamoxifen. There were no clinically apparent pulmonary emboli.
The dose and schedule of administration for tamoxifen was selected based on published pharmacokinetic data.45–47 This regimen resulted in a relatively constant serum tamoxifen level during the 5-day treatment period of approximately 1.5–2.0 μM and slightly higher serum levels of N-desmethyl-tamoxifen (2.0 –2.5 μM). N-desmethyl-tamoxifen, the major tamoxifen metabolite in humans, has been shown to have activity similar to tamoxifen in reversing drug resistance.46, 53 The synergistic actions of tamoxifen with cytotoxic and biologic agents require concentrations in the low micromolar range in some in vitro systems (0.1–2.0 μm), although the optimal dose for synergy varies between different experimental systems and in some cases requires higher concentrations (5–10 μm).31–37 Evidence supports several mechanisms for synergy including regulation of apoptotic mediators,35, 36 multidrug resistance pump inhibition,32 and a unique mechanism for cisplatin synergy.34
Although serum tamoxifen levels reached the targeted range in the current study, there was no improvement of the response rate in the HDT group compared with the identical biochemotherapy regimen in the LDT group (P = 0.462). More disappointing was the CR rate and median survival observed in the HDT group. Only 2 patients of 35 evaluable for response (6%) in the 320-mg tamoxifen group achieved a CR, compared with 10 of 44 patients (23%) at the 20-mg dose level.44 The median survival for the HDT group was 9.5 months compared with 13.4 months for the LDT control group.44 These differences were not statistically significant. However, in the previous (control group) study with LDT there were two independent predictors of a CR to treatment in multivariate analysis: 1) no prior chemotherapy and 2) metastatic disease sites (soft tissue/lymph node/lung vs. other visceral). In the current study, the HDT biochemotherapy group had a comparable distribution of sites of metastases but no patients receiving chemotherapy prior to biochemotherapy (compared with 33% in the historic control group) and thus was a more favorable group and would have been predicted to have a higher response rate. This provides further indication that HDT provided no added benefit to biochemotherapy, and may have been detrimental in producing a CRs.
The failure to observe any beneficial therapeutic effect for HDT may be due to inadequate serum levels. The concentration required for optimal synergistic activity with cisplatin, vinblastine, and IFN-α in in vitro studies varies from 1.0 μM to 10 μM depending on the cell line and other experimental conditions, and the mean tamoxifen concentrations achieved in the current study were only in the lower range of the dose response curve. In addition, tamoxifen has been shown to be highly protein bound with > 98% associated with human serum albumin.54 These results were confirmed by the data in the current study comparing total serum and serum ultrafiltrate levels. The disparity between in vitro cytotoxic synergy and the failure to demonstrate clinical evidence of tumor sensitization may be due to the lower levels of protein present in cell culture leading to less protein binding.55 Thus, the low levels of free tamoxifen in vivo may be inadequate for cytotoxic sensitization. Finally, tamoxifen in the low micromolar range has been reported to have multiple biochemical actions on cell signaling pathways independent of estrogen receptor-related effects and tumor cell sensitization, and one or more of these actions may be responsible for the apparent decreased efficacy.56–58
Although the overall response rate is very similar in the HDT group compared with the control population, the lower CR rate is more important clinically because only these patients have prolonged disease remissions whereas patients with a PR show uniform disease progression within 6–9 months. Overall, the results from the current study support the need for newer, more active pharmacologic agents other than tamoxifen that specifically target distinct drug resistance mechanisms and that should be tested in conjunction with biochemotherapy to improve clinical response and outcome.
The authors thank the entire nursing and support staff of 3-South at Saint John's Health Center for their dedicated care of the patients, with particular recognition to Shirley Edwards, Annette Sy, Phillip Williams, Tess Duenas, and Donni Esposti for their leadership. The authors also thank Ruth Weil and Jackie Kathe for their ongoing support of research in the Division of Medical Oncology at the John Wayne Cancer Institute, Santa Monica, California.