Docetaxel Followed by Gemcitabine and Irinotecan in Solid Tumors

ABSTRACT: Docetaxel (Taxotere), gemcitabine (Gemzar), and irinotecan(Camptosar, CPT-11) are active single agents in a variety of solid tumors. Incombination, synergism may be schedule dependent. Preclinical studies suggestedsynergistic interactions when docetaxel was administered 24 hours beforegemcitabine or irinotecan. The objective of this phase I trial in patients withrefractory solid tumors was to determine the maximum tolerated dose of docetaxelfollowed 24 hours later by gemcitabine and irinotecan. Two different scheduleswere tested: docetaxel escalated by5 mg/m2/cohort from an initial dose of 20 mg/m2 on days 1 and 8 (schedule A) orescalated by 15 mg/m2/cohort from 45 mg/m2 on day 8 only (schedule B). In bothschedules, docetaxel was given over 1 hour. Gemcitabine and irinotecan weregiven on days 2 and 9 (arm A) or 1 and 9 (arm B) at fixed doses of 1,000 mg/m2over 30 minutes and 100 mg/m2 over 90 minutes, respectively. Escalation ofdocetaxel was planned in groups of three patients, with three additionalpatients added at the first indication of dose-limiting toxicity. Four doselevels in arm A and one dose level in arm B have been tested. Seventeen patientswere evaluable in arm A; one died of an unrelated cause on cycle 1, and anotherwithdrew consent before beginning treatment. Five of six patients were evaluablein arm B; one patient inadvertently received G-CSF on cycle 1. Forty-two cycleshave been delivered in arm A (mean; 2.2 cycles/patient), and 25 cycles in arm B(mean, 4.2 cycles/patient); the maximum tolerated dose of docetaxel on arm A was20 mg/m2. The dose-limiting toxicities were grade 3 diarrhea in one patient,grade 3 infection in two patients, and grade 4 neutropenia for > 4 days inone patient at the 25 mg/m2 level. The dose-limiting toxicities on arm Boccurred at the first dose level and included grade 3 diarrhea in one patient,grade 4 diarrhea in one patient, and grade 4 neutropenia for 4 days in anotherpatient. Accrual to schedule B was closed after testing the cohort 1 dose levelbecause testing of a single deescalated docetaxel dose given on day 8 was notconsidered clinically relevant. [ONCOLOGY 15(Suppl 1):37-45, 2001]

Introduction

Irinotecan (Camptosar, CPT-11),gemcitabine (Gemzar), and docetaxel (Taxotere) are chemotherapeutic agents with different intracellulartargets and complementary, rather than overlapping, mechanisms of action.

The taxanes induce polymerization of tubulin’s alpha and betasubunits, resulting in stabilization of microtubules and disruption of the cellcycle.[1,2] Difluorodeoxycytidine triphosphate (dFdCTP), the predominantintracellular metabolite of gemcitabine, is incorporated as a substrate duringDNA synthesis, causing inhibition of DNA elongation and chain termination afterthe addition of another base or another molecule of dFdCTP. In addition, thediphosphate form of gemcitabine leads to a decrease in the normal intracellulartriphosphate pools.[3,4] The active metabolite of irinotecan SN-38 stabilizesthe covalent linkage between the topoisomerase I enzyme and the DNA backboneformed during the enzymatic relaxation of DNA torsional strain. This slows DNAreligation during the catalytic cycle.[5]

All three agents—irinotecan, gemcitabine, and docetaxel—areindividually active in a variety of malignancies as first- and second-linetreatment. In preclinical studies, interactions between these agents have beendemonstrated, with the magnitude of beneficial interaction at least somewhatschedule related. A marked synergistic effect was observed when CAEP cells, acell line derived from squamous cell carcinoma of the lung, were exposed todocetaxel followed 24 hours later by gemcitabine; less synergy was observed withthe reverse sequence.[6] The two sequences also produced moderate synergistickilling of RAL cells, a cell line derived from adenocarcinoma of the lung. Inthis cell line, enhanced synergism was encountered when a 48-hour washout wasused between docetaxel and gemcitabine exposures.[6] A similar sequence,docetaxel given first followed 24 hours later by irinotecan, resulted insynergistic interactions between these two drugs in cancer cell lines.[7,8]

Simultaneous exposure to gemcitabine and irinotecan resulted inantagonism at low concentrations but synergism at concentrations of gemcitabineabove 0.1 mM and irinotecan above 3.2 µM in the SCOG small-cell lung cancercell line. In MCF7 breast cancer cells, synergism occurred at gemcitabineconcentrations of 0.1 to 2 mM and irinotecan concentrations of 0.2 to 10 µM.However, antagonism occurred at high concentrations (ie, > 2 mM ofgemcitabine and > 20 mM of irinotecan).[9]

Phase I data from our institution showed that gemcitabineadministered at 1,000 mg/m² over 30 minutes followed by irinotecan at 100mg/m²over 90 minutes (IrinoGem) could be administered on a day 1 and 8 schedule every3 weeks.[10] This IrinoGem regimen has recently been studied in a phase II trialfor chemotherapy-naive advanced and metastatic pancreatic cancer patients. Inthat trial, almost 90% full doses of both drugs were delivered on days 1 and 8.The regimen has been well tolerated and active with modest toxicity.[11]

To build on the clinical activity of IrinoGem and take advantageof the available preclinical synergy data, we conducted a phase I trial of thecombination of docetaxel, gemcitabine, and irinotecan (the DIG regimen) inpatients with solid tumors. The schedule and doses chosen were based on thepreclinical synergistic interactions and our prior experience delivering almostfull doses of gemcitabine and irinotecan in combination. In this trial, thedoses of gemcitabine and irinotecan were initially fixed at 1,000 mg/m² and 100mg/m², respectively. Docetaxel doses were to be escalated until the maximumtolerated dose for the combination was defined.

Two different schedules were studied: docetaxel was eitheradministered on days 1 and 8, followed 24 hours later by gemcitabine andirinotecan on days 2 and 9 (schedule A), or docetaxel was administered on day 8,with gemcitabine and irinotecan on days 1 and 9 (schedule B). As was done in ourprior studies with the gemcitabine/irinotecan combination, irinotecan was givenimmediately following gemcitabine.

The objectives of the study were to determine the maximumtolerated dose of docetaxel that could be administered with fixed doses ofgemcitabine and irinotecan, to describe the pattern of dose-limiting toxicity,and to define the recommended phase II doses for each of the tested schedules ofthis three-drug regimen.

Patients and Methods

Patient Selection

Adult patients with pathologically confirmed solid tumors thatwere refractory to standard therapy, or for whom no standard therapy of provenefficacy was available, were eligible if they had adequate organ function,performance status of 0 to 2, and resolution of toxic effects from priortherapy. Adequate organ function was defined as granulocyte count of at least1,500/mL, platelet count of at least 100,000/mL, serum creatinine < 2.1mg/dL, and serum bilirubin < 2.1mg/dL. Female patients with child-bearing potential must have had a negativepregnancy test.

Patients were ineligible if they had known bone marrowmetastases, a history of congestive heart failure requiring medical therapy (NewYork Heart Association class III or IV), unstable angina, atrial fibrillation,or myocardial infarction within the 6 months prior to study entry, uncontrolledbacterial, viral, or invasive fungal infection, prior whole pelvic radiation, ora psychiatric condition that would prevent informed consent. Prior chemotherapywith any or all three of the agents under study was allowed. Measurable orevaluable disease was not required. All patients gave written informed consentby signing an informed consent document approved by the Institutional ReviewBoard of the Medical University of South Carolina.

Treatment Plan

Docetaxel was administered as a 60-minute (later amended to a30-minute) intravenous infusion on days 1 and 8 (arm A) or day 8 only (arm B).This was followed 24 hours later, on either days 2 and 9 (arm A) or day 9 (armB) of each 3-week treatment cycle by a 30-minute intravenous infusion ofgemcitabine at 1,000 mg/m² immediately followed by irinotecan at 100mg/m² over90 minutes. On arm B, the day 1 doses of gemcitabine and irinotecan were givenwithout docetaxel pretreatment. The dose levels tested in this phase I trial areshown in Table 1. Cohorts of at least three patients were evaluated at each doselevel. No dose escalation was permitted within individual patients.

Patients were removed from protocol if they demonstratedprogressive disease or allergic reaction with diffuse rash or anaphylaxis, or iftreatment termination was deemed in the best medical interests of the patient.Patients experiencing a dose-limiting toxicity, but also demonstrating clinicalor radiographic response or stable disease with subjective benefit, couldcontinue treatment at the next lower dose level.

Drug Administration

All patients received prophylactic antiemetic therapy chosen bytheir treating physician. Antiemetics generally included an HT3 blocker not onlybefore docetaxel, but also before gemcitabine and irinotecan. Oral dexamethasone at 8 mg was taken the night before and bid on the same day as docetaxelfor a total of three doses (dexamethasone at 20 mg IV was given 30 minutesbefore the infusion of docetaxel if the patient had not taken the dexamethasonethe night before).

Gemcitabine at 1,000 mg/m² was prepared in 250 mL of normalsaline and given over 30 minutes by IV infusion on days 2 and 9 (arm A) or days1 and 9 (arm B) of each treatment cycle. Immediately following completion ofeach dose of gemcitabine, irinotecan at 100 mg/m², prepared in 500 mL of normalsaline or 5% dextrose in water (D5W) solution, was administered over 90 minutesby IV infusion. Docetaxel at the cohort-specific dose was prepared in 100 mL ofnormal saline or D5W, and given over 60 minutes (later amended to 30 minutes) byIV infusion on days 1 and 8 (arm A) of each treatment cycle, or only on day 8 ifthe arm B dose escalation was being employed. The treatment cycles were repeatedevery 3 weeks.

Prophylactic atropine was not routinely given. In patients whodeveloped cholinergic diarrhea during the infusion of irinotecan, 0.5 mg ofatropine was administered IV. These and any other patients who reported diarrheawithin the first 8 hours after completing irinotecan were given prophylacticatropine just before all subsequent doses of irinotecan. Patients wereinstructed to begin taking oral loperamide (Imodium) at the earliest signs ofdiarrhea and/or abdominal cramping that occurred more than 8 hours afterreceiving irinotecan. Loperamide was prescribed as 4 mg orally at the onset ofcramps and/or diarrhea, and then 2 mg every 2 hours around the clock until thepatient was diarrhea-free for at least 12 hours. During the night, patients wereadvised to take 4 mg of loperamide every 4 hours instead of 2 mg every 2 hours.

Growth factors were not permitted during the first cycle oftherapy, unless profound neutropenia and severe or life-threatening infectionwere present. Patients received full supportive care (ie, transfusions of bloodproducts, antibiotics, antidiarrheals, analgesics, etc, as appropriate).

Dose Escalation and Maximum Tolerated Dose

For this phase I trial, the maximum tolerated dose was definedas the dose level immediately below the dose level at which two out of the firstthree patients in any cohort, or at least two out of six patients in anyexpanded cohort, experienced a dose-limiting toxicity. The followingdose-escalation rules were used: Three patients were initially studied at thefirst dose level. If none of these patients experienced a dose-limiting toxicityduring the first cycle of therapy, the dose was escalated to the next higherlevel for the next three patients. If one of three patients at any dose levelexperienced a dose-limiting toxicity, three additional subjects were accrued atthat dose. If none of these three additional patients experienced adose-limiting toxicity, the dose was escalated in subsequent recipients. If oneor more of these three additional patients experienced a dose-limiting toxicity,the maximum tolerated dose was exceeded, and three more study subjects weretreated at the next lower dose. The maximum tolerated dose defined in this trialwas considered the recommended starting dose for subsequent phase II testing.

Dose-Limiting Toxicity

Toxicity was graded according to the National Cancer Institute(NCI) Common Toxicity Criteria. For the purposes of defining the maximumtolerated dose, the dose-limiting toxicity pertained only to toxicity during thefirst cycle of treatment. The dose-limiting toxicity was defined as one or moreof the following: (1) grade 3 and 4 nonhematologic toxicity (excluding nausea,vomiting, fever, and anorexia), (2) grade 4 neutropenia lasting ³ 4 days, and(3) failure to recover neutrophils (³ 1,500/mL) or platelets (³ 100,000/mL) byday 29. During this phase I trial, days 8 and 9 treatment in cycle 1 were givenat full doses if the neutrophil count was > 1,500/mL, platelets ³100,000/mL, and there were no nonhematologic toxicities worse than grade 1. Day8 and 9 treatments in cycle 1 were held in patients with either counts belowthis level or nonhematologic toxicity worse than grade 1 on that day. Doseadjustments for cycle 2 or later are enumerated in Table2.

Patient Evaluation

At enrollment, patients were evaluated with a complete historyand physical examination, including a performance status assessment. Requiredblood counts, serum chemistries, and urinalysis were completed within 14 days ofstudy entry. Any x-ray, scan, computed tomography, magnetic reasonance imaging,or ultrasound that was utilized for tumor measurement in patients withmeasurable or evaluable disease had to have been performed within the 28 daysprior to study entry. During the first cycle of chemotherapy, a physicianmonitored patients at least weekly. For subsequent cycles, patients wereassessed at each chemotherapy visit. A complete blood count (CBC) withdifferential and serum chemistries was performed on days 1 and 8, withadditional CBCs with differential on days 11, 14, 17, and 20 (if indicated) ofeach cycle.

In patients with measurable or evaluable disease, evaluation oftumor response was performed every two cycles. A complete response was definedas disappearance of all clinically detectable malignant disease, and a partialresponse was a ³ 50% decrease in the sum of the products of perpendiculardiameters of all measured lesions. No new lesions could appear, and no existinglesions could enlarge by > 25%. Progressive disease was defined as a ³ 25% increase in the sum of products of measured lesions over the smallest sumobserved (over baseline, if there was no decrease), reappearance of any lesionthat had disappeared, or appearance of any new lesion. Stable disease wasdefined as findings that did not qualify for complete or partial response, orprogressive disease.

Results

Patient Characteristics

A total of 25 patients were registered onto study at theHollings Cancer Center, Medical University of South Carolina (the baselinecharacteristics and tumor types of all 25 patients are listed in Table 3 andTable 4,respectively). Nineteen patients were accrued in arm A: 15 males/4 females,median age: 61 years, with a Zubrod performance status of 0/1/2 (n = 3/12/4,respectively). Seventeen of these patients were evaluable (one died of anunrelated cause on cycle 1, and another withdrew consent before beginningtreatment). Forty-two cycles of schedule A therapy have been delivered (mean:2.5 cycles/patient).

In arm B, six patients were accrued: 2 males/4 females, medianage: 61 years, with a Zubrod performance status of0/1/2 (n = 0/5/1, respectively). Five patients were evaluable for maximumtolerated dose; one patient inadvertently received granulocytecolony-stimulating factor (G-CSF) on cycle 1. Twenty-five cycles of schedule Bchemotherapy have been delivered (mean: 4.1 cycles/patient).

Hematologic Toxicities

Hematologic toxicities during cycle 1, which were dose limitingin this phase I trial, are shown in Table 5. At dose level 1 in arm A (docetaxelat 20 mg/m²), no dose-limiting hematologic toxicities were noted. However, onepatient at dose level 1 with previously treated esophageal cancer experienced agrade 4 neutropenic episode for < 4 days. This patient received cycle 2 onschedule when again he had a grade 4 uncomplicated neutropenia of shortduration. He was taken off study after cycle 2 due to disease progression. Atdose level 2 (docetaxel at 25 mg/m²), one patient with previously treated renalcell carcinoma experienced a febrile neutropenic dose-limiting toxicity, andanother patient with previously treated head and neck cancer had cycle 2 heldfor 1 week due to a neutrophil count of 1,400/mL.

Dose level 3 (docetaxel at 30 mg/m²) was associated with significant hematologic toxicity. One patient withpreviously treated head and neck cancer experienced a grade 3 infection withgrade 4 neutropenia, and died after completing cycle 1 due to diseaseprogression. The other two patients—one with previously treated head and neckcancer and another with previously treated hepatocellular carcinoma—had cycle1, day 8 and 9 therapy held due to low neutrophil counts. At the confirmatorycohort level 2 (docetaxel at 25 mg/m²), three hematologic dose-limitingtoxicities were observed. Two patients—one with previously treated coloncancer and the other with previously treated head and neck cancer—experiencedgrade 3 infection. The patient with colon cancer also experienced grade 4neutropenia lasting > 4 days. The head and neck cancer patient withdrewconsent after completing the first cycle, and the colon cancer patient haddisease progression after two cycles.

On arm B (docetaxel at 45 mg/m² given on day 8 only), onepatient with a previously untreated breast cancer and a new pancreatic mass hadneutropenia for > 4 days. Since her breast cancer shrank during the firstcycle, she received additional chemotherapy at a 25% dose reduction of all threedrugs, per protocol. With the dose reduction, she had uncomplicated grade 4neutropenia for < 4 days on cycle 2. On cycle 3, she had grade 3 neutropeniaon day 11 and grade 4 neutropenia on day 14. Also on cycle 3, she had a redblood cell transfusion on day 9 to correct a hemoglobin level of 6.8 g/dL.

A patient with previously treated non-small-cell lung cancerhad cycle 1, day 8 and 9 therapy held due to low neutrophil counts. This patientcould not be evaluated for cycle 1 dose-limiting toxicity due to aweather-related interruption in his planned evaluation of blood counts. Thispatient was treated on cycle 2 at the full dose and did not experiencedose-limiting toxicities; he was removed from the study after two cycles due todisease progression. Another patient on arm B could not be evaluated fordose-limiting toxicity due to erroneous administration of G-CSF on cycle 1. Thispatient, with previously untreated pancreatic cancer, withdrew consent after onecycle reporting that "she had felt much better" before starting thetreatment with chemotherapy; on cycle 1 she had developed nausea/vomiting for anunspecified number of days.

In arm A, no cumulative hematologic toxicities have beendemonstrated. However, none of the patients had more than six cycles ofchemotherapy on protocol. In arm B, one patient with previously treated non-small-celllung cancer had cumulative hematologic toxicity despite a 25% dose reduction ofdocetaxel and irinotecan due to grade 3 diarrhea experienced during cycle 1. Hewas transfused during cycle 6 for grade 3 anemia, and was admitted to thehospital on cycle 8 with grade 4 thrombocytopenia and severe epistaxis andhematochezia; he stopped therapy after eight cycles with stable disease.

Nonhematologic Toxicities

Nonhematologic toxicities are enumerated in Table6. In doselevel 1 of schedule A, no nonhematologic dose-limiting toxicities were observed.One patient with previously treated non-small-cell lung cancer had reversiblegrade 3 acute renal failure along with grade 2 diarrhea and nausea/vomiting,which occurred on cycle 5. The renal failure was thought to be secondary todehydration, and the patient was taken off protocol following this episode.

At dose level 2, one patient with previously untreated renalcell carcinoma experienced grade 3 diarrhea for 1 day; he did not takeloperamide as instructed. Per protocol requirement, this patient then receivedan additional five cycles at 75% of the initial dose, and he did not haveadditional grade 3 diarrhea. In addition, one patient on cycle 2 with previouslyuntreated malignant mesothelioma experienced grade 3 diarrhea and asthenia. Thedoses of chemotherapy were reduced by 25%, with improvement of thissymptomatology on the following cycles.

Of the three patients at dose level 3, one patient with head andneck cancer had grade 3 nausea/vomiting for 3 days. Another patient withpreviously treated hepatocellular carcinoma had reversible grade 3 asthenia andfatigue, leading to reduction in the doses of all three drugs by 25%. One of thefour patients on cohort dose level 2.1 with previously treated colon cancerexperienced grade 3 diarrhea starting on cycle 1, day 10, and lasting for 3days. This patient did not take loperamide as recommended.

On arm B, two patients had nonhematologic dose-limitingtoxicities. One patient with previously treated non-small-cell lung cancer hadgrade 4 diarrhea, and was admitted to the hospital with orthostatic hypotension;he was removed from the study after one cycle due to disease progression. Theother patient, also with non-small-cell lung cancer, experienced grade 3diarrhea for 14 days, and was also admitted to the hospital. This resulted in a25% dose reduction of docetaxel and irinotecan, and the diarrhea was improved inall subsequent cycles. The patient with previously untreated pancreas cancer whowas erroneously given G-CSF during cycle 1 had grade 3 nausea and vomiting forseveral days during the first cycle.

Response Evaluation

The response data for arms A and B are shown in Table7. Nopatients had a confirmed partial or complete response, while a total of fivepatients experienced stable disease for ³ 4 cycles. Five additional patientshad some evidence of initial tumor shrinkage, but then either progressed rapidlyor were removed from study for other reasons. For example, at dose level 1 ofschedule A, one patient with heavily previously treated non-small-cell lungcancer showed tumor shrinkage for five cycles. She had multiple, too small tocount, small lung nodules, many of which significantly decreased in size andsome that disappeared. She was removed from protocol due to renal toxicity.

At dose level 2, two patients showed tumor shrinkage: one had apreviously untreated malignant mesothelioma that had tumor shrinkage after twocycles and stable radiologic findings after four cycles when he was taken offprotocol; the other, with a previously treated composite non-small-cell lungcancer and small-cell lung cancer, had tumor shrinkage after the first course ofchemotherapy, but declined further treatment due to personal reasons. At doselevel 3, one patient with previously treated head and neck cancer demonstrated> 50% tumor regression after two cycles of therapy; this patient was takenoff protocol after his third cycle of chemotherapy because of worsening of hisperformance status and nutritional problems.

On arm B, the patient with an untreated breast cancer and acoincident pancreatic mass of uncertain etiology had rapid reduction in herbreast cancer without shrinkage of the pancreatic mass. She declined additionalprotocol treatment after four cycles of chemotherapy.

Recommended Doses forPhase II Studies

The maximum tolerated dose of docetaxel given intravenously ondays 1 and 8 followed by the combination of gemcitabine at 1,000 mg/m² andirinotecan at 100 mg/m² given on days 2 and 9, every 3 weeks as described here(schedule A) is 20 mg/m². This regimen is appropriate for phase II testing. Wedo not plan additional phase I testing of our schedule B and cannot recommend aphase II regimen using day 8-only docetaxel in combination with day 1 and 9gemcitabine and irinotecan.

Trial Course

In arm A at dose level 1, no cycle 1 dose-limiting toxicitieswere noted. At dose level 2, one of the three original patients experienced acycle 1 dose-limiting toxicity (febrile neutropenia). This cohort was thenexpanded to include three additional patients (one of these patients died of acause unrelated to the study). Therefore, an additional patient was accrued atthe cohort 2 level. Among these four patients, one experienced 1 day of grade 3diarrhea, but no other dose-limiting toxicities were observed. Due to thedisparate nature of the dose-limiting toxicities noted at dose level 2 plus thebrief, barely grade 3 diarrhea leading to one of the dose-limiting toxicitydesignations, written permission was obtained from the MUSC Institutional ReviewBoard to proceed with dose level 3 accrual.

At dose level 3, all three patients experienced significanttoxicities during cycle 1; one patient had two dose-limiting toxicities—grade3 infection and grade 4 neutropenia for > 4 days. The other two patients wereunable to receive cycle 1, day 8 and 9 chemotherapy due to ³ grade 2neutropenia. No additional patients were entered at this dose level.

In an attempt to confirm dose level 2 as the maximum tolerateddose, four additional patients were then accrued to the docetaxel cohort dose of25 mg/m². Two of these patients experienced dose-limiting toxicities (grade 3diarrhea and grade 4 neutropenia for ³ 4 days), one patient with previouslytreated colon cancer experienced grade 3 infection, and a patient withpreviously treated head and neck cancer also had grade 3 infection. The nextlower cohort dose (docetaxel at 20 mg/m² days 1 and 8, with gemcitabine at 1,000mg/m² followed by irinotecan at 100 mg/m², on days 2 and 9) was then consideredthe maximum tolerated dose.

Because the maximum tolerated dose was the original startingdose of the trial, written approval was obtained from the Institutional ReviewBoard to amend further the study protocol. In this revision, the dose ofirinotecan was reduced to 80 mg/m² to maintain the sequence with gemcitabine at1,000 mg/m² on days 2 and 9. Docetaxel was again given at 25mg/m² on days 1 and8. The first patient enrolled at this dose level experienced no dose-limitingtoxicities; we plan to continue accrual with docetaxel at 25 mg/m² on days 1 and8 along with gemcitabine at 1,000 mg/m² followed by irinotecan at a modifieddose of 80 mg/m² on days 2 and 9.

In arm B, significant dose-limiting toxicities were observed atthe starting docetaxel dose of 45 mg/m². The dose-limiting toxicities in arm Bwere grade 3 diarrhea in one patient with previously treated non-small-celllung cancer, grade 4 diarrhea in another patient with previously untreated non-small-celllung cancer, and prolonged grade 4 neutropenia in a previously untreated breastcancer patient with a concomitant pancreatic mass. In addition, a patient withpreviously untreated pancreas cancer had grade 3 nausea/vomiting. It was feltthat dose deescalation was not worth pursuing at this time, and arm B endedafter the first dose level had been investigated.

Discussion

Our previous phase I work[10] with gemcitabine and irinotecandefined the recommended phase II dose of gemcitabine at 1,000 mg/m² over 30minutes followed by irinotecan at 100 mg/m² over 90 minutes. Doses of each drugwere given weekly for 2 consecutive weeks (days 1 and 8) every 3 weeks, and thedose-limiting toxicity was diarrhea; myelosuppresion was modest. These findingssuggested the possibility that an additional drug could be combined with theIrinoGem regimen.

Based on preclinical data from our institution and thesubstantial single-agent activity of docetaxel in several solid tumors, wedeveloped two schedules of the three-drug regimen of docetaxel, irinotecan, andgemcitabine for phase I testing. The findings of this phase I work suggest thatthe combination of low doses of docetaxel on days 1 and 8, with full doses ofIrinoGem on days 2 and 9, is feasible without growth factor support. However,toxicity is not insignificant, and attempts at escalation of the dose ofdocetaxel above the initial cohort dose level of 20 mg/m² resulted in an excessof dose-limiting toxicities.

In our phase I trial of IrinoGem,[10] no episodes of febrile orprolonged neutropenia were observed during the first cycle of therapy at anydose level. In a phase II trial of IrinoGem in pancreatic cancer,[11] only a2.2% rate of grade 4 neutropenia (and no neutropenic fevers) was reported duringthe 394 cycles administered in 45 patients. However, the addition of evendocetaxel at 25 mg/m² on days 1 and 8 to full doses of IrinoGem on days 2 and 9led to several episodes of grade 4 neutropenia lasting > 4 days and/or grade3 infection. Disappointingly, this marked augmentation in hematologic toxicityand infectious episodes was not accompanied by a similarly dramatic increase inefficacy.

As single agents, docetaxel, irinotecan, and gemcitabine allshowed schedule-dependent dose-limiting toxicities. On weekly or "3 weeksout of 4" schedules, gemcitabine was associated with dose-limitingmyelosuppression.[12,13] For irinotecan, severe neutropenia (78% grade 3/4) wasthe dose-limiting toxicity when the agent was given in an every-3-week scheduleby Merrouche and colleagues.[14] With docetaxel, severe neutropenia occurred onthe every-3-week or the days 1 and 8 every-3-week schedule.[15] Weekly × 4every-6-week administration of irinotecan had severe diarrhea as thedose-limiting toxicity, despite the use of an aggressive regimen of loperamideat the onset of diarrhea.[16,17] In the study by Hainsworth et al[18] assessingweekly docetaxel, grade 4 asthenia was limiting at dose levels that rarelyproduced severe myelosuppression.

Grade 3 asthenia and fatigue were not noted in either our phaseI or II experience with IrinoGem. However, with the addition of low-dosedocetaxel, these toxicities became evident in later cycles of therapy. In theDIG regimen, 1 of 10 patients at a docetaxel dose of 25 mg/m², and another 3patients at 30 mg/m², experienced grade 3 asthenia and fatigue. In addition,grade 1 or 2 asthenia was frequently observed at docetaxel dose levels ³ 25mg/m² on days 1 and 8. This toxicity was reminiscent of the current experienceusing 6 consecutive weeks every 8 weeks of single-agent docetaxeladministration,[18] but occurred at lower dose levels than in prior work.

In the phase I study of weekly docetaxel, the incidences ofasthenia and fatigue were clearly dose dependent. Grade 3 asthenia was thedose-limiting toxicity in 2 of 10 patients at a docetaxel dose of 43 mg/m²; themaximum tolerated dose of docetaxel was 36 mg/m². At lower doses, fatigue andasthenia were never dose limiting. In follow-up phase II trials testing the sameweekly schedule of docetaxel in elderly lung cancer[19] and metastatic breastcancer patients,[20] the incidence of severe fatigue and asthenia was 14% and10% at respective docetaxel doses of 36 and 40 mg/m².

In a trial of single-agent docetaxel administered on days 1 and8 on an every-3-week cycle, asthenia was also dose dependent, but not doselimiting. In the phase I trial of Tomiak et al,[15] among 32 patients withrefractory solid malignancies treated at doses between 20 and 110 mg/m² percourse, 10 of 23 assessable patients receiving ³ 50mg/m² on days 1 and 8 experienced severe asthenia. However, no episodes ofsevere asthenia and fatigue were observed at lower dose levels. Our data suggestthat the triple combination may potentiate this toxicity, and we will carefullyassess this toxicity in additional patients treated with this combination.

As reported from the studies of Rowinsky and coworkers[16] andRothenberg et al,[17] diarrhea is the dose-limiting toxicity of weeklysingle-agent irinotecan. However, significant diarrhea is a much less commontoxicity with single-agent docetaxel or gemcitabine. For example, in a study of730 patients with metastatic breast cancer (and normal liver function testvalues) given docetaxel at 100 mg/m², severe diarrhea was reported in 6% ofpatients. In the gemcitabine database of 973 patients treated at starting dosesranging from 800 to 1,250 mg/m², the incidence of severe diarrhea was only 1%.However, combinations may produce enhanced diarrheal toxicity.

Spridonidis and collaborators[21] reported that with gemcitabineat a fixed dose of 800 mg/m² on days 1, 8, and 15, and escalated doses ofdocetaxel on day 1 every 4 weeks, grade ³ 2 diarrhea occurred in 17% ofpatients treated with docetaxel at 75 mg/m² and in 33% treated at 100mg/m² inphase I investigation. In the two-drug IrinoGem combination, diarrhea was thedose-limiting toxicity at an irinotecan dose of 115 mg/m² on days 1 and 8. Withthe DIG regimen, diarrhea was also dose limiting in two patients in arm A at 25mg/m² and in two patients in arm B at the initial dose of docetaxel at 45mg/m².

This phase I trial accrued predominantly pretreated patients,and previous exposure to all three drugs under study did not exclude patientsfrom study. Among the 25 patients accrued, five had radiologic tumor reductionsthat did not reach the formal criteria for a partial response, and twoadditional patients had stable disease for more than six cycles. The patientswith tumor shrinkage did not have confirmed responses for a variety of reasons(none related to tumor regrowth).

Full doses of IrinoGem (30-minute infusion of gemcitabine at1,000 mg/m² immediately followed by 90-minute infusion of irinotecan on days 2and 9) combined with a 30-minute infusion of docetaxel at 20 mg/m² on days 1 and8 can be used as the starting doses for this triplet in phase II trials.Myelosuppression should be manageable at this dose level, but cumulativeasthenia and fatigue will be more prominent than with IrinoGem alone. Currently,we plan to evaluate the escalation of the docetaxel dose to 25 and 30 mg/m² ondays 1 and 8, with gemcitabine at 1,000 mg/m² immediately followed by irinotecanat a reduced dose of 80 mg/m² on days 2 and 9.

References:

1. Gueritte-Voegelein F, Guenard D, Lavelle F, et al:Relationships between the structure of taxol analogues and their antimitoticactivity. J Med Chem 34:992-998, 1991.

2. Ringel I, Horwitz SB: Studies with RP 56976 (Taxotere): Asemisynthetic analogue of taxol. J Natl Cancer Inst 83:288-291, 1991.

3. Lund B, Kristjansen PEG, Hansen HH: Clinical and preclinicalactivity of 2¢,2¢-difluorodeoxycytidine (gemcitabine). Cancer Treat Rev19:45-55, 1993.

4. Plunkett W, Huang P, Searcy CE, et al: Gemcitabine:Preclinical pharmacology and mechanism of action. Semin Oncol 23(suppl 10):3-15,1996.

5. Kawato Y, Aonuma M, Hirota Y, et al: Intracellular roles ofSN-38, a metabolite of the camptothecin derivative CPT-11, in the antitumoreffect of CPT-11. Cancer Res 51:4187-4191, 1991.

6. Zoli W, Ricotti L, Dal Susino M, et al: Docetaxel andgemcitabine activity in NSCLC cell lines and in primary cultures from human lungcancer. Br J Cancer 81:609-615, 1999.

7. Vanhoefer U, Hilger RA, Harstrick A, et al:Schedule-dependent cytotoxic interactions of Irinotecan (SN-38) and docetaxel inhuman non-small-cell lung cancer (NSCLC) (abstract 649). Proc Am Assoc CancerRev 131, 1999.

8. Okishio K, Kudoh S, Hirata K, et al: Schedule dependentadditive effect of docetaxel and irinotecan in vitro (abstract 2252). Proc Jpn JCancer Res 86:619, 1995.

9. Bahadori HR, Rocha Lima CS, Green MR, et al: Synergisticeffect of gemcitabine and irinotecan (CPT-11) on breast and small cell lungcancer cell lines. Anticancer Res 19:5423-5428, 1999.

10. Rocha Lima CS, Leong SS, Sherman CA, et al: Phase I study ofirinotecan and gemcitabine in patients with solid tumors. Cancer Therapeut2:58-66, 1999.

11. Rocha Lima CS, Savarese D, Bruckner H, et al: Phase II studyof irinotecan and gemcitabine (IrinoGem) in patients with previously untreatedadvanced and metastatic pancreatic cancer (PC) (abstract 1023). Proc Am Soc ClinOncol 19:263a, 2000.

12. Eckardt JR, Von Hoff DD: New drugs in clinical developmentin the United States. Hematol Oncol Clin North Am 8:305-332, 1994.

13. Eckardt J, Eckard G, Villalona-Calero M, et al: Newanticancer agents in clinical development. Oncology 9:1321-1331, 1995.

14. Merrouche Y, Extra J, Abigerges D, et al: Highdose-intensity of irinotecan administered every 3 weeks in advanced cancerpatients: A feasibility study. J Clin Oncol 15:1080-1086, 1997.

15. Tomiak E, Piccart MJ, Kerger J, et al: Phase I study ofdocetaxel administered as a 1-hour intravenous infusion on a weekly basis. JClin Oncol 12:1458-1467, 1994.

16. Rowinsky EK, Grochow LB, Ettinger DS, et al: Phase I andpharmacological study of the novel topoisomerase I inhibitor7-ethyl-10-[4-(1-piperidino)-1-piperidino] carbonyloxycamptothecin (CPT-11)administered as a 90-minute infusion every 3 weeks. Cancer Res 54:427-436, 1994.

17. Rothenberg ML, Kuhn JG, Burris HA 3rd, et al: Phase I andpharmacokinetic trial of weekly CPT-11. J Clin Oncol 11:2194-2204, 1993.

18. Hainsworth JD, Burris HA 3rd, Erland JB, et al: Phase Itrial of docetaxel administered by weekly infusion in patients with advancedrefractory cancer. J Clin Oncol 16:2164-2168, 1998.

19. Hainsworth JD, Burris HA 3rd, Litchy S, et al: Weeklydocetaxel in the treatment of elderly patients with advanced NSCLC. A MinniePearl Cancer Research Network Phase II trial. Cancer 89:328-33, 2000.

20. Burstein HJ, Manola J, Younger J, et al: Docetaxeladministered on a weekly basis for metastatic breast cancer. J Clin Oncol18:1212-1219, 2000.

21. Spridonidis CH, Laufman LR, Jones J, et al: Phase I study ofdocetaxel dose escalation in combination with fixed weekly gemcitabine inpatients with advanced malignancies. J Clin Oncol 16:3866-3873, 1998.