Efficacy of Oxaliplatin in the Treatment of Colorectal Cancer

ABSTRACT: Colorectal cancer is one of the leading causes of cancer death. The mainstay of chemotherapy in colorectal cancer patients for the past 40 years has been fluorouracil (5-FU). Oxaliplatin (Eloxatin) is a novel platinum compound with promising activity in colorectal cancer. As a single agent, oxaliplatin has produced response rates of 12% to 24% in patients with previously untreated advanced colorectal cancer, and 10% to 11% in patients with relapsed or refractory advanced colorectal cancer. In phase II trials, oxaliplatin combined with 5-FU, with or without leucovorin, was associated with response rates of 60% and higher when used as front-line therapy, and when used in patients with relapsed or refractory advanced colorectal cancer, response rates ranged from 25% to 50%. In the front-line setting, two randomized trials of 5-FU and leucovorin, with or without oxaliplatin, demonstrated that the addition of oxaliplatin significantly increases response rate and time to tumor progression, but not survival, over 5-FU plus leucovorin alone. The reasons for this discrepancy are unclear, and several possibilities are being considered. Additional phase III trials are underway to clarify the contribution of oxaliplatin in the treatment of patients with locally advanced and metastatic colorectal cancer. [ONCOLOGY 14(Suppl 11):9-14, 2000]

Introduction

The American Cancer Society estimates that 130,200 cases ofcolorectal cancer are diagnosed in the United States each year. Of these, 32,000are stage IV (ie, metastatic disease) at the time of original diagnosis. Inaddition, each year, another 39,000 patients previously diagnosed with stage Ito stage III disease develop metastatic disease. As a result, approximately71,000 people annually present with metastatic colorectal cancer. Approximately10% of these patients will have isolated hepatic metastases that can besurgically resected, while the remainder will be candidates for systemicchemotherapy.

For nearly 40 years, the mainstay of chemotherapy for patientswith metastatic colorectal cancer has been fluorouracil (5-FU). Threemeta-analyses assessed the efficacy of single-agent5-FU vs biochemical modulation of5-FU with leucovorin, sequential treatment with methotrexate followed by5-FU, and bolus administration vs IV infusion of 5-FU.[1-3] Regimens thatincluded either leucovorin or methotrexate or 5-FU administered as an infusionimproved the response rate significantly (by 50% to 100%), but the improvementin overall survival was modest, from 0.5 to 2 months (Table1). Notably, overallsurvival with 5-FU-based regimens consistently averaged 12 months or less.

A European trial demonstrated that the schedule ofadministration regimen of 5-FU influences both efficacy and safety. This trialrandomized patients with metastatic colorectal cancer and no prior treatment formetastatic disease to treatment with the de Gramont regimen (combiningleucovorin and an IV bolus dose and infusion of 5-FU plus leucovorinadministered over 2 hours for 2 consecutive days every 2 weeks) or the MayoClinic regimen (consisting of leucovorin and a daily bolus injection of 5-FU for5 consecutive days every 4 to 5 weeks).[4]

The response rate with the de Gramont regimen was more thantwice that of the Mayo Clinic regimen (32.6% vs 14.5%; P = .0004), andprogression-free survival was significantly higher (6.4 mo vs 5.1 mo; P = .001).There was a trend toward improved median survival with the de Gramont regimenthat did not reach statistical significance (14.3 months vs 13.1 months; P =.067). However, the de Gramont regimen was associated with lower rates of grade3/4 neutropenia, mucositis, and diarrhea, but higher rates of grade 1 and 2epistaxis and conjunctivitis.

The emergence of two novel cytotoxic agents, irinotecan (CPT-11,Camptosar) and oxaliplatin (Eloxatin), over the past 5 years caused a markedchange in the direction of clinical trials in patients with metastaticcolorectal cancer. Rather than focus on biochemical modulation of 5-FU, recentclinical trials have evaluated the impact of adding each of these agents tostandard therapy. This review will concentrate on data from two randomizedtrials that evaluated 5-FU/leucovorin, with or without oxaliplatin, in thissetting.

Overview of Oxaliplatin

Like other platinum compounds, oxaliplatin inhibits DNAreplication and transcription through the formation of intra- and interstrandDNA adducts. However, unlike cisplatin (Platinol) or carboplatin (Paraplatin),oxaliplatin is a diaminocyclohexane (DACH) platinum (Figure1). The DNA adductsof oxaliplatin are bulkier and more hydrophobic than those of cisplatin orcarboplatin, and these properties contribute to its enhanced activity ininhibiting DNA synthesis and its lack of cross-resistance with cisplatin andcarboplatin.[6-8]

Oxaliplatin platinates DNA much faster than does cisplatin—15minutes for oxaliplatin vs an average of 48 hours for cisplatin. However, unlikecisplatin and carboplatin, oxaliplatin adducts are less susceptible to DNAexcision repair enzymes, and have equivalent activity in DNA mismatchrepair-proficient and -deficient cells in vitro.[9]

Activity Spans Several Cell Lines

In preclinical studies, oxaliplatin demonstrated activityagainst six of eight colorectal cancer cell lines in the National CancerInstitute automated screening panel, whereas cisplatin and carboplatin had noactivity against these cell lines (Figure 2).[7] In in vitro studies,oxaliplatin was synergistic with5-FU and leucovorin.[9-11] In one series, oxaliplatin and 5-FU demonstratedsynergy in 78% of the situations tested, including four human colorectal cancercell lines, three different drug administration sequences, and three differentdurations of exposure to 5-FU.

In phase I studies, oxaliplatin was administered over 2 to 6hours on an every 3-week schedule. The dose-limiting toxicity in phase I studieswas cumulative, reversible paresthesia, and occurred at a dose of 180 to 200mg/m².[12,13] Therefore, the dose of oxaliplatin chosen for phase II studies was130 mg/m² administered on an every 3-week schedule. In order to accommodate theevery 2-week scheduling of the de Gramont regimen, a dose of 85 mg/m² wasselected in order to maintain a stable dose intensity of42.5 mg/m²/wk.

Oxaliplatin as First-Line Therapy for Advanced Colorectal Cancer

The clinical activity of oxaliplatin monotherapy has recentlybeen reviewed.[14] In two phase II trials conducted in patients with previouslyuntreated colorectal cancer, 12% and 24% of those who received 130 mg/m² ofsingle-agent oxaliplatin achieved objective responses. The median durations ofresponse were 6 and 7 months, the median times to tumor progression were 4months in both studies, and the median survivals were 13 and 14.5 months.

Two phase II trials evaluated the combination of 5-FU/leucovorinand oxaliplatin as initial therapy for patients with metastatic colorectalcancer.[15,16] Both were performed using chronomodulated drug administrationschedules, and both reported impressive objective response rates of 59% and 67%.The median progression-free survivals were 8.4 and 11 months, and mediansurvivals were 15 and 18.5 months.

These results prompted the initiation of two randomized trialsof 5-FU/leucovorin, with or without oxaliplatin, as first-line chemotherapy inpatients with metastatic colorectal cancer. In one trial, the drugs wereadministered in a chronomodulated fashion, and in the other, they were givenusing a nonchronomodulated, biweekly schedule.

Phase III Pivotal Trial: de Gramont 5-FU/Leucovorin With orWithout Oxaliplatin

This European multinational, multicenter phase III trial (alsoknown as EFC 2962) evaluated the contribution of oxaliplatin as first-linetreatment in patients with metastatic colorectal cancer.[17] Between August 1995and July 1997, 420 patients were enrolled in the trial. Eligibility criteriaincluded histologically confirmed adenocarcinoma of the colon or rectum,inoperable metastatic disease at enrollment, no prior immunotherapy orchemotherapy for metastatic disease (but adjuvant chemotherapy that had beencompleted at least 6 months prior to enrollment was allowed), at least onebidimensionally measurable lesion on magnetic resonance imaging (MRI) orcomputed tomography (CT) scan, a World Health Organization (WHO) performancestatus (PS) ≤ 2, and adequate serum chemistry values and bone marrow reserve.

Patients randomized to the control arm were treated withleucovorin 200 mg/m² IV over 2 hours, followed by5-FU 400 mg/m² administered as an IV bolus followed by a 22-hour infusion at 600mg/m² on days 1 and 2 every 2 weeks. Patients randomized to the experimental armreceived the same schedule of 5-FU plus leucovorin, with oxaliplatin 85 mg/m² IVover 2 hours on day 1 only.

The primary end point of the trial was progression-freesurvival. Secondary end points included response rate as determined by anindependent review panel based on a confirmatory scan obtained at 4 weeks,overall survival, quality of life (QOL), and safety of the regimens. The nullhypothesis was that there was no difference in progression-free survival, andthe alternative hypothesis was that there would be a 43% difference in medianprogression-free survival (from 7 to 10 months), which was considered aclinically significant improvement. The study was designed to have ³ 80% powerto detect a significant difference at the 0.05 level.

Trial Results

The addition of oxaliplatin to 5-FU/leucovorin significantlyimproved median progression-free survival over5-FU/leucovorin alone (8.2 months vs 6.0 months; hazard ratio, 0.67, 95%confidence interval [CI], 0.54 to 0.83; P = .0003) (Figure3).[17] Amultivariate analysis identified three independent prognostic factors thatcontribute to an improved progression-free survival: low baseline lactatedehydrogenase (LDH) level, good performance status, and randomization to the5-FU/leucovorin plus oxaliplatin treatment arm. An analysis of secondary endpoints revealed that the independently verified, radiographically confirmedobjective response rate was significantly higher in the experimental arm than inthe control arm-50.0% [95% CI, 46.1% to 54.9%] vs 21.9% [95% CI, 17.9% to25.9%]; P = .0001). Although a trend toward longer median survival wasidentified in the oxaliplatin-treated group, 16.2 vs 14.7 months, thisdifference did not achieve statistical significance (death hazard ratio: 0.83[95% CI, 0.65 to1.06]; P = .12).

In an attempt to understand the discrepancy between theimprovements in progression-free and overall survival, a multivariate analysiswas performed based on protocol-specified baseline prognostic factors. Thisanalysis identified five factors that were significant predictors of overallsurvival: WHO PS (risk ratio, 1.50; P = .0001), number of metastatic sites (riskratio, 1.17; P = .0029), alkaline phosphatase (risk ratio, 1.34; P = .0062), LDH(risk ratio, 1.94; P = .0001), and treatment arm (risk ratio, 0.80; P = .0001).

Quality of Life Parameters

Clinical benefit from oxaliplatin was also measured by assessingquality-of-life (QOL) parameters using the European Organization for Researchand Treatment of Cancer (EORTC) QLQ-C30 scale.[18] Overall, the addition ofoxaliplatin to a 5-FU/leucovorin regimen did not compromise patients’ QOL, andthere was no deterioration in the rate of change in performance status.[17] Inaddition, the time to treatment failure (TTF), which includes time to tumorprogression and discontinuation of therapy due to toxicity, death, or patientwithdrawal, was significantly improved with the addition of oxaliplatin. The TTF was 6.7 months for patients treated with the 5-FU/leucovorinplus oxaliplatin combination vs 5.4 months for patients who received5-FU/leucovorin alone as front-line therapy (hazard ratio, 0.74 [95% CI, 0.61 to0.90]; P = .003) (data on file, Sanofi-Synthelabo).

Phase II/ III Supportive Trial: Chronomodulated Schedule

A phase II/III randomized clinical trial (EFC 2961) wasperformed to evaluate chronomodulated IV delivery of5-FU 700 mg/m²/d and leucovorin 300 mg/m²/d (with peak infusion at 4:00 am) ondays 1 to 5 every 3 weeks, with or without oxaliplatin 125 mg/m², administeredover 6 hours on day 1 every 3 weeks in patients with metastatic colorectalcancer who had not been treated previously for metastatic disease.[19] Inclusioncriteria were similar to those of the phase III oxaliplatin clinical trial (EFC2962) described above. The primary end point was objective response rates.Secondary end points included progression-free and overall survival.

Objective Response Rates

A total of 100 patients were randomized to each treatment arm.The two treatment groups were comparable in most baseline demographic anddisease characteristics. The objective response rate (which includes only thosepatients with responses lasting at least 9 weeks and verified by an independentradiologic review committee) was significantly higher in patients treated with5-FU/leucovorin plus oxaliplatin (53%; 95% CI, 42% to 63%) than in patients inthe 5-FU/leucovorin arm (16%; 95% CI, 9% to 24%; P < .0001). The lowerobjective response rates reported in this trial compared with the previous trialmay be because the confirmatory scan was done at 9 weeks rather than at 4 weeksas in the previous trial.

Progression-Free Survival

The addition of oxaliplatin also had a beneficial impact on thesecondary efficacy end point of progression-free survival. Patients treated withoxaliplatin had a progression-free survival of 8.7 months (95% CI, 7.4 to 9.2 months)that was significantly longer than the 6.1-month progression-free survival forpatients who received 5-FU/leucovorin alone as front-line therapy (95% CI, 4.1months to 7.4 months; P = .048). Median survival for the two treatment groupswas similar—19.4 months (95% CI, 15.4 months to 23.4 months) for patients whoreceived the three-drug regimen vs 19.9 months (95% CI, 14.0 months to 25.7months) for those treated with the two-drug regimen (P = not significant). Theestimated 2- and 3-year survival rates were 45% and 30%, respectively, for the5-FU/leucovorin arm, and 37% and 23.5%, respectively, for the 5-FU/leucovorinplus oxaliplatin arm. Regression analysis revealed that the number of organs involved bytumor (P = .0017), PS (P = .001), and the percentage of the liver involved bytumor (£ 25% vs > 25%; P = .0013) were significant prognostic indicators ofsurvival.

Although first-line chemotherapy was mandated by the protocol,subsequent treatment was not, and patients and physicians were encouraged topursue the most aggressive therapy. Of interest is that 57% of the patients inthe 5-FU/leucovorin control arm were treated with 5-FU/leucovorin plus oxaliplatin at the time of tumor progression. In patients who responded to treatment,surgical resection of remaining disease was aggressively pursued. Followingchemotherapy, 21% of patients in the control arm and 32% of patients in theinvestigational arm were able to undergo metastasectomy.

The excellent median survival reported in this trial is likely areflection of the totality of treatment, including the aggressive surgicalapproach, the effect of protocol-specified front-line chemotherapy, and,potentially, of subsequent second-line therapies.

No Increase in Overall Survival

The findings from both of these randomized studies (EFC 2961 andEFC 2962) were very consistent with respect to documenting the beneficial impactof oxaliplatin on objective response rate and progression-free survival. Equallyconsistent was the lack of a detectable impact of front-line oxaliplatin onprolonging overall survival. There are several possible explanations for thisdiscrepancy.

The first is that oxaliplatin is a weak drug, and that itseffects are too transient to have an impact on overall survival. While thiscould explain the higher objective response rate observed in theoxaliplatin-treated patients, a weak or inactive drug would not be expected toimprove progression-free survival by the 37% to 43% observed in these twotrials.

A second explanation is that oxaliplatin’s activity as part ofsecond- and third-line therapy may have obscured any survival advantage obtainedfrom its use as front-line therapy. In fact, 57% of patients assigned to thecontrol arm of the study reported by Giacchetti and colleagues[19] received anoxaliplatin-containing regimen at the time of tumor progression.

A third possibility is that the use of other salvage therapies,such as irinotecan and surgery, could have influenced survival without affectingthe more proximal end points of the study, such as objective response rate andprogression-free survival.

A final possibility is that subtle, but clinically important,differences in baseline patient characteristics could have obscured the impactof oxaliplatin on survival. When these baseline differences were taken intoaccount by Cox regression analysis, overall survival was significantly enhancedby the addition of oxaliplatin to front-line chemotherapy in the EFC 2962 studyof de Gramont and coworkers (death hazard ratio, 0.80; P = .0001).[17] However,which, if any, of these factors may have contributed to the outcome of these twostudies cannot be determined with certainty. Only prospective studies thatcontrol for each of these factors would be able to answer these questions withany degree of certainty.

Oxaliplatin as Second-Line Therapy for AdvancedColorectal Cancer

Three studies of single-agent oxaliplatin in patients withrecurrent colorectal cancer have been reviewed recently.[14] In two of thesestudies, oxaliplatin was administered in standard fashion (130 mg/m² over 2hours). In the other trial, oxaliplatin was given on a chronomodulated schedule(30 mg/m²/d ´ 5 days administered as a 24-hour infusion with the peak infusionrate at 4:00 pm). Objective responses were reported in 10% to 11% of patientsand median survival was 8 to 10 months.

The majority of studies in the second-line setting haveevaluated the activity of oxaliplatin combined with 5-FU, with or withoutleucovorin. Overall, objective response rates ranging from 24% to 55% have beenreported for this combination in patients with relapsed colorectal cancer.[20]The progression-free survival in these studies ranged from 6 to 10 months andoverall survival ranged from 7 to 17 months. While these results are veryencouraging, a number of important questions remain unanswered.

First, how much does 5-FU, with or without leucovorin, add tothe activity of salvage oxaliplatin? While response rate, progression-freesurvival, and overall survival data for the combination appear to comparefavorably with data generated with single-agent oxaliplatin in this setting, nohead-to-head comparative trials have ever addressed this question in thesecond-line setting. Second, does chronomodulation of oxaliplatin, 5-FU, andleucovorin significantly affect the activity and/or toxicity of thiscombination? Third, how active is second-line oxaliplatin in patients whoprogress following first-line treatment with irinotecan, 5-FU, and leucovorin?Finally, how much of the second-line antitumor activity is attributable to achange in the dose and schedule of administration of 5-FU, and how much is dueto the addition of oxaliplatin?

One phase II trial (EFC 2917) addressed the last question, inpart, by adding oxaliplatin to the same 5-FU plus leucovorin schedule on whichthe patient had progressed.[21] Objective responses were obtained in 19 of 172patients (RR = 11%), median progression-free survival was 4.3 months, and mediansurvival was 10.1 months.

A phase III trial (EFC 4584) opened in North America recentlythat may help to address several of these questions. In this trial, 780 patientswith progressive metastatic colorectal cancer following treatment withirinotecan, 5-FU, and leucovorin on the weekly × 4 schedule will be randomizedto one of three treatment arms (see Figure 4). One group will receive bolus plusinfusional 5-FU and leucovorin on an every 2-week basis (de Gramont regimen,FU5LV2). Another will receive single-agent oxaliplatin 85 mg/m² q2wk. A thirdgroup will receive the combination of oxaliplatin, 5-FU, and leucovorin q2wk(folfox4).

The primary end point of this trial is survival; it has a powerof 0.88 to detect a 3-month difference in survival (8 months vs 11 months) witha two-sided significance level of .05. It will also collect data regarding timeto onset or deterioration of tumor-related symptoms. The investigatorsanticipate that accrual should be completed in late 2001, and results willbecome available by late 2002.

Conclusions

Oxaliplatin is the first platinum analog to demonstrateconsistent antitumor activity against advanced colorectal cancer. Phase IIItrials have demonstrated that the addition of oxaliplatin to 5-FU plusleucovorin more than doubled objective response rates and improvedprogression-free survival by 45% compared to 5-FU plus leucovorin alone.Surprisingly, however, these advantages did not translate into improvements inoverall survival. Although several possible explanations have been proposed,including differences in baseline patient characteristics and the impact ofsubsequent salvage therapies on survival, it is not clear how much each of thesecontributed to this observation.

Oxaliplatin is also active when used as second-line therapyfollowing progression on 5-FU and leucovorin. A new question has emergedregarding the activity of oxaliplatin in patients who progress followingfront-line therapy with irinotecan, 5-FU, and leucovorin. A phase III trial hasbeen initiated to evaluate the relative contributions of changing the dose andadministration schedule of 5-FU and leucovorin, the impact of single-agentoxaliplatin, and the combination of both factors on survival. Information fromthis and other clinical trials will help to clarify the role of oxaliplatin inthe treatment of patients with advanced colorectal cancer.

References:

1. Advanced Colorectal Cancer Meta-Analysis Group: Modulation offluorouracil by leucovorin in patients with advanced colorectal cancer: Evidencein terms of response rate. Advanced Colorectal Cancer Meta-Analysis Project. JClin Oncol 10:896-903, 1992.

2. Advanced Colorectal Cancer Meta-Analysis Group: Meta-analysisof randomized trials testing the biochemical modulation of fluorouracil bymethotrexate in metastatic colorectal cancer. Advanced Colorectal CancerMeta-Analysis Project. J Clin Oncol 12:960-969, 1994.

3. Meta-Analysis Group in Cancer: Efficacy of continuousinfusion of fluorouracil compared with bolus administration in advancedcolorectal cancer. Meta-Analysis Group in Cancer. J Clin Oncol 16:301-308, 1998.

4. de Gramont A, Bosset JF, Milan C, et al: Randomized trialcomparing monthly low-dose leucovorin and fluorouracil bolus plus continuousinfusion for advanced colorectal cancer: A French Intergroup Study. J Clin Oncol15:808-815, 1997.

5. Saris CP, van de Vaart P, Rietbroek RC, et al: In vitroformation of DNA adducts by cisplatin, lobaplatin, and oxaliplatin in calfthymus DNA in solution and in cultured human cells. Carcinogenesis 17:2763-2769,1996.

6. Mamenta EL, Poma EE, Kaufmann WK, et al: Enhanced replicativebypass of platinum-DNA adducts in cisplatin-resistant human ovarian carcinomacell lines. Cancer Res 54:3500-3505, 1994.

7. Rixe O, Ortuzar W, Alvarez M, et al: Oxaliplatin,tetraplatin, cisplatin, and carboplatin: Spectrum of activity in drug-resistantcell lines and the cell lines of the National Cancer Institute’s AnticancerDrug Screen panel. Biochem Pharmacol 52:1855-1865, 1996.

8. Schmidt W, Chaney SG: Role of carrier ligand in platinumresistance of human carcinoma cell lines. Cancer Res 53:799-805, 1993.

9. Raymond E, Faivre S, Woynarowski JM, et al: Oxaliplatin:Mechanism of action and antineoplastic activity. Semin Oncol 25(2 suppl 5):4-12,1998.

10. Raymond E, Buquet-Fagot C, Djelloul S, et al: Antitumoractivity of oxaliplatin in combination with fluorouracil and the thymidylatesynthase inhibitor AG337 in human colon, breast, and ovarian cancers. AnticancerDrugs 8:876-885, 1997.

11. deBraud F, Munzone E, Nole F, et al: Synergistic activity ofoxaliplatin and 5-fluorouracil in patients with metastatic colorectal cancerwith progressive disease while on or after fluorouracil. Am J Clin Oncol21:279-283, 1998.

12. Extra JM, Espie M, Calvo F, et al: Phase I study ofoxaliplatin in patients with advanced cancer. Cancer Chemother Pharmacol25:299-303, 1990.

13. Caussanel JP, Levi F, Brienza S, et al: Phase I trial of5-day continuous venous infusion of oxaliplatin at circadian rhythm-modulatedrate compared with constant rate. J Natl Cancer Inst 82:1046-1050, 1990.

14. Becouarn Y, Rougier P: Clinical efficacy of oxaliplatinmonotherapy: Phase II trials in advanced colorectal cancer. Semin Oncol 25(supp5):23-31, 1998.

15. Lévi F, Misset JL, Brienza S, et al: A chronopharmacologicphase II clinical trial with fluorouracil, folinic acid, and oxaliplatin usingan ambulatory multichannel programmable pump. High antitumor effectivenessagainst metastatic colorectal cancer. Cancer 69:893-900, 1992.

16. Lévi F, Zidani R, Brienza S, et al: A multicenterevaluation of intensified, ambulatory, chronomodulated chemotherapy withoxaliplatin, fluorouracil, and leucovorin as initial treatment of patients withmetastatic colorectal carcinoma. Cancer 85:2532-2540, 1999.

17. de Gramont A, Figer A, Seymour M, et al: Leucovorin andfluorouracil with or without oxaliplatin as first-line treatment in advancedcolorectal cancer. J Clin Oncol 18:2938-2947, 2000.

18. Aaronson NK, Ahmedzai S, Bergman G, et al: The EuropeanOrganization for Research and Treatment of Cancer QLQ-C30: A quality-of-lifeinstrument for use in international clinical trials in oncology. J Natl CancerInst 85:365-376, 1993.

19. Giacchetti S, Perpoint B, Zidani R, et al: Phase IIImulticenter randomized trial of oxaliplatin added to chronomodulatedfluorouracil-leucovorin as first-line treatment of metastatic colorectal cancer.J Clin Oncol 18:136-147, 2000.

20. Bleiberg H, de Gramont A: Oxaliplatin plus fluorouracil:Clinical experience in patients with advanced colorectal cancer. Semin Oncol25(suppl 5):32-39, 1998.

21. Van Cutsem E, Szanto J, Roth A, et al: Evaluation of theaddition of oxaliplatin to the same Mayo or German 5-FU regimen in advancedrefractory colorectal cancer (ARCRC) (abstract 900). Proc Am Soc Clin Oncol18:234a, 1999.