The Taxanes: Dosing and Scheduling Considerations

ABSTRACT: Optimal dosing and scheduling are among the most important issues being addressed in clinical studies of the taxanes. The results to date indicate that there may not be a single administration schedule that produces optimal antitumor efficacy. Instead, the specific doses of the taxanes relative to each schedule and the overall aggressiveness of the dosing schedule should be considered. There appears to be a threshold taxane dose or concentration below which only negligible antitumor activity is observed, as well as a plateau dose or concentration above which no further antitumor activity occurs. The doses at which both threshold effects and plateauing of dose-response curves occur seem to be inversely proportional to the duration of the administration schedule. For paclitaxel (Taxol), it appears that comparable antitumor effects are achieved with both short (1- and 3-hour) and prolonged (24- and 96-hour) schedules as long as equitoxic dosing regimens are used. The majority of clinical studies with docetaxel have used a somewhat aggressive dosing schedule, 100 mg/m² over 1 hour, which marks the outer edge of the dosing envelope, but nonrandomized trial results suggest a dose-response relationship in the 60- to 100-mg/m² dosing range. [ONCOLOGY 11(Suppl):7-19, 1997]

Introduction

The determination of optimal dosing and scheduling has been an importantobjective during the development of the taxanes. This issue pertains toboth paclitaxel (Taxol) and docetaxel (Taxotere). The clinical developmentof docetaxel has largely involved a single administration schedule (1-hourinfusion) and a narrow dosing range (60 to 100 mg/m², with most studiesusing 100 mg/m² over 1 hour every 3 weeks). The range of paclitaxeldoses and schedules, on the other hand, has been broad (ranging from 135to 250 mg/m² over 1 to 96 hours every 3 weeks).

Impressive antitumor activity has been reported recently for paclitaxelon disparate administration schedules, which leads to the question of whetheran optimal dosing schedule truly exists for the taxanes. Fortunately, theresults of several prospective randomized studies, in addition to retrospectiveanalyses, may shed light on these questions. This review summarizes pertinentpreclinical, pharmacologic, and antitumor results pertaining to optimaltaxane dosing and scheduling in clinical practice.

In Vitro Cytotoxicity Studies

The results of in vitro studies designed to evaluate taxane dose-responserelationships and optimal taxane scheduling have been reviewed previously.[1,2]Many relevant biologic effects in vitro, such as cytotoxicity, formationof microtubule bundles and mitotic asters, increase in tubulin polymermass, stabilization of microtubules against depolymerization, apoptosis,radiosensitization, antiangiogenesis, and inhibition of chemotaxis andmotility, appear to be directly related to the concentration of the taxanes.[3-14]

The taxanes may induce different intracellular effects, depending onthe drug concentration. They inhibit proliferation of cells by inducinga sustained mitotic block at the metaphase/anaphase boundary at concentrationsmuch lower than those required to increase microtubule mass and microtubulebundle formation.[15] Half-maximal inhibition of HeLa cell proliferationand 50% blockade of mitotic metaphase occur at 8 nM of paclitaxel, whereasmicrotubule mass increases half-maximally at 80 nM of paclitaxel, withmaximal effect at 300 nM.[15] At high concentrations, the unique effectsof paclitaxel in increasing microtubule mass and microtubule bundles havebeen associated with growth inhibition.[12,15]

Because these effects have not been noted at the lowest effective concentrations,they could not have accounted for the antiproliferative effects observedat low concentrations. Instead, growth inhibition has been associated withthe formation of an incomplete metaphase plate of chromosomes and an arrangementof spindle microtubules resembling the abnormal organization that occursat low concentrations of the Vinca alkaloids.[16]

Plateau and Threshold Concentrations

As paclitaxel concentrations progressively increase, a plateauing ofdose-response effects has been observed in various cell lines.[12,17-22]In other words, a situation of diminishing returns occurs as paclitaxeland docetaxel concentrations are increased above specific plateau concentrations,the magnitude of which appears to vary between cell lines. The broad clinicalimplication of these results is that there may be a critical plateau concentration,ie, a dose above which toxicity, but not efficacy, increases.

The cumulative results of in vitro studies suggest that the preciseconcentrations at which plateauing occurs depends on the specific treatmentschedule and varies according to tumor type. In addition, there appearto be precise threshold concentrations below which drug effects do notusually occur. Like plateau concentrations, the precise level at whichthreshold effects take place also varies among cell lines. Threshold concentrationstypically are inversely related to the duration of treatment. In essence,these preclinical observations resemble clinical observations to date.

Treatment Duration

For both paclitaxel and docetaxel, treatment duration appears to bethe most critical determinant of in vitro effect. Prolonging the durationof exposure in vitro generally produces much greater cytotoxicity thanincreasing drug concentration.[1,2,9,12,14,17,18, 20,23-25] For example,an 11-fold increase in the duration of paclitaxel exposure was more effectivein increasing the cytotoxic effect of paclitaxel in an LC8A lymphoma cellline than was a 100-fold increase in paclitaxel concentration.[12] Interestingly,this effect appears to be much more pronounced in taxane-resistant celllines.[24,26-29] The taxane concentrations at which cell survival curvesplateau tend to decrease as the treatment durations are prolonged.

For paclitaxel, and probably docetaxel, the effects of increasing microtubulemass are maximal at drug concentrations that are equimolar with tubulinor when the stoichiometry approaches 1 M of paclitaxel per 1 M of polymerizedtubulin dimer.[4,30-33] The binding of paclitaxel to polymerized tubulinis reversible with a binding constant of approximately 0.9 µM. Docetaxel,which most likely shares the same tubulin binding site as paclitaxel, hasa 1.9-fold higher effective affinity for the site.[4] The assembly of guanosinediphosphate- or guanosine triphosphate-tubulin induced by docetaxel alsoproceeds with a critical protein concentration that is 2.1-fold lower thanthat of paclitaxel.[4] In addition, comparative in vitro cellular pharmacologicstudies have demonstrated that higher intracellular levels of docetaxelin P388 murine leukemia cells may also be attributed, in part, to its threefoldslower efflux rate.[4]

These differences may explain the varying cytotoxic potencies of thetaxanes, with median inhibitory concentrations generally much lower fordocetaxel.[4,21,22,33-36] The relative potencies may not necessarily translateinto a greater therapeutic index for docetaxel, since greater potency mayalso result in more severe toxicity at identical drug concentrations invitro. In addition, the taxanes may
not be completely cross-resistant, although differences in potency mayconfound the interpretation of both preclinical and clinical studies regardingcross-resistance.

These schedule-dependent effects have also been documented in studiesdesigned to determine the
in vitro interactions of the taxanes with ionizing radiation.[9] In moststudies, the radiopotentiating effects of the taxanes have been directlyrelated to the duration of taxane exposure prior to radiation.[9] In oneseries of studies involving lung cancer cells, a radiosensitizing effectcould not be demonstrated for treatment durations of less than 6 hoursat any concentration of paclitaxel.[37]

Preclinical Studies

In early studies performed by the National Cancer Institute, paclitaxelwas administered as a suspension and antitumor evaluations were limitedto studies using intraperitoneally (IP) implanted tumors treated with IPdrug administration or human tumor xenografts implanted in the subrenalcapsule and treated subcutaneously.[38] In mice-bearing IP-implanted P388leukemia, paclitaxel administered every 3 hours for three doses (pharmacologicallysimulating a 24-hour infusion schedule) was more effective than other schedules,including those with multiple-drug treatments on days 1, 5, and 9, or dailytreatment for 9 consecutive days.[38] However, these studies were constraineddue to the limited solubility of paclitaxel in the formulation vehiclesused at that time, thereby precluding the design of proper comparativestudies of prolonged schedules and single-dosing schedules.

Lung Cancer Model

When studies designed to evaluate different schedules were later performedby Bristol-Myers Squibb in the M109 lung cancer model using polyoxyethylatedcastor oil (Cremophor EL) or polysorbate 80 formulations, no schedule-dependentdifferences were observed.[39] However, neither the optimal schedule demonstratedin the P388 leukemia studies (every 3 hours × 8 doses) nor prolonged(eg, 3 24-hour) infusion schedules were evaluated.

The schedule-finding studies in the M109 model indicated that both daily× 5 and daily × 7 schedules were superior to multiple dailydosing for longer intervals (2 or 3 days between injections). Significantly,maximal antitumor activity was achieved at doses that were substantiallylower than the maximum tolerated dose (MTD). This was particularly truefor the daily × 7-day schedule, in which dosing at the MTD did notresult in any therapeutic advantage over an equally effective, but lesstoxic, lower dose.[40] These results are consistent with the plateau effectsnoted in the dose-response curves from a variety of cell lines.

Other Models

Nevertheless, dose-response effects have been observed in many otherpreclinical in vitro models. Progressive reductions in vertebral metastaseswere noted in combined immunodeficient mice inoculated with PC-3 ML prostatecancer cells that were previously incubated with increasing paclitaxelconcentrations from 0.1 to 1.0 µM [41] and in mice bearing bladdercancer.[42]

Dose-toxicity relationships have been especially profound. Substantiallygreater toxicity in both rapidly proliferating (lymphoid, myeloid, gastrointestinal)and nonproliferating (peripheral nerve) tissues has been observed on almostall schedules in mice, rats, and dogs.[43-45] The schedule's effects aremore profound in that lower total doses are typically required to induceequivalent toxic effects in animals treated with more intermittent dosingschedules or treatment over more prolonged durations.[38,43]

It is difficult to compare paclitaxel and docetaxel with respect todose and schedule dependency in preclinical studies in animals due to differencesin tumor models, dosing schedules, and the proximity of treatment dosesto the MTD. Nevertheless, results of limited studies with docetaxel haveindicated clear dose-response relationships, particularly with short- andsingle-dosing schedules. Although one may conclude that the type of administrationschedule appears to have minimal impact on docetaxel's antitumor activity,it should be noted that only limited studies with prolonged schedules havebeen performed.[5,46,47]

Pharmacology

Comprehensive reviews of the clinical pharmacology of paclitaxel anddocetaxel have been published previously,[47-51] and only limited pharmacologicaspects relevant to dosing and scheduling issues will be discussed here.Paclitaxel and docetaxel share the following pharmacologic characteristics:large volumes of distribution, rapid and sustained uptake by most tissues,long elimination half-lives, and significant hepatic disposition. The pertinentpharmacokinetic parameters of both agents are summarized in Table1.

Predictions regarding the potential success of various taxane dosesand schedules are often based on whether biologically active drug concentrationscan be achieved and maintained in human plasma. Such extrapolations mayhave several pitfalls, particularly in situations in which drug concentrationsachieved in plasma and peripheral tissues (or tumors) may be disparate.For both paclitaxel and docetaxel, plasma concentrations achieved withalmost any dosing schedule are capable of inducing relevant biologic andcytotoxic effects in vitro (paclitaxel: more than 0.05 µM with 96-hourinfusions, more than 0.3 µM with 24-hour infusions, and more than5 µM with 1- to 3-hour infusions; docetaxel: more than 3 µMwith 1-hour infusions).

Despite extensive plasma protein binding, both paclitaxel and docetaxelare readily cleared from plasma. More important, the volumes of distributionof the taxanes are very large, which is most likely attributable to avidand ubiquitous drug binding to tubulin. Tissue distribution studies inanimals using radiolabeled drug have revealed high tissue/plasma concentrationratios in virtually all tissues except brain and testes, which are generallyconsidered tumor sanctuary sites.[52-55]

In fact, docetaxel concentrations in mice have been demonstrated tobe substantially higher in the implanted tumor tissue than in plasma.[6]Not only are high taxane concentrations achieved in almost all peripheraltissues, but biologically relevant concentrations are maintained for relativelylong periods.[6] In both animals and humans, the results of radiolabeleddrug distribution studies suggest that there is substantial sequestrationof the taxanes in peripheral tissues.[47-55] Approximately 20% of an administereddose is recovered as either parent compound or metabolites from bile andfeces collected for 24 hours after treatment. Most of the total administeredradioactivity is recovered from feces collected for 1 week following treatment.[52-58]Renal clearance is insignificant for both taxanes, whereas hepatic P450mixed-function oxidative metabolism, biliary excretion, fecal elimination,and tissue binding are responsible for the bulk of systemic clearance.[47-58]

These pharmacologic characteristics, particularly wide total body distribution,avid tissue binding, and tissue sequestration, indicate that plasma concentrationsmay underestimate drug concentrations and pharmacologic exposure in peripheraltissues and tumors. This behavior also suggests that short infusion schedulesmay be as effective as prolonged schedules in saturating peripheral tissuesand tumors. The maintenance of effective tissue saturation may also dependon other factors, including the duration of plasma concentrations maintainedabove a critical threshold, the duration of the infusion, and the totaldose of drug, particularly in situations in which tissue binding may notbe avid.

Linear vs Nonlinear Pharmacokinetics

The pharmacokinetic behavior of paclitaxel appeared linear in earlystudies of prolonged administration schedules. However, the results ofpharmacokinetic studies accompanying a National Cancer Institute of Canada-ClinicalTrials Group (NCIC-CTG) pivotal bifactoral randomized trial (BMS 016 orOv.9) demonstrated that the pharmacokinetic behavior of paclitaxel is nonlinear.[48,49]The NCIC-CTG study observed the effects of paclitaxel (135 vs 175 mg/m²,3 vs 24 hours with premedication) in women with recurrent or refractorycarcinoma of the ovary. Pharmacokinetic data from subsequent studies thatwere performed in both children and adults
have confirmed these results.[48,49]

As with all drugs with nonlinear pharmacokinetic profiles, nonlinearor saturable behavior is accentuated on shorter infusion schedules. Bothplasma concentrations and drug exposure increase disproportionately withincreasing doses on shorter schedules. In addition, the pharmacokineticsof drugs like paclitaxel, which truly exhibit a nonlinear pharmacokineticbehavior at higher plasma concentrations, are more likely to be appearlinear with prolonged infusion schedules that yield low plasma concentrations.

When plasma levels are much lower than Km (the Michaelis-Mentenconstant), elimination or distribution processes are not saturated andpharmacokinetics appear linear (first-order). Conversely, nonlinear (zero-order)pharmacokinetics become more apparent with shorter infusion schedules,which result in higher plasma concentrations that approach or exceed theKm of the saturable processes.

Pharmacokinetic modeling of paclitaxel plasma concentration data indicatesthat both saturable distribution and elimination processes account forpaclitaxel's nonlinear pharmacokinetic behavior.[48, 49] Physiologically,nonlinear drug elimination is most likely due to saturable hepatic P450metabolic processes and/or excretion, which accounts for a principal componentof drug disposition. Saturable drug distribution, on the other hand, ismuch more difficult to explain.

Two pharmacokinetic models have been used successfully to describe nonlineardrug distribution. One model assumes that the drug transfer process intoperipheral tissues is saturable, resulting in Km distribution kinetics.An alternate, more physiologic model assumes that there is a limited (therefore,saturable) number of drug-binding sites in the peripheral compartment.This model displays a rate constant for transfer of drug to the peripheralcompartment, varying directly with the number of unoccupied binding sitesin the peripheral compartment.[E.K. Rowinsky, MD, unpublished results]For paclitaxel, the limited number of binding sites may represent bindingsites on beta-tubulin.

Clinical Implications

Paclitaxel's nonlinear pharmacokinetic profile may have several clinicalimplications. For example, dose escalations, especially on shorter administrationschedules, may result in disproportionate increases in both area underthe concentration-time curve (AUC) and peak plasma concentration (Cpeak),as well as disproportionate increases in toxicity. Dose reductions mayhave the opposite effect, resulting in disproportionate decreases in AUCand/or Cpeak, thereby possibly decreasing antitumor activity. In addition,these models predict that tissue sites are effectively saturated at relativelylow paclitaxel concentrations (achieved with paclitaxel doses less than175 mg/m² on a 3-hour schedule), whereas elimination processes areeffectively saturated at higher doses (achieved with paclitaxel doses (3175 mg/m² on a 3-hour schedule).

To characterize tissue saturation as a function of paclitaxel dose,model simulations have been performed using actual plasma concentrationdata and both types of tissue distribution models, taking saturable eliminationprocesses into account.[E.K. Rowinsky, md, unpublished results] The simulationshave demonstrated that the peak drug contents in tissues do not changesignificantly when paclitaxel doses are increased from 135 to 250 mg/m²on both 3- and 24-hour schedules. In addition, these simulations indicatethat tissue saturation is greater with shorter administration schedulesfor all paclitaxel doses, and the rate that tissues become saturated isalso greater with shorter infusion schedules.

Based on this model, peak drug content is approximately 70% of the theoreticalmaximum at a dose of 135 mg/m², 85% at a dose of 175 mg/m², andgreater than 90% at a dose of 250 mg/m² when paclitaxel is administeredover 3 hours. Respective tissue saturation values are approximately 35%,45%, and 55% when paclitaxel is administered over 24 hours. These resultssuggest that increasing drug content in peripheral tissues and tumors byincreasing paclitaxel dose or dose intensity within a clinically relevantdosing range may be tantamount to a situation of diminishing limiting returnsas tissue binding sites become progressively saturated.

It should be pointed out, however, that the precise levels of tissuesaturation that result in maximal cytocidal or toxicologic effects arenot known. There may also be other important determinants of drug effect,such as the duration that any specific degree of tissue saturation is maintained.In other words, a maximal effect may occur with any degree of tissue saturation.In addition, it may be important to consider the duration that any specificdegree of tissue saturation is maintained. Therefore, direct comparisonsbetween prolonged and short treatment schedules with respect to the effectof different degrees of tissue saturation on outcome (ie, cytotoxicity,toxicity) cannot be adequately performed until other critical determinantsof effect are characterized. However, the notion of a "threshold concentration"or a "threshold dose" due to saturable pharmacokinetic processesmay account for the plateauing dose- or concentration-response relationshipsthat have been observed in vitro and in clinical practice.

Dose and Schedulein Clinical Trials

A broad range of dosing schedules has been evaluated for paclitaxel,whereas most clinical trials with docetaxel have used a single-dosing schedule(100 mg/m² over 1 hour). Evaluations of paclitaxel were initiallylimited to the 24-hour schedule because of the high incidence of majorhypersensitivity reactions in patients receiving short schedules withoutpremedication, and the 24-hour schedule was the first to be approved forwomen with recurrent or refractory ovarian cancer. Most available informationabout paclitaxel in various disease settings thus pertains to the 24-hourschedule.

However, prominent antitumor activity has also been observed with 1-and 3-hour schedules of paclitaxel in women with recurrent or refractorybreast and ovarian cancers. This activity has also been apparent in bothchemotherapy-naive and previously treated patients with non-small-celllung cancer and other tumor types. The results of the NCIC-CTG's pivotalOv.9 (or BMS 016) trial that addressed the relative merits of administrationschedule and premedication led to the regulatory approval of the 3-hourschedule in women with recurrent or refractory ovarian cancer, and it stimulatedbroader interest in exploring alternate doses and schedules.[59]

The equivalence of both the 3- and 24-hour schedules of paclitaxel inthis setting, in which the overall basal response rate is low and largenumbers of patients are required to detect small differences, does notnecessarily imply that both schedules are equivalent in other tumor typesand settings, particularly when equivalent doses are used. Intriguing results,consisting of some of the highest response rates achieved with paclitaxel,have been noted on more prolonged schedules (particularly the 96-hour infusion)in women with recurrent or refractory breast cancer.[60,61] Impressiveantitumor activity has been observed in patients with both chemotherapy-naiveand previously treated non-small-cell lung cancer treated with paclitaxelon an even shorter 1-hour schedule.[62] However, it should also be notedthat the impressive results with both schedules occurred with doses thatresulted in a level of neutropenia comparable to that observed in initialpivotal trials using the 24-hour schedule.

Paclitaxel at its MTD

A concern during the early development of paclitaxel was that its limitedsupply would preclude the scope and number of phase II evaluations.[63,64]Because heavily pretreated women with advanced ovarian cancer were ableto tolerate substantially lower paclitaxel doses than untreated or minimallypretreated patients, investigators initially opted to perform most phaseII evaluations of paclitaxel at its MTD with granulocyte colony-stimulatingfactor (G-CSF [Neupogen]) support, as well as limited phase II trials atlower paclitaxel doses without G-CSF. In several clinical situations, suchas in untreated or minimally pretreated women with advanced breast cancer,phase II studies using high doses of paclitaxel and occasional G-CSF wereperformed from the outset.

In essence, the decision to evaluate paclitaxel at its MTD with G-CSFsupport was based on the possibility that the demonstration of unimpressiveactivity at higher paclitaxel doses would obviate the need to perform additionalphase II trials of paclitaxel at lower doses. It was similarly felt thatthe demonstration of negligible activity at submaximal doses would notconclusively prove that paclitaxel was inactive across its entire dosingspectrum, and additional studies evaluating higher drug doses might berequired.

In studies of paclitaxel at its MTD, negligible activity was demonstratedin patients with advanced melanoma and colorectal, renal, and gastric cancers,[63,64]which argued strongly against performing additional phase II trials oflower doses in these tumor types. A similar reasoning was subsequentlyused during the development of docetaxel in that broad phase II evaluationswere performed using an MTD schedule (100 mg/m² over 1 hour).[51,65]This approach results in unequivocally negative conclusions regarding activitywhen high-dose phase II studies are negative. The potential benefits ofthis approach are offset by the dilemma that often arises when significantantitumor activity is observed in nonrandomized phase II studies employingmaximal drug doses with or without hematopoietic colony-stimulating factorsupport. In this situation-as was the case for paclitaxel in phase II trialsin lymphoma and carcinomas of the breast, lung, cervix, and head and neck-furtherclinical investigations may focus solely on the most intensive dosing schedule,in which lower, less toxic doses and/or dose intensity may also producesubstantial antitumor activity.

Clinical Doses

In clinical practice, paclitaxel is usually administered at a dose of175 mg/m² over 3 hours or 135 to 175 mg/m² over 24 hours every3 weeks. These dosing schedules may result in optimal therapeutic indicesin many disease settings, such as in palliating patients with refractoryor recurrent advanced ovarian and breast cancers. It may, however, be moreappropriate to use alternate dosing schedules in other tumor types andsettings, particularly in less heavily pretreated patients and when prolongationof survival is a tangible objective.

A broad range of paclitaxel doses has been evaluated on almost all schedules,and the most appropriate dose for any schedule is currently being studiedin randomized trials in ovarian, breast, and non-small-cell lung cancers.Although several administration schedules for docetaxel have been studiedin phase I evaluations, the agent is most commonly administered at a doseof 100 mg/m² as a 1-hour infusion every 3 weeks. Lower doses (60 to75 mg/m²) on a 1-hour schedule may be associated with a lower incidenceof both hematologic and nonhematologic toxicities; however, the relativetherapeutic advantages of high vs low doses are not clear.

Ovarian Cancer

In the initial five phase II studies of paclitaxel on a 24-hour schedulein women with recurrent or refractory ovarian cancer, which were used forregistration of the agent in the United States, doses ranged from 110 to300 mg/m².[66-70] Individual study reports used somewhat differentcriteria to define patient eligibility and evaluability for response. Inthese reports, response rates were seemingly higher (36% to 48%) in trialsevaluating higher paclitaxel doses (170 to 300 mg/m²) and dose intensitythan in trials in which most patients were treated with lower paclitaxeldoses (response rates, 30% to 37%; 110 to 175 mg/m²) or higher doses,albeit administered in a
less dose-intensive regimen (response rate, 20%; 250 mg/m² with treatmentdelays).[66-70]

G-CSF support was also used from the outset to ameliorate severe neutropeniaand maintain dose intensity in trials that employed high paclitaxel dosesor dose intensity.[68-70] Nonhematologic toxicities, particularly neurotoxicity,occurred more frequently than in those trials that evaluated
lower paclitaxel doses or dose intensity.[66,67]

At first glance, such observations may suggest that higher paclitaxeldoses and dose intensity are optimal in this and other settings. For example,using the correlative methods of Hrynick and Levin, Reed et al retrospectivelyanalyzed the relationships between paclitaxel dose intensity and antitumoractivity in women with advanced ovarian and breast cancers.[71] They demonstratedstrong positive relationships in both settings (P = .022 and .004, respectively).It should be stressed that these trials were nonrandomized and incorporateddiverse patient populations with respect to pertinent demographic and prognosticvariables.

Results of a Meta-analysis

To more appropriately evaluate the effects of paclitaxel dose and doseintensity on the antitumor activity and survival of patients with refractoryor recurrent advanced ovarian cancer, a pooled analysis of individual patientdata (meta-analysis) was performed.[72] The database was derived from theinitial five studies that were used for registration of paclitaxel in theUnited States for recurrent or refractory disease. It audited demographic,categorical response, and successive tumor measurement data from 191 ovariancancer patients who were treated with paclitaxel on a 24-hour scheduleevery 3 weeks.

In this analysis, the probability of achieving a partial or completeresponse was not related to the average paclitaxel dose (odds ratio, 1.0per 10 mg/m²; P = .60). In fact, the probability of responding appearedto decrease with increasing dose intensity (odds ratio, -0.77 per 10 mg/m²/wk;P = .06). A strong negative relationship between the maximum percent reductionin
tumor size and dose intensity was also apparent (odds ratio, -6.1% per10 mg/m²/wk; P = .007). Not only was a negative effect demonstratedin the logistic regression analysis of the data from all studies but alsoa negative effect was found in each individual study.

With respect to overall survival, the analysis suggested a negativerelationship between average paclitaxel dose and survival (hazard ratio,1.06 per 10 mg/m²; P = .0001). There was also a strong negative relationshipbetween dose intensity and survival (hazard ratio, 1.3 per 10 mg/m²/wk;P = .0001). Similar relationships were demonstrated between paclitaxeldose, dose intensity, and progression-free survival. These relationshipswere minimally affected when the analyses were controlled for individualstudy, performance status, number of prior regimens, platinum sensitivity,and response to prior therapy. These results indicate no clear benefitof increasing paclitaxel doses above 135 mg/m² when the drug is administeredas a 24-hour infusion every 3 weeks to women with recurrent or refractoryadvanced ovarian cancer. This is supported by the known saturable pharmacologicbehavior of paclitaxel.

Possible Explanations

There are several possible explanations that may account for the lackof an effect of higher paclitaxel doses on both the categorical responseand percentage reduction in measurable disease (such as tissue saturation,as discussed previously). It is difficult, however, to propose reasonablystrong biologic or pharmacologic explanations for why antineoplastic activitymight decrease with higher paclitaxel dose intensity.

Liebmann et al reported a negative relationship between paclitaxel concentrationand cytotoxicity in a panel of eight human tumor cell lines that were treatedfor 24 hours.[20] Increasing paclitaxel concentrations from 2 to 20 nM/Lsharply increased cytotoxicity. No additional cytotoxicity occurred withpaclitaxel concentrations above 50 nM/L, and treatment with very high drugconcentrations (more than 10 µM) resulted in even less cytotoxicity.Furthermore, the principal constituent of paclitaxel's clinical formulationvehicle, polyoxyethylated castor oil, at a concentration of 0.135%, antagonizedthe cytotoxic effects of paclitaxel. Even in light of these experimentaldata, the potentially negative effects of paclitaxel dose intensity onantitumor activity and survival in the meta-analysis must be interpretedwith considerable caution and awaits confirmation in larger prospectivestudies.

These results argue strongly against the possibility that increasingpaclitaxel dose above 135 mg/m² on a 24-hour schedule or dose intensitymight enhance clinical antitumor activity. The analysis suggests that theoptimal paclitaxel dose intensity for similar patients is at the lowerend of the range of dose intensity used in these investigations, ie, 45mg/m²/wk (or 135 mg/m² every 3 weeks) as a 24-hour infusion.At this juncture, however, it is not known whether these results can begeneralized to describe paclitaxel dose and dose-intensity relationshipsusing other dosing schedules (eg, 3- or 96-hour schedules) and in otherdisease settings.

Two Randomized Trials

To date, the effect of paclitaxel dose on clinical response in womenwith refractory or recurrent ovarian cancer has been evaluated prospectivelyin two randomized trials.[59,73] The salient features of randomized clinicaltrials of the taxanes that focused on dosing and scheduling issues in ovarianand other neoplasms are listed in Table 2.

In Ov.9 (BMS 016), the NCIC-CTG evaluated the effects of two differentpaclitaxel doses (135 vs 175 mg/m²) and two different schedules (24vs 3 hours) on both response and toxicity.[59] With respect to the dosingissue, progression-free survival was significantly, albeit not profoundly,longer in the high-dose arm (19 vs 14 weeks, P = .02), but response ratesand survival were similar. Nevertheless, the two paclitaxel doses werenot very disparate, and the study design and patient numbers precludedcomparisons between each of the four individual treatment arms, so thatpatients treated with both schedules were analyzed together. It is possiblethe patients receiving paclitaxel on the 3-hour schedule, and not the 24-hourschedule, contributed to the significant difference between the two doses.

The dose-response issue has also been assessed in an intergroup studyin which a similar group of patients were treated with 24-hour infusionsof paclitaxel at one of three doses: 135, 175, and 250 mg/m² plusG-CSF. However, the lowest dose arm was terminated after the regulatoryapproval of paclitaxel, which resulted in reduced patient accrual.[73]A preliminary analysis of the results of the study indicates only modestdifferences in response rates. Response rates were 36% vs 28% in the 250-mg/m²-plus-G-CSFand 175 mg/m²-arms, respectively, but there were no differences ineither progression-free or overall survival.[73]

Based on the results of nonrandomized trials, meta-analyses, and thelimited randomized trials to date, there is no compelling reason to administerpaclitaxel as a 24-hour infusion in
women with recurrent or refractory ovarian cancer at doses above 135 to175 mg/m². The cumulative results indicate that paclitaxel doses above135 mg/m² result in either little or no further benefit with respectto obtaining categorical responses and certainly do not prolong disease-freeand overall survival times.

In the Ov.9 study, myelosuppression was substantially more severe inpatients receiving identical doses of paclitaxel over 24 hours comparedwith 3 hours, probably due to longer drug exposure above a critical thresholdconcentration. Toxicologically, the results suggest that the precise paclitaxeldose capable of inducing any given effect depends on the specific administrationschedule.

These findings preclude extrapolation of the relationships between doseand effect (ie, toxicity, disease activity) from one dosing schedule toanother. The limited availability of comparable data with shorter (eg,3-hour) schedules in this disease setting, particularly at paclitaxel dosesabove 175 mg/m², preclude making similar recommendations. Study resultsin this and other disease settings nonetheless indicate that paclitaxeldoses below 175 mg/m² on short schedules may be suboptimal. For docetaxelin this disease setting, there are sufficient impressive results availableonly for 100 mg/m² over a 1-hour dosing schedule [51,65].

Breast Cancer

As in the case of ovarian cancer, the determination of optimal taxanedosing and scheduling has been the principal goal of several clinical trialsin breast cancer. To date, it appears that dose may be a critical determinantof response for both paclitaxel and docetaxel administered on short schedules(eg, paclitaxel over 3 hours and docetaxel over 1 hour).

Paclitaxel Dosing

The results of a randomized trial (BMS 48) of paclitaxel doses of 135vs 175 mg/m² on a 3-hour schedule in women with metastatic breastcancer who had previously been treated with adjuvant chemotherapy only,chemotherapy for metastatic disease only, or chemotherapy in both adjuvantand metastatic settings, revealed no differences in response rates (29%[high dose] vs 22% [low dose]) or in median survival (11.7 [high dose]vs 10.5 months [low dose]). Progression-free survival was longer, however,with the higher dose (4.2 vs 3 months; P = .02).[74] Severe neutropeniawas not common with either dose; the incidences of grade 4 neutropeniawere 19% and 28% in patients treated with paclitaxel 135 and 175 mg/m²,respectively. Although this trial used nondisparate paclitaxel doses, theseresults indicate that neutropenia and antitumor activity may be congruentindices of drug effect.

These results led to the regulatory approval of paclitaxel 175 mg/m²(3-hour schedule) in women with metastatic breast cancer after failureof combination chemotherapy or relapse within 6 months of adjuvant therapy.The Cancer and Leukemia Group B is currently evaluating whether higherpaclitaxel doses achieve greater activity in women with metastatic breastcancer. In this trial (CALGB 9342), patients are being randomized to treatmentwith paclitaxel doses of 175, 210, or 250 mg/m² on a 3-hour schedule.

Docetaxel Dosing

There has been a paucity of studies addressing optimal dosing with docetaxel.This is because the agent has been evaluated almost exclusively at a doseof 100 mg/m² over 1 hour, which induces severe neutropenia in mostpatients. Although substantial activity with less toxicity has been noted
at lower docetaxel doses (60 to 75 mg/m²), the results of nonrandomizedstudies indicate that there may be a positive dose-response relationshipwith docetaxel given on a 1-hour schedule at doses ranging from 60 to 100mg/m². In women with metastatic breast cancer who have progressedon anthracycline-based chemotherapy, response rates with docetaxel on a1-hour schedule have averaged 35% and 47% at doses of 60 and 100 mg/m²,respectively.[51,65,75-77] Respective response rates have averaged approximately55% and 48% in patients who have not received chemotherapy in the metastaticsetting.[78-81] A multicenter randomized trial, in which women with metastaticbreast cancer are receiving docetaxel doses of either 75 or 100 mg/m²on a 1-hour schedule, is ongoing.

Paclitaxel Scheduling

Optimal paclitaxel scheduling was studied in a randomized trial (BMS71) in which women with metastatic breast cancer were treated with paclitaxel175 mg/m² infused over either 3 or 24 hours.[82,83] Women who hadreceived no prior chemotherapy, adjuvant chemotherapy only, or chemo- therapyfor metastatic disease with or without prior adjuvant therapy were eligible.Paclitaxel dose escalations to 200 and 225 mg/m², as well as dosereductions to 135 mg/m², were permitted, depending on the degree ofmyelosuppression during the previous course.

There were no differences in cumulative response rates (29% vs 31%),median progression-free survival (3.5 vs 4.6 months), or survival (9.8vs 13.4 months) between the 3- and 24-hour groups, respectively. However,it was apparent that the starting doses were not equitoxic, which is demonstratedby the fact that 30% and 79% of women in the 3- and 24-hour groups, respectively,developed grade 4 neutropenia. In the 3-hour group, paclitaxel doses wereincreased on one or two occasions in 65% of patients, whereas dose reductionswere required in only 5% of patients. In contrast, 34% of patients in the24-hour group required dose reductions, and paclitaxel doses were escalatedin 36% of patients.

It is also important to note that responses rates in patients receivingpaclitaxel over 3- and 24-hour schedules were more disparate in women whohad no prior therapy (34% vs 57%) compared with those who had adjuvanttherapy only (36% vs 40%) or chemotherapy in both adjuvant and metastaticsettings (24% vs 22%). These results suggest that the more aggressive 24-hourschedule does not result in a significant improvement in outcome in thepalliative setting. However, more aggressive dosing schedules should beconsidered for patients with earlier stages of disease. This does not necessarilyimply that the 24-hour or more prolonged schedules are superior and shouldbe used exclusively. Instead, alternate aggressive dosing schedules, asgauged by their potential to induce a degree of toxicity similar to paclitaxel175 mg/m² over 24 hours, might suffice. For example, paclitaxel dosesranging from 200 to 250 mg/m² on a 3-hour schedule appear to be equivalenttoxicologically to paclitaxel, 175 mg/m² on a 24-hour schedule.

Sigmoid curves that describe the relationships between paclitaxel effect(eg, the percentage decrement in absolute neutrophil counts) as a functionof the duration of drug exposure above a critical threshold concentrationon 1-, 3-, and 24-hour paclitaxel schedules are depicted in Figure1. It should be noted that the magnitude of the effects induced byall three dosing schedules converge. For the percentage decrements in neutrophilcounts, convergence occurs in the 225- to 250-mg/m² dosing range.

It seems reasonable that clinical evaluations of optimal taxane schedulingshould be performed using equitoxic regimens (ie, 1- to 24-hour schedulesusing doses at or near the point of convergence). The design of an ongoingNational Surgical Adjuvant Breast and Bowel Project study (NSABP B26) maybe better suited than BMS 71 to discern whether an optimal paclitaxel scheduletruly exists. In this study, women with metastatic or locally advancedbreast cancer are being randomized to treatment with paclitaxel, 250 mg/m²given over either 3 or 24 hours with G-CSF. These dosing schedules aremuch more likely to be equitoxic compared with those employed in BMS 71,minimizing potential confounding variables.

Prolonged Paclitaxel Schedules

Much attention has been paid to more prolonged (96-hour) paclitaxeldosing schedules since Wilson et al reported an impressive response rateof 48% in women with advanced breast cancer who had either progressed duringor following treatment with anthracycline-based chemotherapy.[60] Mostpatients, however, were treated with paclitaxel at its MTD on this prolongedschedule (140 mg/m² over 96 hours), and patients often required G-CSFsupport.

Building on this experience, Seidman et al demonstrated that 7 (25%)of 28 women with metastatic breast cancer who developed recurrent or progressivedisease following treatment with short schedules of either paclitaxel ordocetaxel responded to paclitaxel, 140 mg/m² over 96 hours with G-CSFsupport, if necessary.[61] Both hematologic and nonhematologic toxicitieswere substantial in this study, and the range of taxane doses that theresponding patients had previously received was not reported. However,it is likely that most patients received prior treatment with paclitaxel,175 mg/m² over 3 hours, a dosing schedule that produces substantiallyless toxicity than does 140 mg/m² over 96 hours.

Although these results are provocative, they do not necessarily indicatethat the 96-hour schedule is superior to shorter taxane schedules, andit is likely that the aggressiveness of the dosing schedule itself mayhave accounted for this impressive activity. Indeed, such an effect mighthave been achieved with shorter taxane schedules, albeit at higher doses.Similar caution should be used in interpreting the preliminary reportsof responses in women with heavily pretreated metastatic breast cancerwho are treated with docetaxel on an aggressive dosing schedule (eg, 100mg/m² over 1 hour) after developing progressive disease during orafter treatment with less aggressive taxane dosing schedules (eg, paclitaxel,135 to 175 mg/m² over 3 hours).[84]

Ongoing Trials

Clinical trials designed to compare the intrinsic antitumor efficacyof the various taxanes should use equitoxic regimens. An ongoing phaseIII multicenter trial (NCI T93-0165) being conducted at Memorial Sloan-KetteringCancer Center, the M. D. Anderson Cancer Center, Swedish Tumor Institute,and British Columbia Cancer Agency is more appropriately designed to addressthe scheduling issue. In this trial, women with metastatic breast cancerare randomized to treatment with equitoxic paclitaxel dosing schedulesat two extremes, 140 mg/m² over 96 hours and 250 mg/m² over 3hours.

In contrast, an ongoing randomized phase III trial (RPR 56976-311) thatis designed to determine whether paclitaxel or docetaxel possesses superioractivity in women with metastatic breast cancer who have received priorchemotherapy in the metastatic setting does not use equitoxic regimens.In this study, patients are being randomized to treatment with either docetaxel,100 mg/m² on an 1-hour schedule, or paclitaxel, 175 mg/m² ona 3-hour schedule. The paclitaxel dosing schedule is substantially lessaggressive in terms of its potential to induce myelosuppression, and possibly,its antitumor activity; therefore, this may be an unfair comparison.

Lung Cancer

Early phase II trials of paclitaxel and docetaxel in patients with bothnon-small-cell and small-cell lung cancers, as well as in other tumor types,were performed using relatively aggressive dosing regimens (paclitaxel,200 to 250 mg/m² over 24 hours, and docetaxel, 100 mg/m² over1 hour). There have been only limited attempts to rigorously address questionsregarding optimal dosing and scheduling using randomized trial designsin these tumor types. For both paclitaxel and docetaxel, comparisons ofthe results of phase II studies involving patients with similar relevantdemographic characteristics suggest that there may be dose-response relationshipsthat parallel dose-toxicity relationships, particularly on short taxaneschedules.

For example, response rates in chemotherapy-naive patients with non-small-celllung cancer treated with docetaxel, 60, 75, and 100 mg/m² on a 1-hourschedule, were 23%, 25%, and 31%, respectively.[85] For paclitaxel givenon a 1-hour schedule to both previously treated and chemotherapy-naivepatients with non-small-cell lung cancer, negligible toxicity and anti-tumoractivity (2 responses in 17 patients [12%]) were reported at the 135-mg/m²dose level. A dose of 200 mg/m² resulted in much greater toxicityand antitumor activity (11 responses in 36 patients [31%]).[62]

A major concern during the development of the cisplatin (Platinol)/paclitaxelregimen was that the MTD of paclitaxel on a 24-hour schedule in combinationwith cisplatin (135 mg/m²) was significantly lower than the paclitaxeldose (250 mg/m²) that was determined to be active in early phase IIstudies in untreated patients with non-small-cell lung cancer. Since neutropeniawas the principal toxicity of paclitaxel combined with cisplatin, phaseI studies of the regimen subsequently focused on using G-CSF to enablefurther dose escalation of paclitaxel combined with cisplatin. In thesetrials, peripheral neurotoxicity precluded repetitive administration ofpaclitaxel doses above 250 mg/m² (day 1) on a 24-hour schedule followedby cisplatin 75 mg/m² (day 2) and G-CSF.

Randomized Trial

In view of the acceptable toxicity profile demonstrated for cisplatin(75 mg/m²) combined with higher doses of paclitaxel (250 mg/m²)plus G-CSF, the Eastern Cooperative Oncology Group (ECOG 5592) conducteda randomized trial of chemotherapy-naive patients with metastatic non-small-celllung cancer. Patients received either standard therapy, consisting of etoposide(VePesid), 100 mg/m² on days 1 to 3, and cisplatin, 75 mg/m²on day 1; or cisplatin, 75 mg/m², combined with either low doses ofpaclitaxel, 135 mg/m² on a 24-hour schedule, or high doses of paclitaxel,250 mg/m² 24-hour schedule, plus G-CSF.[86]

Response rates in both the high- and low-dose paclitaxel arms (26.5%and 32.1%, respectively) were superior (P less than .001) to the rate inthe etoposide arm (12%). Median survival times with high- and low-dosepaclitaxel (9.56 and 9.99 months, respectively) also were significantlylonger (P less than .001) than with etoposide (7.69 months). Despite similarrates of severe neutropenia, fever, and infection in both paclitaxel arms,the overall impact of high-dose paclitaxel with G-CSF support in this diseasesetting was negligible.

In accompanying pharmacodynamic studies, steady-state paclitaxel concentrations(Css) in plasma were measured in courses 1 and 2 in both paclitaxel arms,and Css was related to outcome.[87] Although there was a significant differencein Css between the low- and high-dose paclitaxel
arms (mean ± SD, 0.35 ± 0.16 µM vs 0.94 ± 0.50µM; P less than .001), no significant pharmacodynamic relationshipswere evident between Css and both response and time to disease progression.These results indicated that neither response nor time to disease progressionis influenced by either paclitaxel Css or dose in chemotherapy-naive patientswith non-small-cell lung cancer who are treated with paclitaxel doses rangingfrom 135 to 250 mg/m² (24-hour schedule) followed by cisplatin.

Collectively, these results in non-small-cell lung cancer indicate thatthe relationship between taxane dose and response plateaus at lower taxanedoses with progressively longer infusion schedules, which is similar tothe situation demonstrated in ovarian cancer.

Head and NeckCancer and Other Tumor Types

As in the use of taxanes in patients with advanced non-small-cell lungcancer, in which early phase II trials of paclitaxel and docetaxel wereperformed using relatively aggressive dosing regimens, there has been apaucity of clinical studies designed to rigorously explore dosing and schedulingissues in patients with advanced head and neck cancer and other tumor types.

Perhaps the only attempt to date is a phase III trial (ECOG 1393) thatevaluated the optimal dosing of paclitaxel on a 24-hour schedule in combinationwith cisplatin in patients with advanced squamous cell carcinoma of thehead and neck.[88] A previous phase II trial of paclitaxel, 250 mg/m²on a 24-hour schedule, produced a 40% response rate. Building on thesedata, Forastiere et al randomized patients with metastatic or locally advanceddisease who had not previously received chemotherapy for recurrent diseaseto treatment with cisplatin (75 mg/m²) following either low-dose paclitaxel(135 mg/m² on a 24-hour schedule) or high-dose paclitaxel (200 mg/m²on a 24-hour schedule) plus G-CSF.

A preliminary analysis demonstrated that response rates were identicalin both arms (35%) and there were no differences in survival parameters.In addition, there appeared to be no significant differences in rates ofboth severe hematologic and nonhematologic toxicities. The results of anaccompanying pharmacodynamic study, which is similar to that performedin the ECOG 5592 trial in non-small-cell lung cancer, is undergoing analysis.As is the case with non-small-cell lung cancer, these clinical resultsindicate that there is no advantage to using paclitaxel doses above 135mg/m² on a 24-hour schedule in combination with cisplatin in patientswith advanced head and neck cancer.

Conclusions

Based on the current cumulative results of nonrandomized and randomizedtrials of the taxanes in cancers of breast, ovary, lung, and head and neck,there does not appear to be a single dosing schedule that produces a vastlysuperior clinical outcome. Although there are reports of both impressiveand unimpressive antitumor activity with some paclitaxel schedules in severaldisease settings in nonrandomized studies, the overall aggressiveness ofthe treatment regimen itself must be taken into account.

In both tissue culture and clinical trials, there appears to be a thresholdpaclitaxel dose or concentration, below which only negligible antitumoractivity is observed, and a plateau dose or concentration, above whichminimal, if any, further antitumor effects occur. The doses at which boththreshold and plateau effects occur appear to be inversely related to theduration of the administration schedule used in clinical practice.

For paclitaxel, it appears that comparable antitumor effects can beachieved with either short (1- and 3-hour) or prolonged (24- and 96-hour)schedules as long as equitoxic dosing regimens are used (ie, higher paclitaxeldoses with short infusion schedules). For docetaxel, although there areinsufficient data available for the gamut of possible dosing schedulesrelative to paclitaxel, the impressive results and toxicities noted withthe most common dosing schedule, 100 mg/m² over 1 hour, indicate thatboth plateau and threshold effects are being achieved. It may not, therefore,be necessary to evaluate alternate administration schedules. However, lowerdocetaxel doses on the 1-hour schedule may result in vastly different toxicityprofiles and therapeutic indices.

In disease settings in which the principal therapeutic goal is palliation(eg, women with recurrent or refractory breast and ovarian cancers), thereappears to be little difference in outcome between various dosing schedulesas long as a minimal plateau paclitaxel dose is exceeded (ie, 3 175 mg/m²on a 24-hour schedule or 3 175 mg/m² on a 3-hour schedule). However,the use of higher doses on shorter administration (ie, more than 200 mg/m²on 1- or 3-hour schedules) may be necessary to achieve a maximal therapeuticoutcome in situations in which prolongation of survival is a reasonableexpectation.

References:

1. Arbuck SG, Canetta R, Onetto N, et al: Current dosage and schedulingissues in the development of paclitaxel (TAXOL). Semin Oncol 4(suppl 3):31-39,1993.

2. Arbuck SA, Blaylock B: Dose and Schedule Issues, in McGuire WP, RowinskyEK (eds): Paclitaxel in Cancer Treatment, pp 151-173. Marcel Dekker, NewYork, 1995.

3. Schiff PB, Horwitz SB: Taxol stabilizes microtubules in mouse fibroblastcells. Proc Natl Acad Sci USA 77:1561-1565, 1970.

4. Diaz JF, Andreu JM: Assembly of purified GDP-tubulin into microtubulesinduced by Taxol and Taxotere. Biochemistry 32:2747-2755, 1993.

5. Bissery M-C, Nohynek, Sanderlink G-J, et al: Docetaxel (Taxotere):a review of preclinical and clinical experience: Part I: preclinical experience.Anti-Cancer Drugs 6:339-355, 1995.

6. Ringel I, Horwitz SB: Studies with RP56976 (Taxotere): A semisyntheticanalogue of Taxol.
J Natl Cancer Inst 83:288-291, 1991.

7. Jordon MA, Wendell K, Gardiner S, et al: Mitotic block induced inHeLa cells by low concentrations of paclitaxel (Taxol) results in abnormalmitotic exit and apoptotic cell death. Cancer Res 56:816-825, 1996.

8. Haldar S, Chintapalli J, Croce CM: Taxol induces bcl-2 phosphorylationand death of prostate cancer cells. Cancer Res 56:1253-1255, 1996.

9. Schiff PB, Gubits R, Kashimawo, et al: Paclitaxel with ionizing radiation,in WG McGuire, Rowinsky EK (eds): Paclitaxel in Cancer Treatment, pp 81-90.Marcel Dekker, New York, 1995.

10. Belotti D, Rieppi M, Nicoletti MI, et al: Paclitaxel inhibits motilityof paclitaxel resistant human ovarian cancer cell. Clin Cancer Res 2:1725-1730,1996.

11. Belotti D, Vergani V, Drudis T, et al: The microtubule-affectingdrug paclitaxel has antiangiogenic activity. Clin Cancer Res 2:1843-1849,1996.

12. Rowinsky EK, Donehower RC, Jones RJ, et al: Microtubule changesand cytotoxicity in leukemic cell lines treated with Taxol. Cancer Res48:4093-4100, 1988.

13. Horwitz SB, Cohen D, Rao S, et al: Taxol: Mechanisms of action andresistance. Monogr Natl Cancer Inst 15:63-67, 1993.

14. Bhalla K, Ibrado AM, Tourkina E, et al: Taxol induces intranucleosomalDNA fragmentation associated with programmed cell death in human myeloidleukemia cells. Leukemia 7:563-568, 1993.

15. Jordan MA, Toso RJ, Thrower D, et al: Mechanism of mitotic blockand inhibition of cell proliferation by Taxol at low concentrations. ProcNatl Acad Sci 90:9552-9556, 1993.

16. Jordan MA, Thrower D, Wilson L: Mechanism of inhibition of cellproliferation by Vinca alkaloids. Cancer Res 51:2212, 1991.

17. Lopes NM, Adams EG, Pitts TW, et al: Cell kill kinetics and cellcycle effects of Taxol on human and hamster ovarian cell lines. CancerChemother Pharmacol 32:235-242, 1993.

18. Helson L, Helson C, Malik S, et al: A saturation threshold for Taxolcytotoxicity in human glial and neuroblastoma cells. Anti-Cancer Drugs4:487-490, 1993.

19. Riccardi R, Servidei T, Spiridigliozzi A, et al: Cytotoxicity ofTaxol in neuroblastoma SH-SY5Y and medulloblastoma TE-671 cell lines invitro. 8th NCI-EORTC Symposium on New Drugs in Cancer Therapy, Amsterdam,195, March 15-18, 1994.

20. Liebmann JE, Cook JA, Lipschultz C, et al: Cytotoxic studies ofpaclitaxel (Taxol) in human tumour cell lines. Br J Cancer 68:1104-1109,1993.

21. Rhiou JF, Naudin A, Lavelle F: Effects of Taxotere on murine andhuman tumor cell lines. Biochem Biophys Res Commun 187:164-170, 1992.

22. Hill BT, Whelan RHD, Shellard SA, et al: Differential cytotoxiceffects of docetaxel in a range of mammalian tumour cell lines and certaindrug resistant cell lines in vitro. Invest New Drugs 12:169-182, 1994.

23. Rowinsky EK, Donehower RC, Tucker RW: Microtubule changes and cytotoxicityproduced by Taxol in human ovarian cell lines. Proc Am Assoc Cancer Res28:423, 1987.

24. Georgiadis MS, Russell E, Johnson BE: Prolonging the exposure ofhuman lung cancer cell lines to paclitaxel increases the cytotoxicity.Proc Int Assoc Study Lung Cancer 11:95, 1994.

25. Figg WD, Thibault A, McCall NA, et al: The in vitro activity ofTaxol on three hormone refractory prostate cancer cell lines, PC3, DU145,and PC3M. Proc Am Assoc Cancer Res 35:431, 1994.

26. Zhan Z, Kang Y-K, Regis J, et al: Taxol resistance: In vitro andin vitro studies in breast cancer and lymphoma. Proc Am Assoc Cancer Res34:215, 1993.

27. Lai GM, Chen YN, Mickley LA, et al: P-glycoprotein expression andschedule dependence of Adriamycin cytotoxicity in human colon carcinomacell lines. Int J Cancer 49:696-703, 1991.

28. Cahan MA, Walter KA, Colvin OM, et al: Cytotoxicity of Taxol invitro against human and rat malignant brain tumors. Cancer Chemother Pharmacol33:441-444, 1994.

29. Kelland LR, Abel G: Comparative in vitro cytotoxicity of Taxol andTaxotere against cisplatin-sensitive and -resistant human ovarian carcinomacell lines. Cancer Chemother Pharmacol 30:444-450, 1992.

30. Wilson L, Miller HP, Farrell KW, et al: Taxol stabilization of microtubulesin vitro. Biochemistry 24:5254-5262, 1985.

31. Parness J, Horwitz SB: Taxol binds to polymerized microtubules invitro. J Cell Biol 91:479-487, 1981.

32. Collins CA, Vallee RB: Temperature-dependent reversible assemblyof Taxol-treated microtubules. J Cell Biol 105:2847-2854, 1987.

33. Lavelle F, Fizames C, Gueritte-Voegelein F, et al: Experimentalproperties of RP 56976, a Taxol derivative. Proc Am Assoc Cancer Res 30:2254,1989.

34. Kelland LR, Abel G: Comparative in vitro cytotoxicity of Taxol andTaxotere against cisplatin-sensitive and resistant human ovarian carcinomacell lines. Cancer Chemother Pharmacol 30:444-450, 1992.

35. Braakhuis BMJ, Hill BT, Dietel M, et al:
In vitro antiproliferative activity of docetaxel (Taxotere), paclitaxel(Taxol), and cisplatin against human tumours and normal bone marrow cells.Anticancer Res 14:205-208, 1994.

36. Garcia P, Braguer D, Carles G, et al: Comparative effects of Taxoland Taxotere on different human carcinoma cell lines. Cancer Chemother.Pharmacol 34:335-343, 1994.

37. Young DH, Michelotti EL, Swindell CS, et al: Antifungal propertiesof Taxol and various analogues. Experientia 4:882-885, 1992.

38. National Cancer Institute Clinical Brochure: Taxol (NSC 125973).Division of Cancer Treatment, NCI, Bethesda, MD, 1993.

39. Rose WC: Taxol: A review of its preclinical in vitro antitumor activity.Anticancer Drugs 3:311-321, 1992.

40. Rose, WC: Taxol-based combination chemotherapy and other in vitropreclinical antitumor studies. Monogr Natl Cancer Inst 15:47-53, 1993.

41. Stearns ME, Wang M: Taxol blocks processes essential for prostatetumor cell (PC-3ML) invasion and metastases. Cancer Res 52:3776-3681, 1992.

42. Medalia O, Aronson M, Ringel I, et al: Inhibition of mouse bladdertumor growth
by intravesicular installation of Taxol. Proc Am Assoc Cancer Res 35:325,1994.

43. Rowinsky EK, Cazenave LA, Donehower RC: Taxol: A novel investigationalanticancer agent. J Natl Cancer Inst 82:1247-1259, 1990.

44. Apfel SC, Lipton RB, Arezzo JC, et al: Nerve growth factor preventstoxic neuropathy in mice. Ann Neurol 29:87-90, 1991.

45. Roytta M, Horwitz SB, Raine CS: Taxol-induced neuropathy: Short-termeffects of local injection. J Neurocytol 13:685-671, 1984.

46. Bissery MC, Guenard D, Gueritte-Voegelein F, et al: Experimentalantitumour activity of Taxotere (RP 57976, NSC 628503), a Taxol analog.Cancer Res 51:4845-4852, 1991.

47. Bruno R, Sanderink GJ: Pharmacokinetics of Taxotere. Cancer Surveys17:305-313, 1993.

48. Rowinsky EK, Donehower RC: Antimicrotubule Agents, in Chabner BA,Longo DL (eds): Cancer Chemotherapy, pp 263-296. Lippincott-Raven, Philadelphia,1996.

49. Rowinsky EK, Wright M, Monsarrat B, et al: Taxol: Pharmacology,metabolism, and clinical implications. Cancer Surveys 17:283-304, 1993.

50. Kearns CM, Gianni L, Egorin MJ: Paclitaxel pharmacokinetics andpharmacodynamics. Semin Oncol 22(suppl 6):16-23, 1995.

51. van Oosterom AT, Schrijvers D: Docetaxel (Taxotere), a review ofpreclinical and clinical experience. Part II: Clinical experience. AnticancerDrugs 6:356-358, 1995.

52. Lesser G, Grossman SA, Eller S, et al: The neural and extaneuraldistribution of systemically-administered [3H]paclitaxel in rats. CancerChemother Pharmacol, 1997 (in press).

53. Eiseman JL, Eddington ND, Leslie J, et al: Plasma pharmacokineticsand tissue distribution of paclitaxel in CD2F1 mice. Cancer Chemother Pharmacol34:465-471, 1994.

54. Marland M, Gaillard C, Sanderink G, et al: Kinetics, distribution,metabolism, and excretion or radiolabelled Taxotere (14C-RP 56976) in miceand dogs. Proc Am Assoc Cancer Res 34:393, 1993.

55. De Valeriola D, Brassine C, Gaillard C, et al: Study of excretionbalance, metabolism, and protein binding of 14C-radiolabelled Taxotere(RP56976, NSC628503) in cancer patients. Proc Am Assoc Cancer Res 34:373,1993.

56. Monsarrat B, Alvinerie P, Wright M, et al: Hepatic metabolism andbiliary excretion of Taxol in rats and humans. Monogr J Natl Can Inst 15:39-46,1993.

57. Walle T, Walle UK, Kumar GN, et al: Taxol metabolism and dispositionin cancer patients. Drug Met Disp 23:506-512, 1995.

58. Gaver RC, Deeb G, Willey T, et al: The disposition of paclitaxel(Taxol) in the rat. Proc Am Assoc Cancer Res 34:390, 1993.

59. Eisenhauer E, ten Bokkel Huinink W, Swenerton KD, et al: European-Canadianrandomized trial of Taxol in relapsed ovarian cancer: High vs low doseand long vs short infusion. J Clin Oncol 12:2654-2666, 1994.

60. Wilson WH, Berg SL, Bryant G, et al: Paclitaxel in doxorubicin-refractoryor mitoxantrone-refractory breast cancer: A phase I/II trial of 96 hourinfusion. J Clin Oncol 12:1621-1629, 1994.

61. Seidman AD, Hochhauser D, Gollub M, et al: Ninety-six hour paclitaxelinfusion after progression during short taxane exposure: A phase II pharmacokineticand pharmacodynamic study in metastatic breast cancer. J Clin Oncol 14:1877-1884,1996.

62. Hainsworth JD, Thompson DS, Greco FA: Paclitaxel by 1-hour infusion:an active drug in metastatic non-small cell lung cancer. J Clin Oncol 13:1609-14,1995.

63. Rowinsky EK, Onetto N, Canetta RM, et al: Taxol: the prototypictaxane, an important new class of antitumor agents. Semin Oncol 19:646-662,1992.

64. Rowinsky EK, Donehower RC: Drug therapy: Paclitaxel (Taxol). N EnglJ Med 332:1004-1114, 1995.

65. Cortex JE, Pazdur R: Docetaxel. J Clin Oncol 13:2643-2655, 1995.

66. Sarosy G, Kohn E, Stone DA, et al: Phase I study of Taxol and granulocytecolony-stimulating factor in patients with refractory ovarian cancer. JClin Oncol 10:1165-70, 1992.

67. Kohn EC, Sarosy G, Bicher A, et al: Dose-intense Taxol: High responserate in patients with platinum-resistant recurrent ovarian cancer. J NatlClin Inst 86:18-24, 1994.

68. McGuire WP, Rowinsky EK, Rosenshein NB, et al: Taxol: A unique antineoplasticagent with significant activity in advanced ovarian epithelial neoplasms.Ann Intern Med 111:273-279, 1989.

69. Thigpen T, Blessing J, Ball H, et al: Phase II trial of paclitaxelin patients with progressive ovarian carcinoma after platinum-based chemotherapy:A Gynecological Oncology Group study. J Clin Oncol 12:1748-1753, 1994.

70. Einzig AI, Wiernik P, Sasloff J, et al: Phase II study and long-termfollow up of patients treated with Taxol for advanced ovarian adenocarcinoma.J Clin Oncol 10:1748-1753, 1992.

71. Reed E, Bitton R, Sarosy G, et al: Paclitaxel dose intensity. JInfus Chemother 6:59-63, 1996.

72. Rowinsky EK, Mackey MK, Goodman SN: Meta analysis of paclitaxeldose-response and dose-intensity in recurrent or refractory ovarian cancer.Proc Am Soc Clin Oncol 15:284, 1996.

73. Omura GA, Brady MF, Delmore JE, et al: A randomized trial of paclitaxelat 2 dose levels and Filgastrim (G; G-CSF) at 2 dose levels in platinumpretreated epithelial ovarian cancer (OVCA): A Gynecologic Oncology Group,SWOG, NCTTG and ECOG study. Proc Am Soc Clin Oncol 15:280, 1996.

74. Nabholtz J-M, Gelmon K, Bontenbal M, et al: Multicenter, randomizedcomparative study of two doses of paclitaxel in patients with metastaticbreast cancer. J Clin Oncol 14:1858-1867, 1996.

75. Ravdin P, Burris HA, Cook G, et al: Phase II trial of docetaxelin advanced anthracycline-resistant or anthracenedione-resistant breastcancer. J Clin Oncol 13:2879-2885, 1995.

76. Valero V, Holmes FA, Walters RS, et al: Phase II trial of docetaxel:a new, highly effective antineoplastic agent in the management of patientswith anthracycline-resistant metastatic breast cancer. J Clin Oncol 13:2886-2894,1995.

77. Rhone-Poulenc Rorer: Data on file.

78. Hudis CA, Seidman AD, Crown JPA, et al: Phase II and pharmacologicstudy of docetaxel as initial chemotherapy for metastatic breast cancer.J Clin Oncol 14:58-65, 1996.

79. Chevallier B, Fumoleau P, Kerbrat P, et al: Docetaxel is a majorcytotoxic drug for the treatment of advanced breast cancer: A phase IItrial of the Clinical Screening Cooperative Group of the European Organizationfor Research and Treatment of Cancer. J Clin Oncol 13:314-322, 1995.

80. Trudeau ME, Eisenhauer EA, Higgins BP, et al: Docetaxel in patientswith metastatic breast cancer: A phase II study of the National CancerInstitute of Canada-Clinical Trials Group. J Clin Oncol 4:422-428, 1996.

81. Dieras V, Chevalier B, Kerbrat P, et al: A multicentre phase IIstudy of docetaxel 75 mg/m2 as first-line chemotherapy for patients withadvanced breast cancer: Report of the Clinical Screening Group of the EORTC.Br J Cancer 74:650-656, 1996.

82. Peretz T, Sulkes A, Chollet P, et al: A multicenter randomized studyof two schedules of paclitaxel in patients with advanced breast cancer.Eur J Cancer 31A (suppl 5):S75, 1995.

83. Bristol-Myers Squibb: Data on file.

84. Velero V, Burris HA, Jones SE, et al: Multicenter pilot study ofTaxotere in Taxol-resistant metastatic breast cancer. Proc Am Soc ClinOncol 15:95, 1996.

85. Kunitoh H, Watanabe K, Onoshi T, et al: Phase II trial of docetaxelin previously untreated advanced non-small cell lung cancer. J Clin
Oncol 14:1649-1655, 1996.

86. Bonomi P, Kim K, Chang A, et al: Phase III trial comparing etoposide-cisplatinversus Taxol with cisplatin-granulocyte-colony-stimulating factor versusTaxol-cisplatin in advanced non-small cell lung cancer: An Eastern CooperativeOncology Group trial. Proc Am Soc Clin Oncol 15:382, 1996.

87. Rowinsky EK, Bonomi P, Jiroutek M, et al: Pharmacodynamic studiesof paclitaxel in ECOG 5592: A phase III trial comparing etoposide pluscisplatin vs low-dose paclitaxel plus cisplatin vs high-dose paclitaxelplus cisplatin plus G-CSF in advanced non-small cell lung cancer. ProcAm Soc Clin Oncol, 1997 (in press).

88. Forastiere AA, Leong T, Murphy B, et al: A phase III trial of highdose paclitaxel + cisplatin + G-CSF versus low dose paclitaxel + cisplatinin patients with advanced squamous cell carcinoma of the head and neck:An Eastern Cooperative Oncology Group trial. Proc Am Soc Clin Oncol, 1997(in press).