Breast cancer is the most common noncutaneous malignancy inwomen in industrialized countries. Chemotherapy prolongs survival inpatients with early-stage breast cancer, and maintaining the chemotherapydose intensity is crucial for increasing overall survival. Manypatients are, however, treated with less than the standard dose intensitybecause of neutropenia and its complications. Prophylactic colonystimulatingfactor (CSF) reduces the incidence and duration of neutropenia,facilitating the delivery of the planned chemotherapy doses.Targeting CSF to only at-risk patients is cost-effective, and predictivemodels are being investigated and developed to make it possible forclinicians to identify patients who are at highest risk for neutropeniccomplications. Both conditional risk factors (eg, the depth of the firstcycleabsolute neutrophil count nadir) and unconditional risk factors(eg, patient age, treatment regimen, and pretreatment blood cell counts)are predictors of neutropenic complications in early-stage breast cancer.Colony-stimulating factor targeted toward high-risk patients startingin the first cycle of chemotherapy may make it possible for fulldoses of chemotherapy to be administered, thereby maximizing patientbenefit. Recent studies of dose-dense chemotherapy regimens with CSFsupport in early-stage breast cancer have shown improvements in disease-free and overall survival, with less hematologic toxicity than withconventional therapy. These findings could lead to changes in how earlystagebreast cancer is managed.
ABSTRACT: Breast cancer is the most common noncutaneous malignancy inwomen in industrialized countries. Chemotherapy prolongs survival inpatients with early-stage breast cancer, and maintaining the chemotherapydose intensity is crucial for increasing overall survival. Manypatients are, however, treated with less than the standard dose intensitybecause of neutropenia and its complications. Prophylactic colonystimulatingfactor (CSF) reduces the incidence and duration of neutropenia,facilitating the delivery of the planned chemotherapy doses.Targeting CSF to only at-risk patients is cost-effective, and predictivemodels are being investigated and developed to make it possible forclinicians to identify patients who are at highest risk for neutropeniccomplications. Both conditional risk factors (eg, the depth of the firstcycleabsolute neutrophil count nadir) and unconditional risk factors(eg, patient age, treatment regimen, and pretreatment blood cell counts)are predictors of neutropenic complications in early-stage breast cancer.Colony-stimulating factor targeted toward high-risk patients startingin the first cycle of chemotherapy may make it possible for fulldoses of chemotherapy to be administered, thereby maximizing patientbenefit. Recent studies of dose-dense chemotherapy regimens with CSFsupport in early-stage breast cancer have shown improvements in disease-free and overall survival, with less hematologic toxicity than withconventional therapy. These findings could lead to changes in how earlystagebreast cancer is managed.
Breast cancer is the most commonnoncutaneous malignancyin women in industrializedcountries; it has been estimatedthat it will affect 211,300 women inthe United States in 2003, accountingfor 32% of the new cases of cancer inwomen. Breast cancer incidenceincreases with age. The median age atonset is 63 years, with approximately45% of those who die of the diseasebeing 65 years old or older. Breastcancer is the leading cause of cancerdeath in women after lung cancer, butthe 5-year survival rate is much higherthan that in other cancers.Standard treatment for patientswith early-stage breast cancer includesadjuvant chemotherapy, which confersa survival benefit, while chemotherapyis considered palliative in patients withadvanced disease. The most recentanalyses of the Early Breast CancerTrialists' Collaborative Group foundthat treatment with adjuvant chemotherapyresulted in 23.5% reductionsin the annual risk of the recurrence ofbreast cancer and 15.3% reductions indeath due to breast cancer. A clearbenefit was seen in patients up to 70years old, with anthracycline-containingregimens appearing to be betterthan non-anthracycline-containingregimens.The impact of maintaining the doseintensity of the chemotherapy on disease-free and overall survival has beenshown in trials of adjuvant chemotherapy.A retrospective study byBonadonna and colleagues assessedsurvival with adjuvant chemotherapywith CMF (cyclophosphamide[Cytoxan, Neosar], methotrexate,fluorouracil [5-FU]); follow-up at 20years found that patients receiving less
than the standard dose had suboptimaloutcomes (Figure 1). The investigatorsfound that optimal long-termsurvival was associated with a delivereddose of at least 85% of the referencestandard.The importance of relative doseintensity was confirmed in a prospectiverandomized trial in more than1,500 patients with early-stage breastcancer treated with three different doseintensities (low, moderate, and standard)of adjuvant chemotherapy withcyclophosphamide, doxorubicin, and5-FU. At a median follow-up of 9years, both disease-free and overallsurvival were higher with the standardand moderate dose intensities thanwith the low dose intensity. A recentstudy by the Cancer and LeukemiaGroup B (CALGB) of adjuvant chemotherapyin nearly 6,500 patients,including those described above,showed similar dose-related benefits;the patients treated with what was consideredat the time to be high-doseadjuvant chemotherapy were 12%more likely to remain alive and disease-free at a median follow-up of 9.6years than were those treated with lowdosechemotherapy.Neutropenia and theDelivery of ChemotherapyMaintaining full-dose chemotherapyis often hampered by the occurrenceof myelosuppression, withchemotherapy-induced neutropeniabeing the primary cause of dose delaysand reductions in patients withearly-stage breast cancer. A nationwidesurvey of more than 1,000 patientswith early-stage breast cancertreated at 13 oncology practices foundthat 30% of the patients received lessthan 85% of the standard referencedose. The doses were delayed orreduced in 45% of the patients overall,and neutropenia was the cause in61% of these dose modifications. Amore recent and much larger surveyof practice patterns, in more than20,000 patients with early-stage breastcancer treated with adjuvant chemotherapy,found that 35% of patientshad dose reductions of more than 15%and 25% had a treatment delay of morethan 7 days. Overall, 56% of thepatients were treated with a relativedose intensity of less than 85%, including67% of those older than 65 years.Randomized clinical trials haveshown that preemptive managementwith granulocyte colony-stimulatingfactor (G-CSF) can reduce the durationof severe neutropenia and the riskof its complications.[8,9] Such prophylaxisalso facilitates the delivery ofthe planned dose of chemotherapy ontime. In a trial by de Graaf and colleagues,74% of patients treated withCSF 24 hours after chemotherapywere given at least 85% of the plannedchemotherapy dose intensity, comparedto only 45% of the controls.A meta-analysis of the results of eightrandomized controlled trials of CSFfound that chemotherapy dose reductionsor delays were more than twice
as common in patients who were givenplacebo. Furthermore, the use ofCSF early in therapy affects the likelihoodof neutropenic complications inboth the initial and the subsequentcycles of therapy.Two randomized phase III trialsthat compared the G-CSFs filgrastim(Neupogen) and pegfilgrastim(Neulasta) in patients with breastcancer treated with docetaxel(Taxotere) and doxorubicin found thatpegfilgrastim was comparable tofilgrastim in reducing the incidence offebrile neutropenia and the duration ofgrade 4 neutropenia.[12,13] The patternof less severe neutropenia in thelater cycles was also seen in the pivotaltrials of filgrastim, with the durationof grade 4 neutropenia being reducedfrom 3 days in cycle 1 to 1 dayin the later cycles.[8,9]Colony-stimulating factor has beenshown to be clinically effective, but itsuse in all patients with breast cancertreated with myelosuppressive chemotherapywould not be cost-effective. Itis important to devise strategies forusing CSF in those patients who areat greatest risk for neutropenic complicationsand reduced dose intensity.These strategies will identify thosepatients who are most likely to benefitfrom CSF, which will reduce therisk of neutropenia and make it possiblefor full-dose-intensity chemotherapyto be maintained. Validatedand reliable predictive models shouldbe useful in guiding treatment decisionsand selecting patients for prophylaxiswith CSF.Risk Models for Neutropenia inBreast CancerSilber and colleagues developed arisk model for neutropenic complicationsin patients treated with adjuvantchemotherapy for early-stage breastcancer. Logistic regression modelswere developed for unconditional(pretreatment) factors and conditional(based on the patients' initial hematologicresponse to the chemotherapy)factors. The pretreatment model wasunsuccessful in accurately predictingneutropenia, dose reductions, or dosedelays. The conditional model, however,found a significant associationbetween the depth of the absolute neutrophilcount (ANC) nadir in cycle 1and subsequent neutropenic complications.The value of the first-cycle ANCnadir was validated as a predictor ofcomplications of neutropenia in a secondgroup of patients with early-stagebreast cancer. This retrospectiveanalysis used the same definition ofneutropenic events as Silber and colleaguesbut also examined the incidenceof febrile neutropenia. The firstcycleANC nadir was the only risk factorthat was found to be significantlypredictive (P < .0001) of neutropenicevents in subsequent cycles, with therate of neutropenic complications being30% in patients with a first-cycleANC nadir of 0.5 109/L or less and10% in those with higher nadirs (P =.04). The first-cycle ANC nadir alsopredicted the likelihood of the chemotherapydose intensity being less than85%. The rates of chemotherapy doseintensity of 85% or less were 55% inpatients with a first-cycle ANC nadirof 0.5 109/L or less and 32% in thosewith higher nadirs (P = .05).The Silber model, in which G-CSFis used in all subsequent cycles in patientswith a first-cycle ANC nadir of0.5 109/L or less (categorized as highrisk), was prospectively validated in astudy by Rivera and colleagues.In this study, the rate of low delivereddose intensity-85% or less-in highriskpatients was only 5%, comparedto 12.1% in matched historical controls(Figure 2). The incidences of febrileneutropenia and of hospitalizationdue to febrile neutropenia weresimilar in the high-risk group and thehistorical controls (10.9% and 9.4%,and 4.2% and 4.7%, respectively) (seeFigure 2). Silber and colleagues thenbuilt a cost-effectiveness model basedon the relation between chemotherapydose intensity and disease-free survivalin early-stage breast cancer thathad been shown in CALGB 8541.They found a cost-effectiveness ratioof $34,297 per year of life saved whenG-CSF was used in the 50% of patientswho were classified as high risk by thismodel.A study of pretreatment and posttreatmentfactors in patients withearly-stage breast cancer treated withadjuvant chemotherapy found that severalvariables affect the risk. Thepretreatment factors that predictedhematologic complications in the first
cycle of chemotherapy included agegreater than 65 years, treatment withanthracycline-containing regimens,and low pretreatment blood cellcounts. Posttreatment factors duringthe first cycle that were predictiveof subsequent neutropenic events includedANC nadir less than 0.5 109/L, febrile neutropenia in the firstcycle, and the extent of the drop in hemoglobinlevel.Risk models that have been developedthus far have been based on retrospectiveanalyses and have definedneutropenic complications in variousways. A prospective registry that includespatients with breast cancer hasbeen implemented in order to furtheridentify pretreatment and early treatmentrisk factors for subsequent neutropeniato make early interventionwith first-cycle CSF support possible.Dose-Dense Chemotherapy inEarly-Stage Breast CancerA recent study with dose-densechemotherapy (ie, using standard dosesizes but shortening the intervals betweenthe cycles) has shown better patientoutcomes. The observationthat maintaining the dose intensity ofstandard chemotherapy is associatedwith better clinical outcomes led to theinvestigation of using greater dose intensityin adjuvant chemotherapy forbreast cancer. Dose intensity (theamount of drug delivered per unit oftime) can be increased either by increasingthe dose (dose escalation) orby shortening the interval betweencycles (dose density). The results withdose escalation have thus far beenequivocal.Studies that compared conventional-dose chemotherapy and highdosechemotherapy followed by stemcell transplantation (HDCT/SCT) havemostly been negative in the adjuvantsetting in high-risk patients with nodepositivedisease.[21-24] A few trialshave shown a benefit of HDCT/SCTin a subset of patients. Roche et alshowed an improvement in diseasefreesurvival in those with 7 to 10 positivenodes. Recently, Tallman etal showed a significantly longer timeto recurrence in those assigned to
HDCT/SCT who met strict eligibilitycriteria. Rodenhuis et al showedgreater relapse-free survival in thosewith 10 or more positive nodes andthose with tumors that did notoverexpress HER2/neu. None ofthese trials has shown a benefit in overallsurvival in all patients.The dose-dense approach to increasingdose intensity is based onpreclinical models of the growth ofcancer cells by nonexponentialGompertzian kinetics. In volume-reducedGompertzian cancer models theregrowth of cancer cells between thecycles of cytoreduction is more rapidthan in exponential models. Thus,it is unclear whether simple dose escalationis enough for the success ofadequately planned multicycle regimens,since other strategies, such asdose density, may prove to be morepotent as a therapeutic manipulation.This hypothesis was initially investigatedin pilot studies. Hudis and colleaguesreported the results in a trialin which 42 patients with resectedbreast cancer that involved more thanthree positive nodes were treated withsequential dose-dense chemotherapy,consisting of three cycles of doxorubicin,followed by three cycles ofpaclitaxel, and then three cycles ofcyclophosphamide (A -> T -> C) in14-day cycles with G-CSF support.This regimen proved to be feasible andpromising. The reported actuarial disease-free survival rate was 78% aftera median of 4 years, with only 4 deathsdue to metastatic disease.Because combination chemotherapyis often more toxic than singleagents, Fornier and colleagues comparedthe toxicity of a sequential dosedenseregimen of doxorubicin,paclitaxel, and cyclophosphamide(A -> T -> C) and that of another dosedenseregimen with the same scheduleof doxorubicin, followed by threecycles of concurrent paclitaxel andcyclophosphamide (A -> TC).There was greater toxicity in the concurrentarm, with more hospitalizationsfor febrile neutropenia and morered blood cell transfusions for anemia.Furthermore, the mean delivered doseintensities of the paclitaxel and cyclophosphamidewere significantlygreater in the sequential arm thanin the concurrent one (P = .01 andP = .05, respectively). Thus, dosedensesequential chemotherapy ismore feasible than doxorubicin followedby concurrent paclitaxel and cyclophosphamide.The effect of dose-dense chemotherapywas then tested in a large prospectivephase III study coordinatedby the CALGB for the National CancerInstitute's Breast Intergroup, INTC9741. The study compared sequentialdoxorubicin, paclitaxel, andcyclophosphamide (A -> T -> C) withconcurrent doxorubicin and cyclophosphamidefollowed by paclitaxel(AC -> T), using dose-dense (2-weekly) and conventional (3-weekly)schedules, as adjuvant chemotherapyin 1,973 patients with breast cancer.The dose-dense schedule used G-CSFin all patients to make the 2-weekcycles possible by lessening neutropeniccomplications (Figure 3).The dose-dense regimens resultedin significantly higher 3-year diseasefreesurvival (85% vs 81%; P = .01)and 3-year overall survival (92% vs90%; P = .013), regardless of predictivefactors such as the number of positivenodes, tumor size, menopausalstatus, and tumor estrogen receptorstatus (Figure 4). There was no differencein disease-free survival or overallsurvival between the sequential andconcurrent arms. Grade 4 neutropeniawas more common with conventionaltherapy, occurring in 33% of patientstreated with conventional regimensand 6% of those treated with dosedenseregimens (P < .0001). In addition,fewer cycles were delayed becauseof hematologic toxicity withdose-dense than with conventionaltherapy (15% vs 38%; P < .0001). Inorder to deliver conventional chemotherapyin a dose-dense schedule, CSFsupport is required to reduce neutropeniccomplications.There was more grade 4 neutropeniain the 3-week-cycle arm, but morepatients (13%) required red blood celltransfusions in the 2-week-cycle armof AC -> T than in the three other arms.This was not because of significantgrade 3 or 4 anemia, but was perhapscaused by grade 2 anemia (data notreported). With the availability oferythropoietin, the need for red bloodcell transfusions should be diminished.Furthermore, dose-dense chemotherapysignificantly reduced the occurrenceof contralateral breast cancer(0.3% vs 1.5%, P = .0004).CALGB 9741 showed not only thefeasibility of this approach, but alsothe superiority of dose-dense over conventionalchemotherapy. These findingsare exciting and are consistentwith previous mathematical modelpredictions that shortening the intervalbetween chemotherapy cyclescould result in more-effective eradicationof malignant cells, potentiallyimproving survival. The sequentialapproach tested in CALGB 9741failed to show superiority of singleagentsequential therapy over combineddoxorubicin/cyclophosphamidesequenced into paclitaxel, but showedno disadvantage for uncoupling agentsfrom one another, either. On the basisof current data, practicing oncologistsmay consider treating patients withbreast cancer in this dose-dense fashion.However, extrapolating these datato all regimens outside of a clinicaltrial setting should be done with caution,as unexpected toxicities mayemerge. These findings suggest importantavenues for future research in bothbreast cancer and other chemosensitivetumors, and confirmatory studiesare encouraged.ConclusionsMaintaining the dose of the chemotherapyis important in increasinglong-term survival in patients withearly-stage breast cancer. Neutropenia,the major dose-limiting toxic effect ofmyelosuppressive chemotherapy, canbe limited with early use of CSF. Evidence-based risk models for predictingwhich patients are at greatest riskfor neutropenia and its complicationsmay be an efficient and cost-effectiveway of limiting these complicationsand helping ensure that the chemotherapyis delivered as planned. Moreprospective research is needed to determinewhich factors to use in riskmodels that predict subsequent neutropeniaand its complications. Earlyresults with dose-dense chemotherapyin CALGB 9741 are exciting, showingimproved disease-free and overallsurvival and less grade 4 neutropeniawhen compared to conventionallyscheduled chemotherapy. The findingsof this study suggest that dose-densescheduling with appropriate chemotherapyregimens that require CSFsupport may replace conventional dosingas the new standard of care inearly-stage breast cancer.
The author(s) have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
American Cancer Society: Cancer Factsand Figures 2003. Atlanta, American CancerSociety, 2003.
Early Breast Cancer Trialists’ CollaborativeGroup: Polychemotherapy for early breastcancer: An overview of the randomised trials.Lancet 352:930-942, 1998.
Bonadonna G, Valagussa P, Moliterni A,et al: Adjuvant cyclophosphamide, methotrexate,and fluorouracil in node-positive breastcancer: The results of 20 years of follow-up. NEngl J Med 332:901-906, 1995.
Budman DR, Berry DA, Cirrincione CT,et al: Dose and dose intensity as determinantsof outcome in the adjuvant treatment of breastcancer. The Cancer and Leukemia Group B. JNatl Cancer Inst 90:1205-1211, 1998.
Muss HB, Woolf SH, Berry DA, et al:Older women with node positive (N+) breastcancer (BC) get similar benefits from adjuvantchemotherapy (Adj) as younger patients (pts):The Cancer and Leukemia Group B (CALGB)experience (abstract 11). Proc Am Soc ClinOncol 22:4, 2003.
Link BK, Budd Gt, Scott S, et al: Deliveringadjuvant chemotherapy to women withearly-stage breast carcinoma: Current patternsof care. Cancer 92:1354-1367, 2001.
Agboola O, Crawford J, Dale DC, et al:Most patients treated with adjuvant chemotherapyfor breast cancer receive substantiallyreduced dose intensity (abstract 110). Proc AmSoc Clin Oncol 22:28, 2003.
Crawford J, Ozer H, Stoller R, et al: Reductionby granulocyte colony-stimulating factorof fever and neutropenia induced by chemotherapyin patients with small-cell lung cancer.N Engl J Med 325:164-170, 1991.
Trillet-Lenoir V, Green J, Manegold C, etal: Recombinant granulocyte colony stimulatingfactor reduces the infectious complicationsof cytotoxic chemotherapy. Eur J Cancer29A:319-324, 1993.
de Graaf H, Willemse PH, Bong SB, etal: Dose intensity of standard adjuvant CMFwith granulocyte colony-stimulating factor forpremenopausal patients with node-positivebreast cancer. Oncology 53:289-294, 1996.
Lyman GH, Kuderer NM, DjulbegovicB: Prophylactic granulocyte colony-stimulatingfactor in patients receiving dose-intensivecancer chemotherapy: A meta-analysis. Am JMed 112:406-411, 2002.
Holmes FA, O’Shaughnessy JA, VukeljaS, et al: Blinded, randomized, multicenter studyto evaluate single administration pegfilgrastimonce per cycle versus daily filgrastim as anadjunct to chemotherapy in patients with highriskstage II or stage III/IV breast cancer. J ClinOncol 20:727-731, 2002.
Green MD, Koelbl H, Baselga J, et al: Arandomized double-blind multicenter phase IIIstudy of fixed-dose single-administrationpegfilgrastim versus daily filgrastim in patientsreceiving myelosuppressive chemotherapy. AnnOncol 14:29-35, 2003.
Silber JH, Fridman M, DiPaola RS, etal: First-cycle blood counts and subsequentneutropenia, dose reduction, or delay in earlystagebreast cancer therapy. J Clin Oncol16:2392-2400, 1998.
Thomas ES, Rivera E, Erder MH, et al:Using first cycle nadir absolute neutrophilcount (FCNANC) as a risk factor for neutropenicevents: A validation study (abstract 144).Proc Am Soc Clin Oncol 20:37a, 2001.
Rivera E, Erder MH, Moore TD, et al:Targeted filgrastim support in patients withearly-stage breast carcinoma. Cancer 98:222-228, 2003.
Silber JH, Fridman M, Shpilsky A, et al:Modeling the cost-effectiveness of granulocytecolony-stimulating factor use in early-stagebreast cancer. J Clin Oncol 16:2435-2444,1998.
Agboola O, Crawford J, Dale DC, et al:Risk models for neutropenic complications associatedwith breast cancer adjuvant chemotherapy(abstract 261). Proc Am Soc Clin Oncol21:66a, 2002.
Dale DC, Wolff D, Agboola O, et al:Development of a risk model for neutropeniccomplications based on a prospective nationwideregistry (abstract 2229). Proc Am Soc ClinOncol 22:554, 2003.
Citron ML, Berry DA, Cirrincione C, etal: Randomized trial of dose-dense versus conventionallyscheduled and sequential versusconcurrent combination chemotherapy as postoperativeadjuvant treatment of node-positiveprimary breast cancer: First report of Intergrouptrial C9741/Cancer and Leukemia Group B trial9741. J Clin Oncol 21:1431-1439, 2003.
The Scandinavian Breast Cancer StudyGroup 9401. Results from a randomised adjuvantbreast cancer study with high dose chemotherapywith CTCb supported by autologousbone marrow stem cells versus dose escalatedand tailored FEC therapy (abstract 3). Proc AmSoc Clin Oncol 18:2a, 1999.
Hortobagyi GN, Buzdar AU, TheriaultRL, et al: Randomized trial of high-dose chemotherapyand blood cell autografts for highriskprimary breast carcinoma. J Natl CancerInst 92:225-233, 2000.
Peters WP, Rosner G, Vredenburgh J, etal: Updated results of a prospective, randomizedcomparison of two doses of combinationalkylating agents (AA) as consolidation afterCAF in high-risk primary breast cancer involvingten or more axillary lymph nodes (LN):CALGB 9082/SWOG 9114/NCIC Ma-13 (abstract81). Proc Am Soc Clin Oncol 20:21a,2001.
Crown JP, Lind M, Gould A, et al: Highdosechemotherapy (HDC) with autograft(PBP) support is not superior to cyclophosphamide(CPA), methotrexate and 5-FU (CMF) followingdoxorubicin (D) induction in patients(pts) with breast cancer (BC) and 4 or moreinvolved axillary lymph nodes (4+LN): TheAnglo-Celtic I study (abstract 166). Proc AmSoc Clin Oncol 21:42a, 2002.
Roche HH, Pouillart P, Meyer N, et al:Adjuvant high dose chemotherapy (HDC) improvesearly outcome for high risk (N > 7) breastcancer patients: the Pegase 01 trial (abstract 102).Proc Am Soc Clin Oncol 20:26a, 2001.
Tallman MS, Gray R, Robert NJ, et al:Conventional adjuvant chemotherapy with orwithout high-dose chemotherapy and autologousstem-cell transplantation in high-riskbreast cancer. N Engl J Med 349:17-26, 2003.
Rodenhuis S, Bontenbal M, Beex L, etal: High-dose chemotherapy with hematopoieticstem-cell rescue for high-risk breast cancer.N Engl J Med 349:7-16, 2003.
Norton L, Simon R, Brereton JD, et al:Predicting the course of Gompertzian growth.Nature 264:542-545, 1976.
Norton L. Evolving concepts in the systemicdrug therapy of breast cancer. SeminOncol 24(4 suppl 10):S10.3-S10.10, 1997.
Hudis C, Seidman A, Baselga J, et al:Sequential dose-dense doxorubicin, paclitaxel,and cyclophosphamide for resectable high-riskbreast cancer: Feasibility and efficacy. J ClinOncol 17:93-100, 1999.
Fornier MN, Seidman AD, TheodoulouM, et al: Doxorubicin followed by sequentialpaclitaxel and cyclophosphamide versus concurrentpaclitaxel and cyclophosphamide: 5-year results of a phase II randomised trial ofadjuvant dose-dense chemotherapy for womenwith node-positive breast carcinoma. Clin CancerRes 7:3934-3941, 2001.