Evolving Concepts in the Systemic Drug Therapy
`|
`of Breast Cancer
`
`Larry Norton
`
`Experimental and clinical observations of the prolifera-
`tion of cancer cells and their responses to cytotoxic
`drugs already have had an impact on the design ofanti-
`cancer therapies and it is possible that further under-
`standing of the natural history of tumors will enable
`better treatments to be developed. This review ad-
`dresses several approaches to improving the prognosis
`for patients with breast cancer in which our under-
`standing of tumor dynamics plays an important role.
`Increasing dose intensity can be achieved by dose esca-
`lation (increasing the amount of drug) or increasing
`dose density (reducing the time between treatments).
`The Gompertzian model of tumor growth, which is
`concordant with many experimental andclinical obser-
`vations, can offer an explanation why, although dose
`escalation improves the number ofclinical responses,
`cure is still uncommon.
`In the Gompertzian model,
`smaller tumors grow faster, so tumor regrowth be-
`tween treatment cycles is more rapid when cell kill is
`greatest. Reducing the time available for tumor re-
`growth (increasing dose density), which is now possible
`through theuse of colony-stimulating factors to hasten
`hematopoietic recovery, may have a greater impact
`on clinical outcome than dose escalation. Sequential
`schedules allow optimal doses to be used in dose-dense
`cycles, Several cycles of the optimal dose of agent A,
`followed by several cycles of the optimal dose of agent
`B, may be moreeffective than the simultaneous combi-
`nation of suboptimal doses of A and B. In this cantext,
`agents A and B maybesingle agents or established
`combinations. This tactic allows new agents, such as
`the taxoids, to be used in conjunction with established
`therapies, such as doxorubicin plus cyclophosphamide,
`at optimal doses in dose-intensive regimens. Although
`such regimens may maximize cytoreduction, this may
`not be sufficient to improvesignificantly the long-term
`outcomefor patients with breast cancer. A recenttrial
`using high-dose consolidation chemotherapy with au-
`tologous bone marrow support has thrown doubt on
`the assertion, implicit in Gompertzian cytodynamics,
`that optimal cytotoxic therapy can kill all tumor cells.
`The results of this trial suggest that a consistent num-
`ber of tumorcells remain whether high-dose chemo-
`therapy with autologous bone marrow support is used
`in patients in complete pathologic remission or in pa-
`tients with overt relapse. If there is a lower limit to
`cytoreduction, other approaches must be developed to
`control or prevent the regrowth of residual tumorcells,
`This will require a better understanding of the molecu-
`lar biology of breast cancer and the ability to predict
`and assessthe sensitivity of individual patient’s tumors
`to particular therapies, Factors such as HER2 overex-
`pression are already beinglinked to sensitivity to partic-
`ular agents and the products of oncogenes such as HER2
`maybe targeted by biologic therapies such as mono-
`clonal antibodies. Furthering our understanding of the
`
`biology and behavior of tumor cells may lead to signifi-
`cant improvements in the long-term prognosis for pa-
`tients with early and advanced breast cancer.
`Semin Oncof 24(suppl 10):S10-3-S10-10. Copyright ©
`1997 by W.B. Saunders Company.
`
`NCOLOGYis the field of medicine that is
`most closely aligned to cytokinetics, the sci-
`ence of the proliferation of cells and the growth
`of cellular populations. Much has been learned
`about the proliferative behavior of cancercells by
`examining the biochemistry and cell biology of
`tumors, their natural histories, and their responses
`to anticancer drugs, especially the agents classified
`as cytotoxics. The concepts so derived already
`have had a significant impact on the design of
`cancer therapies. There is some hope that by tak-
`ing the principles already established to the next
`stage of understanding, significant advances in the
`treatmentof patients with malignant tumorsofall
`stages may be made, with breast cancer as a prime
`example.
`This brief review will address several potentially
`useful approaches to improving the prognosis for
`patients with breast cancer. These approachesin-
`volve dose intensity, which has been rigorously
`defined.! Thefirst approach constitutes dose esca-
`lation,
`ie,
`increasing the doses of therapeutic
`agents. This is the most common contemporary
`method of increasing dose intensity, and often is
`erroneously regarded as synonymous with dose in-
`tensity. The second approach involves dose den-
`sity; administering more drug by reducing the in-
`terval between treatment
`cycles. The
`third
`approach involves sequential scheduling, as op-
`posed to other scheduling strategies, as ameans of
`increasing dose intensity by increasing dose den-
`sity. This could be a more effective,
`less toxic
`method of optimizing cancercell kill than other
`styles of chemotherapy administration, The newer
`
`
`From the Department of Medicine, Memorial Sloan-Kettering
`Cancer Center, New York, NY.
`Address reprint requests to Larry Norton, MD, Department of
`Medicine, Memorial Sloan-Kettering Cancer Center, 1275 York
`Ave, New York, NY 10021.
`Copyright © 1997 by W.B. Saunders Company
`0093-7754/97/2404-1002$05 00/0
`
`Seminars in Oncology, Vol 24, No 4, Supp! [0 (August), 1997: S10-3-S10-10
`
`S10-3
`
`(cid:43)(cid:82)(cid:86)(cid:83)(cid:76)(cid:85)(cid:68)(cid:3)(cid:89)(cid:17)(cid:3)(cid:42)(cid:72)(cid:81)(cid:72)(cid:81)(cid:87)(cid:72)(cid:70)(cid:75)(cid:3)
`Hospira v. Genentech
`(cid:44)(cid:51)(cid:53)(cid:21)(cid:19)(cid:20)(cid:26)(cid:16)(cid:19)(cid:19)(cid:27)(cid:19)(cid:24)(cid:3)
`IPR2017-00805
`(cid:42)(cid:72)(cid:81)(cid:72)(cid:81)(cid:87)(cid:72)(cid:70)(cid:75)(cid:3)(cid:40)(cid:91)(cid:75)(cid:76)(cid:69)(cid:76)(cid:87)(cid:3)(cid:21)(cid:19)(cid:25)(cid:24)
`Genentech Exhibit 2065
`
`

`

`$10-4
`
`LARRY NORTON
`
`‘DOSE ESCALATION AND DOSE DENSITY
`
`concept of the limits of cytoreduction, whichhas
`arisen recently on the basis of a very provocative
`clinical trial in high-dose chemotherapy of meta-
`static (stage IV) disease, also will be discussed.”
`_ This latter discussion leads to a consideration of
`the need for the integration of chemotherapy with
`other treatment modalities able to prevent or to
`control tumor regrowth after cytotoxic cytoreduc-
`tion,
`
`growth, the doubling time becomes progressively
`longer as the tumor grows larger. This means that
`small tumors grow faster than larger tumors, but
`also have more cells in their mitotic cycle and are
`thus more susceptible to chemotherapy. Figure 1,
`left, illustrates the pattern ofcell kill with conven-
`tional doses of therapy given in a conventional
`schedule widely spaced in time, eg, every 4 weeks.
`This pattern has been established and collaborated
`by many animal studies and is concordant with
`clinical observations.’ The proportion of cells
`killed increases as the overall number ofcells de-
`It has long been knownthat for many drugs and
`creases; this is the “Gompertzian” phenomenon.
`many tumors, the escalation of the dose level of
`therapy apparently increases the number of cancer
`Gompertzian kinetics predict, however,
`that re-
`cells that are killed.’ It therefore is not surprising
`growth will be more rapid as cell numbers decrease,
`a conceptthat also has been collaborated by exper-
`that retrospective analyses and a few prospective
`imental and clinical observations.’
`trials have suggested that dose intensity is some-
`The mechanisms
`underlying Gompertzian
`times a positive prognostic factor for the outcome
`of patients treated with chemotherapy.** The term
`growth are still obscure, but it is easy to imagine
`a reason for its persistence throughout evolution.
`“dose intensity” was formulated by Hryniuk as the
`amountof drug divided by the size of the patient
`Normalorgans follow Gompertzian-type kinetics in
`their embryonic and childhood development and
`divided by the period of therapy,
`ie, milligrams
`per square meter (of body surface area) per week.’
`in their response to toxins. Gompertzian growth is
`an effective way for an organ to respondto perturba-
`Increasing the dose level was for many years the
`only way of increasing the intensity of therapy.
`tion:
`the greater the perturbation, the faster the
`This is because the other variable within a treat-
`recovery. A tumor, sharing this mechanism of re-
`sponse to injury,
`is predisposed to rapid recovery
`ment schedule, the interval between doses, could
`not be manipulated because recovery from mucosal
`from cytotoxicity, limiting our ability to reduce the
`and hematologic toxicities is not, within broad
`number of cancer cells sufficiently to prohibit re-
`limits, dependent on the doses of chemotherapy
`growth. Increasing the dose of therapy might in-
`crease cell kill per cycle; however, since it does
`used. Although regimens using lower doses of
`not prevent the regrowth of cells between cycles of
`agents have been associated,
`in both anecdotal
`therapy, regrowth occurs, leading to eventual re-
`experiences and randomized trials, with reduced
`response rates and even survival rates, it cannot
`lapse of the disease (Fig 1, center), This might ex-
`plain thefailure of dose escalation to alter the prog-
`be assumed that an increasing dose-response curve
`nosis dramatically for most patients with breast
`is universal among agentsor diseases. With chis in
`cancer.A major effort in clinical investigation
`mind, the results are awaited of a series of large
`trials by the National Surgical Adjuvant Breast
`over several decades has led to the development of
`and Bowel Project that examine the alkylating
`highly sophisticated dose-escalated regimens, It is
`universally agreed that such regimens increase the
`agent cyclophosphamide at doses ranging between
`600 and 2,400 mg/m’ in the adjuvant chemother-
`likelihood of anticancer response and the probabil-
`ity of complete remission.’ The long-term results,
`apy of breast cancer(B. Fisher, personal communi-
`however, have been disappointing, since cureisstill
`cation), Moreover, even if a strictly increasing
`dose-effect relationship were assumed,
`this does
`uncommon. The major impedimentto cure is pet-
`haps the regrowth phenomenon described above.
`not prove that clinical results could be improved
`by simply increasing the dose level.
`In addition, as will be explained below, a second
`This second, somewhat cryptic point may be impediment may beafinite limit to the magnitude
`
`illustrated by computer simulations of the impact
`of the cytoreduction accomplished by a short-course
`of various chemotherapy plans, for example, by
`chemotherapy regimen of even the highest possible
`dose level. For these reasons, we should perhaps
`a Gompertzian tumor that, although idealized,
`is
`modelled on clinical data? In Gompertzian
`now recognize thatit is unlikely that dose escalation
`
`

`

`SYSTEMIC DRUG THERAPY OF BREAST CANCER
`
`S10-5
`
`Higher-dose therapy
`
`Dose-densetherapy
`
`/ Lower-dose therapy
`t
`t tt
`
`wz
`
`7010
`
`108
`
`Cellnumber. 104
`
`108
`
`Fig I. Mathematical models
`of tumor cytoreduction and re-
`growth following conventional,
`dose-escalated, and dose-dense
`treatment regimens.
`
`10?
`
`24
`
`"0
`
`8
`
`16
`
`24
`
`0
`
`8
`
`16
`
`24
`
`Time (in weeks)
`
`alone will make a dramatic difference to the out-
`come of treatment for breast cancer.
`Our discussion thus far has concerned the most
`common method of increasing dose intensity,
`ie,
`increasing dose level. Now, however, for the first
`time, it is possible to increase dose intensity by a
`second manipulation: by decreasing the time inter-
`val between dose administrations. The advance per-
`mitting this new manipulation is the in vitro produc-
`tion of hematopoietic colony-stimulating factors
`capableoffacilitating recovery of the bone marrow.
`Theuse of agents such as granulocyte colony-stimu-
`lating factor (G-CSF) allows myelotoxic agents to
`be given more frequently. This increases dose inten-
`sity even when the doses of the agents are not esca-
`lated. Hence, even when the dose-response curve
`for an agentis flat, a dose-dense schedule assures
`that more drug is given per unit of time, resulting
`in the death of more cancer cells. The reason for
`the theoretical superiority of dose density may relate
`to the temporallimits imposed on regrowth between
`cycles (Fig 1, right), Alternatively, dose density may
`be advantageous because more sustained exposure
`to growth dysregulatory agents may permanently im-
`pair growth-promoting intracellular signaling? In
`addition, G-CSFallows some measure of dose escala-
`tion. When the dose-responserelationship rises, re-
`sulting in increased cell kill with higher doses of
`agents, the combination of dose escalation and dose
`density might be an extremely efficient method of
`increasing cell kill, by eradicating whole populations
`of cancer cells and thereby improving therapeutic
`results (Fig 1, right).
`
`SEQUENTIAL SCHEDULES
`If dose intensity can be increased using both
`dose escalation and dose density,
`the question
`
`arises of how agents should be used in combina-
`tion. The answer to this question is embedded in
`our discussion of dose density, but it does not nec-
`essarily involve the use of hematopoietic cytokines
`to shorten cycle lengths. We have, since the dawn
`of modern medical oncology, become used to the
`concept of simultaneous combination chemother-
`apy. When it wasoriginally proposed, the idea of
`using several toxic drugs simultaneously was not
`automatically accepted by the therapeutic commu-
`nity. Indeed, its establishment as the cornerstone
`of medical oncology represents a victory of intel-
`lectual persistence, clinical research, and clinical
`experience, over preconceived concepts and in-
`trinsic conservatism.®” Simultaneous combina-
`tions have proved to be moreeffective than single
`agents, mostly in regimens that giveall the drugs
`in adequate dosage. The simultaneous combina-
`tion of two or more drugs, however, often necessi-
`tates a reduction in the dose of each agent com-
`pared with its dose level as a single agent.
`Let us examine, for example, the situation in
`which the dose of a drug has to be reduced by 50%
`in order for it to be added to a combination. For
`treatment to be effective, it would be expected
`that half of the dose of the agent would be more
`than half as efficacious as a full dose. If this is not
`the case, 50% of drug A plus 50% of drug B might
`be less efficacious, in terms of cell kill, than 100%
`of drug A or 100% of drug B. The reduced efficacy
`might compromise the possibility of effecting a
`cure. As an extension of this concept, if A + B
`can be given at full dosage of each component,
`and C + D also can be given at full dosage, A +
`B and C + D can then be considered as single
`units. However, if in order to make a combination
`A+B+C+4+D, 50% of A +B and 50% of C
`
`

`

`$10-6
`
`LARRY NORTON
`
`Alternating therapy
`ABABABAB
`
`Sequentialtherapy
`AAAABBBB
`
`ox,WHVH
`
`
`
`Cellnumber a ocy
`
`104
`
`Aw
`
`
`
`
`
`0
`
`8
`
`16
`
`24
`
`0
`
`8
`
`16
`
`24
`
`Time (in weeks)
`
`Fig 2, Mathematical models of tumor cytoreduction and
`regrowth following alternating and sequential dose-dense cyto-
`toxic treatment regimens. Brokenlines indicate cells sensitive
`to treatment A;solid lines indicate cells sensitive to treatment
`B.
`
`+ D must be given, the same issues of reduced
`efficacy may apply.
`Sequentialand alternating schedules are two ap-
`proaches to overcoming the problem ofdoselevel.
`Such schedules need not be based on the use of
`single agents (A, B, C, or D), but may feature both
`simultaneous combinations (A + B and C + D)
`and single agents. This approach is illustrated by
`the Milan adjuvant
`trial
`in which single-agent
`doxorubicin and the established combination of
`cyclophosphamide/methotrexate/fluorouracil
`(CMF) were given in alternating or sequential
`schedules.’° A theoretical construct relevant to an
`understanding of this study is examined below.
`Figure 2 illustrates the solution of a mathemati-
`cal model simulating the use of alternating and
`sequential schedules. Both types of schedule are
`capable of delivering single agents or combinations
`at full dosages. There is, however, an important
`difference between the schedules. Alternating
`schedules are less dose dense than sequential
`schedules because the interval between the like
`components of the alternation are relatively long:
`in an alternation of ABABAB,the cycles of A are
`lengthenedto allow the Bs to be given.If a cancer
`cell population is sensitive to A but not to B, the
`A-sensitive population has a long time to recover
`between administrations of A. This applies equally
`to cells sensitive to B but resistant to A. As a
`result, both the A-sensitive and B-sensitive cells
`receive treatment that is not dose dense, and both
`populations escape eradication (Fig 2, left).
`
`Sequential schedules, on the other hand, are
`more dose dense for each agent(Fig 2, right). Cells
`sensitive to agent'A may be eliminated over sev-
`eral cycles of therapy when A is administered at
`optimal doses. The cells sensitive to agent B re-
`main unaffected until a dose-dense regimen of
`agentB is introduced.If, however, the B-sensitive
`cells are proliferating slowly during the part of the
`regimen when agent A alone is being used, the
`loss, in terms of the increase in number of B-sensi-
`tive cells present at the end of treatment A,is
`small compared with the gains to be expected by
`then using treatment B in a dose-dense fashion,
`Thevalidity of the scenario constructed above
`is supported by the results of the Milan adjuvant
`trial in patients with high-risk primary breast can-
`cer (with three or more positive axillary lymph
`nodes). Doxorubicin followed by CMF shows
`greater efficacy than doxorubicin alternating with
`CME."° Anoldertrial by the Cancer and Leukemia
`Group B (CALGB) has demonstrated that the
`combination of vinblastine, doxorubicin, thiotepa,
`and the steroid fluoxymesterone, given sequen-
`tially after a regimen of CMFplus vincristine plus
`prednisone (CMFVP), improves outcomefor high-
`tisk patients compared with CMFVP alone.'!
`Other sequential regimens have shown advantages
`in breast cancer and other diseases.’
`If sequential regimens using non—cross-resistant
`agents are able to improve substantially the out-
`comeof patients with breast cancer, many possibil-
`ities are opened for the integration of new agents
`and for the improvementin the efficacy of estab-
`lished agents by manipulations, such as biochemi-
`cal modulation and biotherapy. For example, the
`taxoids, docetaxel
`(Taxotere; Rhéne-Poulenc
`Rorer, Antony, France) and paclitaxel (Taxol;
`Bristol-Myers Squibb Oncology, Princeton, NJ),
`are two highly active agents that have been added
`only in the last decade to ourlist of useful drugs
`for the treatmentof advanced breast cancer.!” Both
`drugs have an element of non—cross-resistance
`with the other most active agents, especially the
`anthracyclines. In a pilot study by Hudis et al at
`the Memorial Sloan-Kettering Cancer Center, pa-
`tients with high-risk stage II/IIIA breast cancer
`were treated with a sequential single-agent sched-
`ule involving doxorubicin followed by paclitaxel
`and high-dose cyclophosphamide (ATC).¥ All
`agents were administered every 2 weeks for three
`cycles each, as facilitated by the use’ of G-CSF.
`
`

`

`SYSTEMIC DRUG THERAPY OF BREAST CANCER
`
`$10-7
`
`The ATC regimen was given to 42. patients with
`four or more positive axillary nodes (median,eight;
`range, four to 25), and more than 80% of these
`patients are disease free at 42 months.
`Other investigators are using a similar strategy.
`Currently, in the United States, the CALGBis
`coordinating a randomized, prospective Intergroup
`trial
`(with the Eastern Cooperative Oncology
`Group and the Southwest Oncology Group) of the
`treatment of node-positive patients with operable
`breast cancer with the combination of doxorubicin
`and cyclophosphamide (AC) for four cycles at the
`conventional cycle length of 3 weeks, followed
`by no additional chemotherapy, or four cycles of
`paclitaxel at a conventional dose and schedule.
`(Doxorubicin is given at one of three dose levels
`to investigate the issue of dose escalation as well
`as sequential scheduling.) The National Surgical
`Adjuvant Breast and Bowel Project is engaged in
`a similar trial, using docetaxel, which also investi-
`gates the timing of surgery. Because the issue of
`short-course, very high-dose chemotherapy with
`autologous bone marrow support (ABMS)is still
`unresolved in the adjuvant chemotherapy ofbreast
`cancer, the Intergroup is currently comparing the
`ATC regimen with ABMSin patients with be-
`tween four and nine involved axillary lymph
`nodes. Patients with 10 or more involved nodes
`are being entered into two trials to investigate
`whether ABMS is more effective than conven-
`tional combination chemotherapy. These patients
`may, however, enter the ATC versus ABMStrial
`when their accrual goals have been met in 1997.
`Atpresent, the Intergroup is contemplating a trial
`for node-positive patients who are not candidates
`for or who do notelect participation in the ATC
`versus ABMStrial. In this trial, patients will be
`randomized to ATC administered in a 2-week cy-
`cle (with G-CSF) or a 3-week cycle (without G-
`CSF) or to AC followed by paclitaxel as described
`above, also in 2- or 3-week cycles. This two-by-
`two design will enable analysis of the advantages
`of dose density by cycle length manipulation versus
`the advantages of dose density by sequencing.
`These examples of dose density trials also illus-
`trate that the taxoids are now part of many major
`investigations aimed at the optimization of chemo-
`therapy in the operable (stages I] and IITA) adju-
`vant setting. It is theoretically possible that the
`application of similar concepts of dose intensity
`(escalation and density) and sequencing might add
`
`to the possibility of also treating locally advanced
`(stage IIIB) and stage IV disease. It is to be ex-
`pected that the taxoids will play a major role here,
`especially considering their lack of cross-resistance
`with the anthracyclines.
`
`THE LIMITS OF CYTOREDUCTION
`
`the above arguments were based on the
`All
`Gompertzian model of tumor growth and regres-
`sion, and the exponential model of Skipper etal,
`from whichit derived,’ In the Skipper and Norton-
`Simon models,
`the degree of cell kill is strictly
`related to the tumorsize at the time of therapy
`and the dose of therapy, with no limits on the
`magnitude of cytoreduction until the last cancer
`cell is eliminated. Therapythatis started at a small
`tumorsize will result in a smaller number of surviv-
`ing cancer cells than whenit is started at a larger
`tumor size. Should the cancer cells regrow from
`their point of maximum cytoreduction, the degree
`of cell kill could be estimated by measuring the
`time it takes for the cells to reach a population size
`sufficient to be identified as a recurrence. Hence,it
`may be expected thatif the therapy is used to treat
`a smaller numberof cells, resulting in a smaller
`residual volume after treatment,
`it would take
`longer for the tumor to relapse than a tumor
`treated ata larger initial size. A recently presented
`preliminary report of a clinical trial in the high-
`dose chemotherapy of stage IV disease has cast
`serious doubts overthis central hypothesis.’ Analy-
`sis of the results of this trial suggest that a small but
`consistent number of tumorcells survive maximal
`cytoreductive therapy, whether this treatmentis
`given at the time of complete remission or at the
`time of relapse from complete remission. Should
`this hypothesis be verified, massive cytoreduction
`alone cannot be expected to cure all individuals
`with stage IV disease. Further improvements in
`clinical outcomeswill therefore depend on the in-
`tegration of other disease-control modalities that
`inhibit the regrowth of this residual population of
`tumorcells.
`In this clinical trial, patients with metastatic
`breast cancer were given intensive induction che-
`motherapy using doxorubicin, 5-fluorouracil, and
`methotrexate.” Patients who achieved a complete
`response (98 of 423 patients) were randomly as-
`signed to receive either immediate consolidation
`therapy with high-dose chemotherapy or to be
`allowed to relapse, at which time the same form
`
`

`

`$10-8
`
`LARRY NORTON
`
`of high-dose chemotherapy was used. The high-
`dose treatment consisted of cyclophosphamide/cis-
`platin/carmustine with ABMS. As expected, the
`patients receiving immediate ABMShad a longer
`initial disease-free survival than those observed
`who were untreated until relapse (0.9 years v 0.3
`years; P = .008). However, overall survival was
`better in the group observed until relapse (3.2 years
`v 1.9 years; P = .04). The disease-free and overall
`- survival measured from the time of ABMSwas the
`same both in patients who underwent ABMSat
`complete remission and in those who were so
`treatéd at recurrence. Because the outcomeafter
`the time of ABMS was the same in both groups,
`the total disease-free and overall survival in pa-
`tients receiving delayed ABMS had improved by
`the magnitudeof the disease-free survival after the
`induction complete remission plus the magnitude
`of
`the disease-free and overall survival after
`ABMS.Since approximately 10% of the patients
`randomized to the observation (delayed ABMS)
`group werestill in initial complete remission after
`7 years, they did not require ABMS,butstill con-
`tributed to the total results of delayed ABMS.
`From a practical point of view, the results of
`this study suggest that should a stage IV patient
`achieve a complete remission from conventionally
`dosed chemotherapy, she would be better treated
`by ABMSat the time of recurrence (should she
`relapse) than by immediate ABMS. Thetheoreti-
`cal
`implications are equally important. As dis-
`cussed above,time to recurrence is a good measure
`of the numberof residual cancercells that remain
`after a treatment. In this study, the time to recur
`after ABMS was the sameirrespective of whether
`the high-dose treatment was given at the time of
`complete remission (small number of tumor cells)
`or at the time of recurrence (a much larger number
`of tumorcells), suggesting that the numberof can-
`cer cells remaining after ABMS was the same in
`both cases, even though the numberofcells pres-
`ent at the start of therapy was dramatically differ-
`ent. This means that there must be a finite limit
`to the cytoreduction caused by chemotherapy.
`Oncea critical numberofcells is achieved, chemo-
`therapy cannot further reduce the population size.
`If this is true, it would explain why the results of
`high-dose programs in breast cancer are so consis-
`tent in spite of wide variations in the agents used
`and the styles of administration.’ It would also
`explain whythe overall results are somewhatdisap-
`
`pointing in terms of the small numberof patients
`whoarestill disease free many years after ABMS.
`The optimistic side of this analysis, however, is
`that massive cytoreductions are possible, and are
`indeed routinely achieved, even when high-dose
`chemotherapy is used at the time of macroscopic
`disease. This could be extremely meaningful for
`planning novel therapies. If there is a maximum
`degree of tumor eradication that can be achieved
`using chemotherapy, other treatments must be
`added to prevent the regrowth of residual tumor
`cells. Defining the limits of cytoreduction and
`learning how to control or eradicate residual dis-
`ease may be our biggest challenge in the next de-
`cade.
`
`PATIENT SELECTION
`
`Preventing regrowth of residual cancercells will
`require a better understanding of the cellular and
`molecular biology of breast cancer. Several hints
`at this relate to the sensitivity of different breast
`cancers to cytotoxic therapies. The CALGB has
`published the interesting observation concerning
`patients whose tumors overexpress the oncogene
`product and putative growth factor receptor, HER2
`(which may work by dimerization with other re-
`ceptors of the epidermal growth factor family,
`rather than by ligand-receptor interaction). Pa-
`tients with tumors that overexpress HER2 mayde-
`tive the most benefit from higher doses of the
`adjuvant combination of cyclophosphamide/doxo-
`tubicin/5-fluorouracil.'* The interpretation of this
`observation is notstraightforward, because it could
`mean that HER2 expression indicates specialsensi-
`tivity (and, hence, better results as more drug is
`given) or special resistance (higher doses are re-
`quired to achieve an effect), At the Memorial
`Sloan-Kettering Cancer Center, Seidman et al
`have found a relationship between HER2 expres-
`sion and response to taxoids that indicates in-
`creased sensitivity.” Using the 4D5 antibody to
`the extracellular domain of the HER2 gene prod-
`uct,
`this group detected overexpression in 51 of
`126 patients, with stage [V breast cancer who were
`treated with either paclitaxel or docetaxel. HER2
`overexpression was strongly correlated with poor
`prognostic factors but, paradoxically, with better
`response to taxoids (Table 1). Even within the
`subsets most associated with poor prognosis (vis-
`ceral-dominant disease, extensive prior chemo-
`therapy, and low performance score), HER2 over-
`
`

`

`SYSTEMIC DRUG THERAPY OF BREAST CANCER
`
`
`
`Table |. HER2 Expression, and Sensitivity to Taxoid
`Drugs in Stage IV Breast Cancer
`
`No.of Patients (%)
`No. ofpatients treated
`126
`Taxoid administered
`Paclitaxel
`Docetaxel
`HER2 overexpression*
`
`106 (84)
`20 (16)
`51 (40.5)
`OddsRatio for Response:
`HER2*/HER2~
`
`Variable
`
`2.467
`Visceral-dominant disease
`2.32F
`Extensive prior chemotherapy
`
`Low performance score [93+
`
`* Scored > 2+ in comparison with control sections using
`immunostaining with monoclonal antibody 4D5.
`FP < .005.
`Data from Seidmanetal,'*
`
`
`expression predicted better response. Should this
`observation be corroborated, it mightindicate that
`it will be possible to choose optimal drugs for an
`individual patient.
`Of perhaps even greater potential impactis the
`relationship between HER2 and chemotherapy
`sensitivity, which may indicate an association be-
`tween mitotic regulation by signal transduction
`and the biologic effects of classic cytotoxic ther-
`apy. Such an association has long been anticipated
`by cytokinetic reasoning.*!® This possibility is
`closely allied with research aiming to exploit
`growth factor receptors directly for anticancer ef-
`fect. For example, Baselga and Mendelsohnstudied
`the humanbreast tumor xenograft, BT-474, which
`overexpresses HER2.'’ The growth of this xeno-
`graft can be inhibited, but not cured, by paclitaxel.
`Treatment with the murine anti-HER2 mono-
`clonal antibody, 4D5, also inhibited tumor growth,
`producing some cures. When animals were given
`both paclitaxel and the antibody,
`tumor growth
`was profoundly inhibited, and approximately half
`of the treated animals had no detectable tumorat
`autopsy. Similar results were seen with doxorubi-
`cin, but not with all agents. A humanized 4D5
`antibody also has been used as a single agent in
`patients with previously treated stage IV breast
`cancer exhibiting HER2 overexpression.'* Thera-
`peutic responses were seen, including one patient
`whose complete remission has lasted for over
`3 years. Such therapeutic manipulations could the-
`
`$10-9
`
`oretically overcomethelimits of cytoreduction de-
`scribed in the previous section or suppress tumor
`regrowth after chemotherapeutic cytoreduction.
`Antibodies to the epidermal growth factor recep-
`tor, tyrosine kinase-inhibiting drugs, ras-farnesyla-
`tion inhibitors, and vaccines against MUC-1 and
`other antigens could have a similar role in the
`future.
`
`CONCLUSIONS
`
`In this brief review, several novel approaches
`to the chemotherapy of cancer, especially breast
`cancer, have been considered. Three strategies for
`increasing dose intensity have been discussed: dose
`escalation, dose density, and sequential schedul-
`ing. Dose escalation has somereal advantages, and
`if doses are escalated enough, it may be possible
`to achieve a minimum residual
`tumor burden.
`Nevertheless, dose escalation has limitations, and
`increased dose density already has been shown by
`clinical trials to be an effective way of achieving
`a high cell kill. Sequential scheduling, which is
`one way of achieving increased dose density,is also
`rational, and the combination of dose density via
`shortened cycle length (with G-CSF) and sequen-
`tial scheduling has yielded promising preliminary
`results, which will now be subject to prospective
`evaluation in randomized clinicaltrials. Sequential
`scheduling has allowed the easy integration of new,
`highly effective drugs, such as the taxoids, into the
`therapeutic plan. The selection of patients most
`likely to be sensitive to a treatment could greatly
`improve response rates to that treatment, and the
`combination of therapies such as chemotherapy
`and monoclonal antibodies to growth factor recep-
`tors may have the potential further to improve the
`outcome of treatment.
`
`REFERENCES
`
`1. Hryniuk WM: Average relative dose inten