Dose Intensity of Chemotherapy for Childhood Cancers

Transcription

1 Dose Intensity of Chemotherapy for Childhood Cancers MALCOLM SMITH, JEFFREY ABRAMS, EDWARD L. TRIMBLE, RICHARD S. UNGERLEIDER Clinical Investigations Branch, Cancer Therapy Evaluation Program, National Cancer Institute, Bethesda, Maryland, USA Key Words. Drug dose-response relationship Ewing s sarcoma Rhabdomyosarcoma Neuroblastoma Alkylating agents Acute lymphocytic leukemia Child Infant ABSTRACT Since the formulation of the dose-intensity concept of anticancer therapy in the mid-1980s, the concept that more is better has gained general acceptance among pediatric oncologists. However, recently published clinical trials results for adults with breast cancer, germcell tumors, and ovarian cancer raise questions about the value of dose intensification in improving outcome. Given the differences in sensitivity between pediatric and adult tumors to cytotoxic agents, the results from these adult trials suggest a need for caution but do not suggest that evaluations of dose intensity for pediatric tumors are unwarranted (especially for a tumor such as Ewing s sarcoma, which is especially sensitive to alkylating agents). Since dose-intensive therapies have significant shortand long-term costs for the patient, it is important to obtain reliable data concerning possible benefits of this INTRODUCTION Conventional wisdom among pediatric oncologists seeking to optimize treatment has been that more is better. The broad acceptance of this paradigm is the result of a confluence of factors. The observations by pediatric investigators that significant improvements in outcome could be achieved for children with acute lymphoblastic leukemia (ALL) [1-3] and non-hodgkin s lymphoma (NHL) [4-6] through the use of increasingly intensive therapies was followed by the formulation by others of the dose-intensity concept, and the application of this concept to retrospective analyses of clinical trials [7-9]. These analyses often correlated superior outcome with regimens that were more dose-intensive. Additionally, improvements in supportive care, and, in particular, the clinical development of the hematopoietic growth factors (HGFs), suggested that more intensive therapies could be given with relative safety. Introduction of the recombinant HGFs created strategy. Toward this end, NCI-sponsored randomized clinical trials evaluating the role of dose intensification have been initiated for several tumors of children (including neuroblastoma, germ-cell tumors, Ewing s sarcoma, and brain tumors in infants). These trials should be completed and reported in the next two to three years, and they may make unique contributions in defining the benefits and limitations of dose intensity as a cancer treatment strategy. In the long term, however, the utility of dose intensification is limited for children with cancer because of the inherent toxicities associated with its application. Identification of agents that more specifically target tumor cells is essential. Fortunately, pediatric tumor cells do have unique biological characteristics that may be susceptible to targeting for therapeutic benefit. The Oncologist 1996;1: an atmosphere of excitement that further enhanced enthusiasm for pursuing more aggressive therapies. To illustrate, a 1988 publication predicted that the clinical availability of the recently cloned myeloid colony-stimulating factors created the prospect of a revolution in medicine as profound as the introduction of antibiotics a half a century ago [10]. Nearly a decade later, enthusiasm for increasingly intensive therapies has diminished, particularly among medical oncologists. It seems useful to critically evaluate what has been learned about the strategy of applying more intensive therapies to the treatment of cancer, with a special focus on comparing experiences in childhood versus adult cancers. This review begins with a discussion of some of the limitations of the dose-intensity concept and describes several ways in which it may be misapplied. This background information provides a framework within which to evaluate several retrospective studies that suggested a benefit for dose-intensive therapies for Correspondence: Malcolm Smith, M.D., Ph.D., Clinical Investigations Branch, Cancer Therapy Evaluation Program, National Cancer Institute, Room 741, EPN, Bethesda, MD 20892, USA. Telephone: ; Fax: ; smithm@dct.nci.nih.gov Received August 19, 1996; accepted for publication September 1, AlphaMed Press /96/$5.00/0 The Oncologist 1996;1:

2 Smith, Abrams, Trimble et al. 294 pediatric tumors, and to assess selected randomized clinical trials in adult patients that address the role of dose intensity. The review continues by considering clinical trials in children that have addressed (or are addressing) the contribution of dose intensity to improved outcome, and closes with suggestions for targeting cellular derangements of cancer cells as a successor strategy to that of dose intensification. For the purposes of this review, clinical trials evaluating changes in both cumulative dose and dose intensity are considered to be addressing dose-intensity questions. While purists might dispute this approach, many of the trials evaluating dose intensity also raise the cumulative dose of cytotoxic agents, and a negative result in a trial that increases both dose and dose intensity makes it unlikely that a positive result would be obtained if dose intensity alone were increased. CAVEATS IN THE INTERPRETATION OF STUDIES EVALUATING DOSE INTENSITY The application of the concept of dose intensity to experimental therapy is not always straightforward, and it is important to review some of the pitfalls involved before proceeding. Dose intensity is defined as the amount of drug delivered per unit time and is usually standardized to body surface area as mg/m 2 /week. This definition, while easily applied when considering only one agent, is problematic when evaluating combination chemotherapy regimens. Hryniuk and Bush described a methodology for applying the dose-intensity concept to multiagent regimens in which one regimen is declared to be the standard, and the dose intensities of other regimens are calculated relative to this standard [11]. However, a key assumption in defining this average relative dose intensity (ARDI) for a multiagent regimen is that all of the drugs of the standard regimen contribute equally to the efficacy of the regimen. A specific example illustrates why this may be problematic. A decade ago a multiagent regimen for osteosarcoma containing bleomycin, cyclophosphamide, and dactinomycin (BCD) would likely have been considered the standard regimen in an exercise to evaluate the concept of dose intensity in the treatment of osteosarcoma [12]. A regimen lacking these agents, but which made effective use of cisplatin and doxorubicin (now considered the most active agents for osteosarcoma), might have a lower calculated ARDI than a regimen which used BCD extensively and minimized use of cisplatin and doxorubicin. Since the relative contributions of agents to the efficacy of multiagent regimens used for pediatric tumors is seldom known, the ARDI concept is very difficult to apply to tumors treated with combination chemotherapy. Dose Intensity and Non-Additive Combined Drug Actions In evaluating dose intensity in multiagent regimens, it is also necessary to assume that the efficacy of one agent is not modulated (either positively or negatively) by combination with another agent. For many regimens, this is undoubtedly the case, but clear examples of both positive (i.e., synergistic) and negative interactions exist. For example, experiments with cultured tumor cells have demonstrated synergistic cytocidal activity when topotecan is used with DNA-damaging agents (e.g., cisplatin and alkylating agents) [13, 14]. The clinical correlate of this in vitro observation is that combinations of topotecan with other agents produce far more hematopoietic toxicity than would be anticipated based on the doses of the agents used singly. When topotecan is combined with relatively low doses of either cisplatin [15], carboplatin [16], or cyclophosphamide [17], the tolerable dose of topotecan is less than 50% of its maximally tolerated dose as a single agent. Unquestioning application of the principle of dose intensity would lead to the avoidance of drug pairs that produce synergistic cytotoxicity, since this interaction would limit the doses of each agent that could be given. While in vitro evidence for synergy does not necessarily predict for improved clinical efficacy for a drug combination [18], it nonetheless provides a valid lead which may warrant clinical evaluation of the drug combination in spite of the lower doses required when the agents are given together. The clinical observation that cisplatin preceding etoposide produces higher response rates than cisplatin following etoposide provides evidence that specific interactions between cytotoxic agents can enhance antitumor activity [19]. Dose Intensity and Cell-Cycle Specific Agents Agents with cell-cycle specificity require action during explicit cell-cycle phases in order to kill cells. Prolonged duration of systemic exposure above a minimum threshold concentration helps to assure that a high proportion of cells is exposed to the agent during the critical cell-cycle phase. The antitumor activity of the same total dose of etoposide (500 mg/m 2 ) given as either a single 24-h infusion or divided into five daily 2-h infusions illustrates how poorly the dose-intensity concept applies to cell-cycle specific agents. Patients receiving etoposide as a 24-h infusion had a response rate of 10%, compared with a response rate of 89% for patients receiving the same total dose of etoposide, but given daily as a 1-h infusion for five days [20]. This and other studies have demonstrated that the primary determinant of etoposide antitumor activity is not dose-per-unittime, but rather the duration that steady-state plasma etoposide concentrations are maintained above 1-2 µg/ml [20-23]. For cytarabine, another cell-cycle specific agent, a murine leukemia model system was used to show that single-dose cytarabine is less effective than fractionated cytarabine (every 3 h for eight doses), even when a 10-fold higher total dose of cytarabine is given as a single dose [24]. The only way in which the doseintensity concept can be meaningfully evaluated for cell-cycle specific agents is when the schedule is maintained constant for the regimens compared.

3 295 Dose Intensity of Chemotherapy for Childhood Cancers Dose Intensity and Saturable Physiological Processes Agents such as prednisone and L-asparaginase are especially important in the treatment of pediatric ALL [25, 26], and both of these agents illustrate the inapplicability of the dose-intensity concept to agents which act via a saturable mechanism. For example, L-asparaginase acts by depleting asparagine [27]; doses greater than those required for complete depletion of asparagine increase the dose intensity of asparaginase without further beneficial effect. Similarly, prednisone exerts its antitumor effect by interacting with specific glucocorticoid receptors in leukemia cells; once the required number of receptors has been occupied by prednisone, higher doses of the drug produce no additional antitumor effect [26, 28]. For both agents, a more relevant measure of the intensity with which the agents are used in a particular regimen would be the achieved duration of saturation of the targeted process. The Costs to the Patient of Dose-Intensive Therapy While focusing on improving outcome by increasing the intensity of therapy, it must be remembered that both tumorous and normal tissues will be affected. Thus, the increased short- and long-term toxicities associated with this strategy must be weighed against its benefits, especially in children, for whom a full and normal life after completion of therapy is the goal. It is often not possible to differentiate between those toxicities related to dose intensity and those related to cumulative dose, since commonly both of these variables are increased. One clear example of the cost to patients of more intensive therapies are the studies comparing standard- and high-dose cisplatin, in which neurotoxicity and ototoxicity are markedly higher in patients receiving high-dose cisplatin [29-34]. To date, there is little evidence for clinical benefit from the use of these high cisplatin doses [29, 31, 33, 34]. The acute toxicities associated with extreme and prolonged myelosuppression can also be severe and may neutralize any benefit seen from intensification. Early evaluations of more aggressive chemotherapy regimens in children produced toxic death rates approaching 10% [35], but increased experience with the supportive care measures required to give these regimens safely and the availability of G-CSF have led to reductions in the frequency of infection-related deaths [36, 37]. However, intensive therapies require more days of intravenous antibiotics and hospitalization for the treatment of febrile neutropenia and documented infections [1]. Repeated courses of high-dose chemotherapy can also lead to apparent stem cell depletion, which is clinically manifested as increasingly severe thrombocytopenia and progressively delayed hematopoietic recovery following cytotoxic treatments [37, 38]. In a murine model, repeated treatments with myelosuppressive drugs followed by stimulation with the HGFs induce damage to the host stem cell compartment [39]. The long-term consequences of some degree of stem cell depletion in children are not known. The increased mutagenic burden associated with higher doses of chemotherapy might also lead to higher frequencies of second malignant neoplasms. As an example, a surprisingly high frequency of secondary leukemia was observed within the first two years of initiating treatment among women receiving dose-intensive doxorubicin and cyclophosphamide [40]. These and other adversities highlight the need for careful monitoring for early and late toxicities in clinical trials evaluating dose-intensity questions, and also highlight the need for carefully conducted randomized trials to reliably assess the outcome associated with dose-intensive therapy. Dose Intensity and Alkylating Agents Although a number of limitations of the dose-intensity concept have been given, it should be noted that it applies best to alkylators (as single agents or with fixed doses of other agents). The alkylating agents generally demonstrate a log-linear relationship between dose and cytotoxic activity in preclinical models, and schedule of administration is not a significant determinant of their antitumor activity [24]. Furthermore, animal models provide evidence for alkylating agents that dose intensity (and not cumulative dose) is the dominant treatment design variable with respect to degree of therapeutic response and the likelihood of cure [24]. SELECTED RANDOMIZED CLINICAL TRIALS EVALUATING DOSE INTENSITY IN ADULT PATIENTS While tumors occurring in adults are generally less chemosensitive than those occurring in children, at least some adult cancers are chemoresponsive (e.g., germ-cell tumors, breast cancer, and small-cell lung cancer), and the observations about the role of dose intensity for these tumors may have relevance to the treatment of children. As will be apparent from the examples below, there have been more failures of the dose-intensity principle than successes, although some positive results have been reported. Platinum Dose and Dose Intensity Evaluations of the effect of cisplatin and carboplatin dose intensification on antitumor activity have occurred primarily in the setting of advanced ovarian cancer and testicular cancer. One of the few positive trials of dose intensity in ovarian cancer was performed by the Scottish Gynecology Cancer Trials Group, in which cyclophosphamide was given with cisplatin for six courses, with patients randomized to either 50 mg/m 2 or 100 mg/m 2 of cisplatin per course. The benefit was modest (four-year progression-free survival rate of 27.4% versus 17.0%), and patients in the high-dose

4 Smith, Abrams, Trimble et al. 296 arm experienced significantly more chronic neurotoxicity [32]. While the Scottish trial varied both cumulative cisplatin dose and cisplatin dose intensity, a randomized Gynecologic Oncology Group trial maintained cumulative dose constant and varied only dose intensity (50 mg/m 2 8 courses versus 100 mg/m 2 4 courses) [34]. For this patient population with bulky ovarian epithelial cancers, there was no discernible improvement in patient outcome as a result of doubling the dose intensity, although there was more severe toxicity (hematological, gastrointestinal, infectious, and renal). Others have reported similar results [33, 41]. A retrospective analysis of multiple phase I and II trials evaluated the impact of carboplatin systemic exposure area under the curve (AUC) on response rate for patients with ovarian cancer [42]. In all patient subsets, the likelihood of complete or overall response did not increase significantly above a carboplatin AUC of 5 to 7 mg/ml min, suggesting the absence of a dose-response effect once a threshold systemic exposure is reached. Results of randomized trials evaluating carboplatin dose intensity have appeared only in abstract, but do not appear particularly promising [43, 44]. In one of these studies, patients were randomized to receive carboplatin at an AUC of either 6 or 12 mg/ml min. The response rates were virtually identical for the low- and high-dose carboplatin treatments: 57% (95% confidence interval [CI]: 52-75) and 63% (95% CI: 46-69), respectively. However, thrombocytopenia and febrile neutropenia were much more common among patients receiving the higher dose [43]. Randomized evaluations of the role of cisplatin dose intensity for germ-cell tumors suggest that while substandard doses result in inferior results, there appears to be no benefit for exceeding standard cisplatin doses. A randomized comparison between cisplatin doses of 75 mg/m 2 and 120 mg/m 2, both given with vinblastine and bleomycin, demonstrated a clear advantage in favor of high-dose over low-dose therapy, both in complete response rate (63% versus 43%) and survival [45]. A subsequent randomized study compared cisplatin doses of 100 mg/m 2 and 200 mg/m 2 in combination with etoposide and bleomycin [31]. In this study, there was no benefit for the higher dose in terms of complete response rate, disease-free survival, or survival, although patients receiving the higher dose of cisplatin experienced much higher rates of ototoxicity (32% versus 0%) and neurotoxicity (26% versus 1%) than patients receiving standard-dose cisplatin. Breast Cancer The anthracyclines are among the most active classes of agents against breast cancer, with objective response rates as single agents as high as 60% to 80% [46]. Thus, breast cancer is an appropriate tumor in which to evaluate the dose-response relationship of anthracyclines. A retrospective analysis of multiple phase II trials of epirubicin that evaluated dose levels ranging from 50 mg/m 2 to 180 mg/m 2 in patients with advanced breast cancer suggested a strong dose-response relationship, particularly for doses from 90 mg/m 2 to 180 mg/m 2 [46]. Randomized comparisons of different dose levels of single agent doxorubicin or epirubicin have shown a higher rate of objective responses for higher dose levels [47-49]. However, the anthracycline dose levels evaluated in these studies (as low as 35 mg/m 2 of doxorubicin and mg/m 2 of epirubicin) included levels below the standard doses for these agents. This points out a methodological problem of these types of dose-intensity studies: if standard-dose cytotoxic drug treatment is more effective than no treatment, it is possible to choose a dose level low enough to be indistinguishable from no treatment, and thus a threshold effect may be misinterpreted as a dose-response effect [50]. In a randomized trial that evaluated a wide range of epirubicin doses (40 mg/m 2 to 135 mg/m 2 ), the two higher doses (90 mg/m 2 and 135 mg/m 2 ) were significantly more effective than the lower doses, but there was no difference in efficacy between the two higherdose arms, suggesting that escalation beyond 90 mg/m 2 does not provide additional benefit. However, approximately 20%- 30% of patients with early progressive disease at the lower epirubicin doses did have objective responses when treated subsequently with a higher dose of epirubicin (135 mg/m 2 ), in support of a dose-response relationship. A recent South African trial compared two cycles of highdose cyclophosphamide (2.4 gm/m 2 ), mitoxantrone (35-45 mg/m 2 ), and etoposide (2.5 gm/m 2 ) versus six to eight cycles of conventional dose cyclophosphamide (600 mg/m 2 ), mitoxantrone (12 mg/m 2 ), and vincristine (1.4 mg/m 2 ) as first-line treatment for metastatic breast cancer [51]. The high-dose arm included either autologous bone marrow or peripheral blood stem cell rescue. The complete response rates were 51% for the high-dose arm compared with 4% for the conventional-dose arm, and this resulted in a survival advantage (90 weeks versus 45 weeks). However, the asymmetry of the two treatment arms (etoposide versus vincristine), and the unexpectedly poor response rates and survival observed in the conventional-dose arm, indicate that confirmation of the superiority of high-dose therapy over standard-dose treatment is required. The strategy of dose intensity has also been tested in the earlier stages of breast cancer. The Cancer and Leukemia Group B (CALGB) conducted a trial in which 1,572 women with node-positive, stage II breast cancer were randomized to different dose levels and dose intensity of adjuvant chemotherapy [52]. One cohort received 400 mg/m 2 cyclophosphamide and 40 mg/m 2 of doxorubicin once every 28 days and 400 mg/m 2 of fluorouracil twice every 28 days, for

5 297 Dose Intensity of Chemotherapy for Childhood Cancers six cycles. Another cohort received a higher dose-intensity regimen (50% higher doses of the three drugs) but received only four cycles (equal cumulative dose), and the third cohort of women received half the total dose used in the other two groups and at half the dose intensity used in the second group. The women in the first two cohorts had significantly longer disease-free survival (p < 0.001) and overall survival (p = 0.004) than those receiving the regimen with lower dose intensity and lower cumulative dose. These results are consistent with either a dose-response effect or a threshold level for dose and dose intensity. An important additional observation from this study was that patients randomly assigned to the high-dose regimen had significantly longer disease-free and overall survival if their tumors had c-erbb-2 overexpression, while there was no clear evidence of a dose-response effect in patients with minimal or no c- erbb-2 expression [53]. While this finding requires confirmation, it highlights the important principle that a dose-response effect may be limited to tumors with certain biological characteristics. This has relevance to childhood cancer in that a dose-response relationship may exist for certain adult and pediatric tumors, but not for others, and that an answer to the dose-intensity question obtained in one tumor setting may not be applicable to others. The role of cyclophosphamide dose and dose intensity in adjuvant regimens for breast cancer was evaluated in a large randomized trial National Surgical Adjuvant and Breast Project (NSABP) B-22, n = 2,238 in which women were randomized to receive either standard-dose cyclophosphamide (600 mg/m 2 ) for four courses or escalated-dose cyclophosphamide (1,200 mg/m 2 ) for two or four courses. All patients additionally received the same dose and schedule of doxorubicin (60 mg/m 2 for four courses). The early results from the study indicate no difference in three-year disease-free survival or survival among the three arms, suggesting that within the dose range evaluated there is no dose-response relationship [54]. The successor study (NSABP B-25) extended the cyclophosphamide dose range evaluated to 2,400 mg/m 2 given for either two or four courses, but efficacy results from this study are not yet available. RETROSPECTIVE ANALYSES OF THE ROLE OF DOSE INTENSITY FOR PEDIATRIC TUMORS Several retrospective analyses published in the early 1990s provided evidence that treatments with increased dose intensity may improve outcome for specific tumors of childhood. The significance of these reports was not that they demonstrated a role for regimens with higher dose intensity in the treatment of these tumors, but rather that they suggested hypotheses about dose intensity and the role of specific agents that could be tested in subsequent randomized clinical trials. Cheung and Heller analyzed the dose intensities of five drugs (teniposide, cisplatin, cyclophosphamide, doxorubicin, and vincristine) as used during induction chemotherapy in 44 clinical trials for children ( 1 year) with stage IV neuroblastoma [55]. In their analysis, the dose intensities of teniposide and cisplatin had the greatest influence on clinical outcomes (i.e., proportion of major response, median survival, and median progression-free survival), while those of cyclophosphamide and doxorubicin were less significant. Although many of the protocols in the analysis extended treatment to more than one year, the duration of treatment had no positive influence. There was a striking positive correlation between the mean relative dose intensity for the regimens and clinical response, survival, and progression-free survival. Dose-intensity analyses of published Ewing s sarcoma and osteogenic sarcoma trials were used to determine which agents were most closely associated with a favorable response [56]. For the osteogenic sarcoma trials, the percentage of patients with more than 90% tumor necrosis following neoadjuvant chemotherapy was the end point for analysis. Doxorubicin dose intensity appeared to be the most important determinant of favorable response for osteogenic sarcoma, while increasing dactinomycin dose intensity was associated with a lower likelihood of favorable response. The negative impact of dactinomycin dose intensity may result from diminished doxorubicin dose intensity in those regimens with a higher dactinomycin dose intensity. For Ewing s sarcoma, the primary comparison of dose intensity was between the two treatment arms of the Intergroup Ewing s Study (IESS) -II [56]. The IESS-II trial randomized patients with non-metastatic, non-pelvic Ewing s sarcoma to one of two regimens which differed only in the dose, schedule, and timing of the four drugs (vincristine, dactinomycin, cyclophosphamide, and doxorubicin) [57]. The experimental treatment arm (Treatment 1) was termed a high-dose intermittent regimen because cyclophosphamide was given as a single dose every three weeks rather than the weekly lower dose used in Treatment 2. In both treatment arms, either doxorubicin or dactinomycin was given every six weeks; the arms differed, however, in that Treatment 2 alternated doxorubicin and dactinomycin throughout the entire protocol (i.e., doxorubicin given every 12 weeks), while in Treatment 1 the doxorubicin was given only during the first nine months of treatment and dactinomycin was given during the last nine months of treatment. The planned dose intensity of the two treatment arms was similar for all four drugs for the total treatment duration; however, during the first nine months of therapy, the doxorubicin dose intensity of Treatment 1 was 250% that of Treatment 2, while dactinomycin was not delivered during the first 36 weeks in Treatment 1. Patients receiving the higher doxorubicin dose intensity during the first 36 weeks of

6 Smith, Abrams, Trimble et al. 298 treatment had a five-year disease-free survival of 68% compared with 48% for the treatment arm with lower doxorubicin dose intensity (p = 0.03). PROSPECTIVE CLINICAL EVALUATIONS OF DOSE INTENSITY IN CHILDREN In the past five years there have been numerous evaluations in children with cancer of multiagent chemotherapy regimens using doses of agents significantly exceeding those commonly used in the 1980s. A number of investigators have demonstrated the relative safety of using cyclophosphamide doses as high as 4 gm/m 2 in combination with other agents (e.g., vincristine, doxorubicin, or etoposide) [36, 37, 58]. Dose escalation of ifosfamide has occurred as well, but is limited by unacceptable levels of nephrotoxicity at higher doses [59-62]. Dose escalation of agents with stem cell toxicity (e.g., melphalan, thiotepa, carboplatin) has generally been limited to the bone marrow transplant setting because of prolonged, cumulative thrombocytopenia. However, it is now possible to give multiple courses of these agents at high doses by the use of peripheral blood progenitor cell (PBPC) support [63]. Thus, the use of regimens with high-dose intensity is feasible in children, but the question is whether there is clinical benefit to this strategy. As noted earlier, dose-intensive therapies have significant short- and long-term costs for the patient, and it is important to obtain reliable data concerning possible benefits of this strategy. Pediatric Acute Lymphoblastic Leukemia and B Cell Lymphoma Outcome for children with pediatric ALL and B cell NHL has improved dramatically since the 1970s. How much of this improvement can be attributed to increased dose intensity? For pediatric ALL, identification of effective means of controlling central nervous system leukemia led to marked improvements in outcome in the 1970s [64]. Subsequently, there have been incremental improvements in outcome so that currently over 70% of children survive five years from diagnosis [65]. Several randomized studies provide insight into how this improvement has been achieved. The Children s Cancer Group (CCG) demonstrated that adding one course of treatment ( delayed intensification ) with a regimen including doxorubicin and cyclophosphamide along with other agents (e.g., dexamethasone and cytarabine) improved outcome when compared with a similar treatment program without delayed intensification. The Pediatric Oncology Group (POG) also observed that the addition of cytotoxic agents (including daunomycin, cytarabine, and teniposide) to an antimetabolitebased regimen improved outcome for children with high-risk ALL [66]. In both of these regimens, the superior regimens were clearly more myelotoxic than the less effective standard therapies, and hence could be considered more intensive. However, in neither case can the improvement be credited to an increase in dose intensity, but rather must be attributed to the introduction of one or more additional effective agents. Another POG trial demonstrated that for standard-risk ALL patients, high-dose methotrexate (1 gm/m 2 ) was superior to low-dose oral methotrexate when given in the context of an antimetabolite-based regimen [67]. While this might seem to address a dose-intensity question, the concept of dose intensity is not easily applied to high-dose methotrexate with leucovorin rescue. The improved outcome with high-dose methotrexate is most easily ascribed to better penetration of sanctuary sites [68] and to better saturation of the polyglutamylated methotrexate pool [69]. The UK Medical Research Council (MRC) trial UKALL X does suggest a benefit for additional courses of intensification therapy [70]. In this study, 1,171 children who were treated with intensive induction therapy were randomized to receive additional intensification therapy at five weeks, 20 weeks, both, or neither. The five-year disease-free survival was 71% (95% CI: ) for children randomized to two blocks of intensification therapy, this being significantly better than the 62% (95% CI: ), 61% (95% CI: ), and 57% (95% CI: ) rates for the groups randomized to one intensification block at five weeks, one at 20 weeks, and no intensification, respectively. The additional treatment appeared to be beneficial to all patients, even those traditionally deemed at lower risk of relapse. A CCG study which also randomized patients to one or two blocks of intensification therapy has completed accrual and will provide an opportunity for confirmation of the UKALL X trial. Perhaps for no other malignancy is the role of intensive therapy more apparent than for small noncleaved-cell lymphoma (SNCCL) and B cell (surface immunoglobulin-positive) ALL (B-ALL). Current therapy allows the cure of 80%-90% of children with advanced SNCCL and B-ALL. The most effective regimens induce profound myelosuppression, and treatment courses are given as quickly as hematopoietic recovery allows [71-74]. Comparison of these effective regimens with an older, less-effective regimen [75] suggests that success has been achieved not by maximizing the dose intensity of a single agent, but rather by adding other effective agents (e.g., high-dose methotrexate, high-dose cytarabine, etoposide or teniposide) and by giving successive treatment courses as rapidly as possible. This approach has not only resulted in a very high cure rate, but has had the added benefit of limiting the long-term toxicity that might result from an emphasis on any single agent. Pediatric Germ-Cell Tumors The CCG and POG jointly conducted a randomized trial (open to accrual from February, 1990 to February,

7 299 Dose Intensity of Chemotherapy for Childhood Cancers 1996) for children with high-risk malignant germ-cell tumors comparing two platinum-based regimens, one using a cisplatin dose of 200 mg/m 2 /course and the other using a dose of 100 mg/m 2 /course. Each course of therapy additionally included bleomycin and etoposide, and four to six courses of therapy (depending on response after the first four courses) were to be given. The design for this study increased both the cisplatin cumulative dose and the cisplatin dose intensity, so that any improvement in outcome or increase in toxicity observed could be due to either or both of these factors. Outcome data from this study will answer the question of whether the increased ototoxicity and neurotoxicity associated with high-dose cisplatin is balanced by a significant improvement in survival. Ewing s Sarcoma POG and CCG are also collaborating on an Intergroup study for nonmetastatic Ewing s sarcoma comparing a 48- week standard regimen using five drugs (vincristine, doxorubicin, cyclophosphamide, ifosfamide, and etoposide) to a 30-week dose-intensified regimen using the same chemotherapeutic agents. Although the two treatment regimens have very similar cumulative doses of all five agents, the doseintensified regimen uses significantly higher doses during the first four courses of therapy of cyclophosphamide (4.2 gm/m 2 /course) and ifosfamide (12 gm/m 2 /course) than in the standard regimen (1.2 gm/m 2 /course and 9 gm/m 2 /course, respectively). Thus, during the first four courses of therapy, the dose-intensified regimen has an approximately twofold higher alkylating agent dose intensity. This study opened to accrual in February, 1995, and will require approximately five years of patient entry in order to reliably detect an improvement in event-free survival from 58% to 70%. For metastatic Ewing s sarcoma, attempts at dose escalation have not shown apparent benefit. Sixty children with metastatic Ewing s sarcoma received escalated doses of cyclophosphamide (2.2 gm/m 2 /course), doxorubicin (90 mg/m 2 /course), and ifosfamide (14 gm/m 2 /course) as part of a five-drug regimen that also included vincristine and etoposide [76]. However, the two-year event-free survival (26%) and survival (35%) for these patients did not differ appreciably from that achieved in a patient population receiving the same five agents at standard doses. These results should not necessarily diminish enthusiasm for the dose-intensity evaluation in nonmetastatic patients, since interventions which are not successful in the metastatic setting may prove successful for nonmetastatic Ewing s sarcoma (as witnessed by the benefit observed for the ifosfamide/etoposide combination in patients with nonmetastatic disease but not in patients with metastatic disease) [76, 77]. Infants with Brain Tumors Infants with brain tumors represent a special problem given the severe sequelae associated with irradiation of the immature brain. One strategy to address this problem has been to use chemotherapy following maximum surgical resection, in the hopes of delaying radiotherapy until the child is older and the brain less susceptible to its toxic effects [78]. The approach was only partially successful, since approximately two-thirds of patients developed progressive disease during chemotherapy, indicating the need for more effective chemotherapy. POG investigators are conducting a randomized evaluation of the role of increased dose intensity in improving outcome for children <3 years of age with brain tumors. Within 2-4 weeks following the initial surgical attempt of maximal safe resection, patients are randomized to receive either the standard dose-intensity regimen or a more dose-intensive regimen. The standard dose-intensity regimen includes courses of cyclophosphamide and vincristine and courses of cisplatin and etoposide which are given at four-week intervals. The dose-intensive regimen includes the same combinations given at three-week intervals, and with twofold higher doses of cyclophosphamide and 20% higher doses of cisplatin and etoposide. The overall planned increase in dose intensity is approximately 2.7-fold for cyclophosphamide and approximately 1.3- to 1.6-fold for the other agents. This study began accrual in March, 1992, and should complete accrual in Hematopoietic Stem Cell Rescue for Dose Intensification Despite years of investigation, the role of myeloablative therapy in pediatric oncology remains uncertain. Recent results from randomized trials for children with newly diagnosed AML have shown that autologous bone marrow transplantation (BMT) is no better than conventional chemotherapy, but that allogeneic BMT appears to improve outcome [79-81]. Randomized trials of the role of autologous BMT for ALL have not been conducted in children, but a recent case-control study was unable to document benefit for autologous BMT [82]. For neuroblastoma, conflicting results from nonrandomized clinical trials have been reported. POG investigators observed no benefit for autologous BMT using a Mantel-Byar analysis to minimize the time-to-transplant bias [83]. CCG investigators, on the other hand, reported that subsets of patients receiving autologous BMT had improved outcome, although patient selection bias rather than treatment effect may explain these results [84]. The CCG randomized trial comparing autologous BMT to intensive, standard-dose chemotherapy completed accrual in 1996, and the results from this

8 Smith, Abrams, Trimble et al. 300 important trial will provide an unbiased comparison of these two treatment strategies. However, even if autologous BMT is superior to conventional chemotherapy, new treatment approaches are needed for diseases like advanced-stage neuroblastoma, since the long-term survival following transplantation is unacceptably low [85]. CONCLUSIONS The great treatment successes for childhood cancer have occurred predominantly for those tumors that are markedly more sensitive to therapeutic agents than are normal tissues. Pediatric ALL, in which apoptosis is induced by asparagine depletion, glucocorticoids, and anti-metabolite-directed therapy, illustrates this point, since the therapeutic index for these treatments is very large. The pursuit of dose intensity as a treatment concept is de facto recognition that for certain pediatric tumors, available agents may have a small therapeutic index so that the limits of patient tolerance must be tested in order to achieve significant benefit. Should the lack of benefit observed for dose intensity in a number of clinical trials in adult patients dissuade pediatric oncologists from further evaluations of dose intensity? Given the differences in sensitivity between pediatric and adult tumors to cytotoxic agents, the results from adult trials suggest a need for caution but do not suggest that evaluations of dose intensity for pediatric tumors are unwarranted. For example, childhood cancers such as rhabdomyosarcoma and Ewing s sarcoma are especially sensitive to alkylating agents, as witnessed by the very high objective response rate ( 90%) observed in previously untreated patients to ifosfamide (either as a single agent or given with etoposide) [86, 87]. Chemoresponsive adult tumors such as breast cancer and small-cell lung cancer have lower response rates to ifosfamide than these pediatric tumors [88-90], while other adult tumors such as soft-tissue sarcoma and non-small-cell lung cancer have yet lower response rates [91, 92]. Increasing alkylating agent dose intensity when treating pediatric cancer patients has the potential for achieving higher rates of cure, and clinical trials evaluating dose intensity for pediatric tumors may make unique contributions (given the sensitivity of specific tumors to alkylating agents) in defining the benefits and limitations of dose intensity as a cancer treatment strategy. In the long term, however, the application of the principle of dose intensification is unlikely to provide the revolution that was predicted in 1988; therefore, identification of agents that more specifically target tumor cells is essential. Pediatric tumor cells do have unique biological characteristics that can be targeted for therapeutic benefit. For example: Ewing s sarcoma and alveolar rhabdomyosarcoma both have tumor-specific fusion genes that produce messenger RNAs and fusion proteins that are only present in tumor cells [93]. These fusion molecules may be targeted directly (e.g., through the use of antisense probes or other genetic mechanisms) [94, 95] or indirectly by stimulating the immune system to recognize and kill cells expressing the fusion proteins [96]. Pediatric tumors may also be dependent on specific signal transduction pathways that can be modulated to decrease tumor cell growth and to decrease tumorigenic properties. For example, the insulin-like growth factors (IGFs) appear to promote the survival and growth of a number of pediatric tumors [97, 98], and inhibition of signaling through the IGF- I receptor represses both the in vitro and in vivo growth of human rhabdomyosarcoma cell lines [98, 99]. Agents which block the IGF signaling pathway may find application in treatment of rhabdomyosarcoma and other pediatric solid tumors that employ this signaling pathway for cell growth. Modulation of signaling pathways may alter the sensitivity of tumor cells toward cytotoxic chemotherapy, since in some tumor cells the activity of specific tyrosine kinase signaling pathways appears related to resistance to cytotoxic agents. For example, c-erbb2-overexpressing lung cancer cells require tyrosine kinase activity for their chemoresistant phenotype, and tyrosine kinase inhibitors such as emodin can sensitize these cells to chemotherapeutic drugs [100]. The c-erbb-2 tyrosine kinase receptor also appears to play a role in tumor cell chemoresistance for breast cancer cells, and downmodulation of erbb-2 expression enhances chemosensitivity to cytotoxic agents like cisplatin [101]. The tyrosine kinase activity of the ABL proto-oncogene appears to protect cells against cytotoxic agents, and in chronic myeloid leukemia cell lines, lowering cellular levels of the BCR ABL fusion protein markedly increases the sensitivity of the cells to chemotherapy and other inducers of apoptosis [102, 103]. The ras signal transduction pathway appears to be especially important for some pediatric myeloid leukemias [ ]. Agents that inhibit ras signal transduction [107, 108] are now entering clinical evaluation and should be considered for evaluation in children. Monoclonal antibodies (with or without toxin conjugates) are a form of targeted therapy that may be especially beneficial for childhood cancers. The relative specificity of the monoclonal antibodies, combined with a toxicity profile that is generally non-overlapping with conventional cytotoxic agents, make these agents attractive for pediatric evaluation [109].

9 301 Dose Intensity of Chemotherapy for Childhood Cancers The dramatic advances in understanding the molecular biology and physiology of pediatric tumor cells have laid the groundwork for development of these targeted therapeutics. Evaluations of dose intensity are best viewed as an important and necessary stepping stone toward a future of more refined and tumor-specific treatments for childhood cancer. REFERENCES 1 Tubergen D, Gilchrist G, O Brien R et al. Improved outcome with delayed intensification for children with acute lymphoblastic leukemia and intermediate presenting features: a Childrens Cancer Group Phase III Trial. J Clin Oncol 1993;11: Gaynon P, Steinherz P, Bleyer W et al. Improved therapy for children with lymphoblastic leukemia and unfavorable presenting features: a follow-up report of the Childrens Cancer Group Study CCG-106. J Clin Oncol 1993;11: Niemeyer CM, Reiter A, Riehm H et al. Comparative results of two intensive treatment programs for childhood acute lymphoblastic leukemia: the Berlin-Frankfurt-Munster and Dana-Farber Cancer Institute protocols. Ann Oncol 1991;2: Patte C, Michon J, Frappaz D et al. Therapy of Burkitt and other B-cell acute lymphoblastic leukaemia and lymphoma: experience with the LMB protocols of the SFOP (French Paediatric Oncology Society) in children and adults. Baillieres Clin Haematol 1994;7: Murphy SB, Bowman WP, Abromowitch M et al. Results of treatment of advanced-stage Burkitt s lymphoma and B cell (SIg+) acute lymphoblastic leukemia with high-dose fractionated cyclophosphamide and coordinated high-dose methotrexate and cytarabine. J Clin Oncol 1986;4: Reiter A, Schrappe M, Ludwig W-D et al. Chemotherapy in 998 unselected childhood acute lymphoblastic leukemia patients. Results and conclusions of the multicenter trial ALL-BFM 86. Blood 1994;84: Levin L, Simon R, Hryniuk W. Importance of multiagent chemotherapy regimens in ovarian carcinoma: dose intensity analysis (comments). J Natl Cancer Inst 1993;85: Hryniuk W. The importance of dose intensity in the outcome of chemotherapy. In: Hellman S, DeVita V, Rosenberg S, eds. Important Advances in Oncolology. Philadelphia: J.B. Lippincott Company, 1988: Surbone A, DeVita VT, Jr. Dose intensity. The neglected variable in clinical trials. Ann NY Acad Sci 1993;698: Gasson D, Golde J. Hormones that stimulate the growth of blood cells. Sci Am 1988;259: Hryniuk W, Bush H. The importance of dose intensity in chemotherapy of metastatic breast cancer. J Clin Oncol 1984;2: Link M, Goorin A, Miser A et al. The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremity. N Engl J Med 1986;314: Kaufmann S, Peereboom D, Buckwalter C et al. Cytotoxic effects of topotecan combined with various anticancer agents in human cancer cell lines. J Natl Cancer Inst 1996;88: Chou TC, Motzer RJ, Tong Y et al. Computerized quantitation of synergism and antagonism of taxol, topotecan, and cisplatin against human teratocarcinoma cell growth: a rational approach to clinical protocol design (comments). J Natl Cancer Inst 1994;86: Rowinsky E, Grochow L, Kaufmann S et al. Sequencedependent effects of topotecan and cisplatin in a phase I and pharmacokinetic study. Proc Am Soc Clin Oncol 1994;13:142a. 16 Heideman R, Kuttesch J, Stewart C et al. A phase I trial of a fixed systemic exposure (AUC) of carboplatin (CARBO) with continuous infusion (CI) topotecan (TOPO) in pediatric solid tumors. Proc Am Soc Clin Oncol 1995;14:447a. 17 Saylors R, Stewart C, Wall D et al. Phase I trial of topotecan plus cyclophosphamide in children with refractory or recurrent solid tumors: a Pediatric Oncology Group study. Proceedings ASPH/O 1995;4:24-25a. 18 Greco W, Faessel H, Levasseur L. The search for cytotoxic synergy between anticancer agents: a case of Dorothy and the ruby slippers? J Natl Cancer Inst 1996;88: Maksymiuk AW, Jett JR, Earle JD et al. Sequencing and schedule effects of cisplatin plus etoposide in small-cell lung cancer: results of a North Central Cancer Treatment Group randomized clinical trial. J Clin Oncol 1994;12: Slevin M, Clark P, Joel S et al. A randomized trial to evaluate the effect of schedule on the activity of etoposide in small-cell lung cancer. J Clin Oncol 1989;7: Clark PI, Slevin ML, Joel SP et al. A randomized trial of two etoposide schedules in small-cell lung cancer: the influence of pharmacokinetics on efficacy and toxicity. J Clin Oncol 1994;12: Joel SP, Slevin ML. Schedule-dependent topoisomerase II-inhibiting drugs. Cancer Chemother Pharmacol 1994;34 (suppl):s84-s Sonnichsen DS, Ribeiro RC, Luo X et al. Pharmacokinetics and pharmacodynamics of 21-day continuous oral etoposide in pediatric patients with solid tumors. Clin Pharmacol Ther 1995;58: Skipper H. Dose intensity versus total dose of chemotherapy: an experimental basis. In: DeVita V, Hellman S, Rosenberg S, eds. Important Advances in Oncology Philadelphia: J.B. Lippincott Company, 1991: Pui C-H, Crist W. Biology and treatment of acute lymphoblastic leukemia. J Pediatr 1994;124: Gaynon PS, Lustig RH. The use of glucocorticoids in acute lymphoblastic leukemia of childhood. Molecular, cellular,