Review. Is 'one cycle every three or four weeks' obsolete? A critical review of dose-dense chemotherapy in solid neoplasms. K. Fizazi & L.

Transcription

1 Annals of Oncology 11: S 2000 Klimer Academic Publishers. Printed in the Netherlands. Review Is 'one cycle every three or four weeks' obsolete? A critical review of dose-dense chemotherapy in solid neoplasms K. Fizazi & L. Zelek Department of Medicine, Institut Gustave-Roussy, Villejuif, France Summary Background: Shortening the interval between cycles is one means of increasing the dose intensity of chemotherapy, and can be supported by biological and mathematical rationales. Our objective was to assess the clinical relevance of the rapid repetition of regimens (so-called 'dose-dense chemotherapy') in various solid neoplasms. Design: The medical literature was reviewed in accord with Mulrow's recommendations. Randomised studies comparing frequently-repeated chemotherapy to standard regimens as well as open studies are described and critically examined. Results: Dose-dense regimens were widely found to be feasible. In small-cell lung cancer, survival of patients receiving dose-dense regimens was better than that of patients treated by standard chemotherapy in three trials, two of which reached significance, when these intensive regimens allowed better dose intensity. In poor-prognosis germ-cell tumors, a dose-dense regimen was not better than standard therapy, perhaps because of an excessively high toxicity-related death rate. However, recent phase II studies have provided encouraging results. In early breast cancer, the one published randomized study in the adjuvant setting showed only a trend towards better disease-free survival in node-positive women receiving a weekly-repeated regimen. Two randomized trials failed to show any benefit in the neoadjuvant setting with a dose-dense regimen. No evidence of a benefit was provided in metastatic breast cancer. In advanced colorectal cancer, evidence of an improvement in survival with weekly or bi-weekly 5-FU-leucovorin compared to a classic monthly schedule has recently been shown in two randomized trials, and dose-dense regimens are recognized as standard therapy in many countries. studies of dose-dense regimens have also shown high response rates and long survival in many neoplasms, including Ewing's sarcoma, gestational trophoblastic disease, ovarian carcinoma and gastric cancer. Conclusions: A considerable amount of experience has been gained with frequently-repeated regimens. A few randomized trials have demonstrated a benefit for survival on standard chemotherapy in small-cell lung cancer and advanced colorectal cancer. However, this benefit appears to be weak. The combination of dose-dense chemotherapy regimens with new anti-cancer strategies based on our insights into the mechanisms of oncogenesis is a challenge on the eve of the millennium. Key words: chemotherapy, dose, dose-density, dose-intensity, high-dose chemotherapy Introduction Chemotherapy is used in a wide spectrum of neoplasms. Historically, the treatment schedule and timing of chemotherapy was dictated by neutrophil recovery kinetics. With most myelosuppressive agents used alone or in combination, peripheral granulocyte nadirs are reached in about days and recovery is achieved in about 21 days [1]. Except in the case of nitrosoureas, these data have since been extensively substantiated. Bone marrow stem-cell proliferation is at a maximum when the granulocyte nadir is reached and recycling of chemotherapy at that point, and particularly of cycle-specific agents, may lead to considerable hematopoietic damage. It was decided that treatment intervals should be prolonged and, since the 1960s, it has been a rule of thumb to design most chemotherapy combinations to be repeated every three or four weeks. This applies to the bestknown regimens such as MOPP, CHOP, CMF, FAC and PVB. The interval was often fixed in weeks and not in days, for practical reasons. However, the possibilities of repeating therapy between the tenth and twenty-first days were not assessed. More recently, certain authors have developed new theories derived from basic concepts of chemotherapy, to optimize the practical use of this anticancer tool. First, Goldie and Coldman hypothetized that the emergence of drug-resistant clones within tumours was genetically programmed [2]. They recommended the use of a dose-intensive strategy to circumvent cell mutations. Their theory was that this could be obtained by rapidly alternating the treatment plan to preclude dose-limiting toxicity, so that the intervals between cycles could be reduced. Both this theory and the classic log-kill model [3] were challenged by Norton and Simon [4]. These latter authors agreed that a rapid reduction of the tumor

2 134 burden could best be achieved if the entire treatment was administered over as short a period as possible. They developed a mathematical kinetic model of the effective level of therapy to support these assumptions, later designated the Norton-Simon model. However, contrary to Goldie and Coldman, they recommended a sequential administration of chemotherapy instead of an alternated schedule. This theory was subsequently confirmed by a randomized Italian trial which found a remarkable advantage in a sequential as opposed to an alternating design of a regimen comprising doxorubicin and CMF in stage II breast cancer [5]. Both Goldie's and Norton's models outlined the short timing of recycling. Shortening the interval between chemotherapy cycles could be achieved in a number of ways. Firstly, this could be done without additional therapies by administering full conventional doses or even higher doses after a short interval, whatever the granulocyte count at the time of recycling. This has been performed safely in advanced breast cancer [6]. Secondly, hematopoietic growth factors (HGF) can provide support. This strategy has been developed worldwide, but to date there is no evidence of an improvement in neutropenia-related mortality with the use of cytokines [7]. The third method consists in reinjecting autologous peripheral blood stem cells after some or all of the cycles. One of the major advantages of this approach is that not only is the time to granulocyte recovery reduced, but also that of platelets and red cells counts, compared to the HGF currently available [8]. Finally, the current and future development of chemoprotectors may enhance tolerance and further, reduce the interval between courses of chemotherapy. Of note, some anti-tumor compounds have been developed with a weekly schedule. They include old drugs such as bleomycin, vincristine, methotrexate and 5-fluorouracil (5-FU), which tend not to give rise to myelotoxicity, or if so, only in a moderate form, but also, and more recently, drugs such as vinorelbine and gemcitabine, which are consistently myelotoxic. Shortening the gap between chemotherapy cycles has been tried and developed for about 15 years in efforts to increase the 'dose-intensity' of cytotoxic agents. This concept has been actively developed in the clinic, because a dose-response curve has been demonstrated for many agents in vitro and in vivo in animal models. The dose-intensity of chemotherapy can be increased by higher drug doses, by reducing the time between cycles, or both. In the clinic, most investigators have focused on increasing dose-intensity by using massive doses of chemotherapy with bone marrow or peripheral stem-cell transplantation as consolidation therapy. This strategy, however, has not been translated into a gain in overall survival in the few solid tumors for which randomised studies are available [9, 10]. Some authors believe that rapidly-repeated high-dose chemotherapy would afford higher, more durable complete remissions than a singlecycle high-dose regimen given as consolidation [11]. Compared with this latter strategy, the theoretical advantages of a rapidly-repeated approach are 1) the early increase in dose-intensity proven decisive in some neoplasms [12], 2) a decrease in mortality associated with the use of massive doses that has resisted attempts to obtain rates below 5% in many experiences, and 3) a decrease in the cost of treatment management, since hospitalization in a single isolated room is no longer required [13]. This is a review of the results reported recently with shortened intervals between courses of chemotherapy in solid tumors. Methods As recommended [14], the literature was reviewed according to published methodology quality criteria. Mulrow's recommendations for reviewing medical articles were used [15]. These criteria include: purpose, data identification, data selection, validity assessment, qualitative synthesis, quantitative synthesis, summary, future directions. Data identification Data were obtained by the two investigators using the following procedures: - a computer-assisted Medline search covering a 10-year period using the key words 'dose-intensity', 'dose-dense', 'densified' and 'chemotherapy'; - the Proceedings of the American Society of Clinical Oncology (ASCO) were hand-searched from 1993 to 1998 (6-year period) to identify dose-dense regimens; - the abstracts of the 1994, 1996 and 1998 issues of the European Society of Medical Oncology (ESMO) meetings were handsearched to identify dose-dense regimens, - the subject index of the Annals of Oncology, Cancer, European Journal of Cancer, Journal of Clinical Oncology and Journal of the National Cancer Institute were systematically examined from the January 1993 through the December 1998 issues for the following key-words: 'dose-density', 'dose-intensity', 'highdose chemotherapy', 'chemotherapy'. Data selection Chemotherapy regimens recycled with intervals of less than three weeks were identified. Articles or abstracts reporting fewer than 20 patients were excluded from the analysis, except when rare neoplasms were the subject of the report Meeting abstracts, which did not result in a full publication during the three succeeding years, were not selected for analysis. In addition, reports on hematological malignancies were excluded and the review was sub-entitled 'A critical review of dose-dense chemotherapy in solid neoplasms'. The reports were classified according to the tumor type (e.g., breast cancer) and the remaining articles were grouped together in the chapter 'miscellaneous neoplasms'. Validation assessment The assessment of the data was standardized as follows: - randomized studies on dose-density were described extensively and were submitted to a critical analysis appearing first in each chapter; - phase II or feasibility studies were described more succinctly; - patient selection criteria were examined; - toxicity-related deaths were mentioned, when they were reported by authors; - the dose-intensity received or the relative dose-intensity were mentioned when reported. Qualitative synthesis We attempted to point out limitations and inconsistencies found in reports in each chapter. Summary and future directions Data are summarized in the abstract and in the conclusion.

3 135 Table I. Randomised studies of dose-dense chemotherapy in small-cell lung cancer (SCLC). Author, year [reference] Study type Main selection criteria No. of patients included Intensive regimen Interval Received dose-intensity Response and survival Sculier [16] Souhami, 1994 [17] Steward, 1998 [18] Furuse, 1998 [19] Thatcher, 1998 [20) Different regimens compared Different regimens compared (PE/IA versus CAV/PE) Same regimen in the two arms Interval four versus three weeks Different regimens compared (CODE versus CAV/PE) Same regimen in the two arms Interval three versus two weeks Karnofsky index ^ 60 Age < 75 years Limited SCLC or extensive SCLC with a good prognosis Age < 75 years Good or intermediate Prognosis SCLC according to Cerny Age ^ 70 years Extensive SCLC Age < 75 years PS 0-2 SCLC with PS Doxorubicin 25 mg/m 2 dl Etoposide 120 mg/m 2 dl Cyclophosphamide 0.5 g/m 2 dl Vindesine 3 mg/m* d8 Cisplatin 60 mg/m 2 d8 Vincristine 2 mgdl5 MethotrexatelOO mg/m 2 dl5 + leucovorin No growth factor Cisplatin 50 mg/m 2 dl Etoposide 75 mg/m 2 dl. 2 Ifosfamide 2000 mg/m 2 d8 Doxorubicin 25 mg/m 2 d8 No growth factor VICE + GM-CSF Ifosfamide 5000 mg/m 2 dl Carboplatin 300 mg/m 2 dl Etoposide 120 mg/m 2 i v. dl mg/m 2 p.o. d3 Vincristine 0.5 mg/m' dl5 CODE regimen + G-CSF Cisplatin 25 mg/m 2 Vincristine 1 mg/m 2 (w 1,2,4,6) Doxorubicin 40 mg/m 2 (w 1,3.5.7) Etoposide 240 mg/m 2 (w 1,3,5,7) ACE regimen + G-CSF: Doxorubicin 40 mg/m 2 Cyclophosphamide 1000 mg/m 2 Etoposide 120 mg/m 2 I v mg/m 2 po. 1 week dl = d22 1 week dl = dl5 3 weeks 1 week 2 weeks Significantly lower in the 'intensive' arm (weekly timing not respected) Dose-intensity difficult to compare (different drugs in the two arms) Significantly better dose-intensity in the intensive (3-week) arm Dose-intensity twice as good in the CODE arm for drugs common to both arms Increased by 33% in the fortnightly arm Significantly better complete response rate than that of standard arm in patients with localised disease No benefit in survival for the weekly regimen No difference in response rates No difference in survival No difference in response rates Significantly better survival in the three-week arm than in the four-week arm (/> = ) No difference in complete response rates Trend towards better survival No difference in response rates Significantly better survival in the dosetense arm (P = 0.04) Results Dose-density in small-cell lung cancer (SCLC) A question that persists in oncology is whether a higher dose-intensity of chemotherapy can increase survival in SCLC [12, 16-23]. Studies of dose-intensity in SCLC have produced conflicting results: some authors showed no correlation with survival, whereas others suggested that survival could potentially be improved by increasing dose-intensity. The major importance of high-dose cisplatin and cyclophosphamide during the first chemotherapy cycle has been demonstrated in a randomised trial [12]. To our knowledge, five randomised studies have compared dose-dense regimens to conventional chemotherapy [16-20]. Their results are summarised in Table 1. First, European investigators reported that the adjunction of vindesine and cisplatin on day 8 and combination vincristine-methotrexate on day 15, to a three-drug combination (doxorubicin, cyclophosphamide and etoposide), repeated every three weeks had no effect on survival [16]. These disappointing results could have been due to the lower dose-intensity of the intensive regimen compared to that of the standard regimen, because of longer delays between courses in the former. Furthermore, most toxicity-related deaths were observed in the intensive arm. Basically this regimen was not feasible because intervals were longer and the toxicityrelated death rate was high. A second trial undertaken by a British team also compared a weekly to a three-weekly regimen. Patients did not receive prophylactic hematopoietic growth factors, nor had they in the previous trial. Results regarding dose-intensity are difficult to assess, as some drugs were used in only one of the two arms. No difference was detected in response nor in survival in this large study which included 438 patients [17]. More recently, a third trial conducted by Europeans focused on reducing time intervals in the V-ICE regimen (ifosfamide, carboplatin, etoposide and vincristine) from four to three weeks by incorporating GM-CSF. Three hundred patients were enrolled and the final results were recently published [18]. They show that the three-week schedule allows an increase of 26% in dose-intensity received, compared to the four-week arm. This increase in dose-intensity yielded a significant improvement in survival (P = ) without increased toxicity. Twoyear survival was 33% and 18% in the intensive arm and the standard arm, respectively. This difference remained significant after adjustment for prognostic factors. Surprisingly, the complete response rate and median dura-

4 136 tion of response were similar in the two arms and thus, the mechanism by which dose intensification conferred the survival benefit is unclear. A Japanese randomised trial recently compared the weekly CODE schedule + G-CSF to the standard CAV/PE regimen [19). Although the dose-intensity received in the weekly arm was twice that of the standard arm, there was no difference in complete response rates. The two-year survival rate was slightly better in the intensive arm but short of statistical significance (P = 0.10). However, the exclusion of patients with localized SCLC, the number of toxicity-related deaths and the fact that this trial was not powerful enough to detect a difference in survival of less than 15% could account for lack of a significant impact of the densification of chemotherapy on survival. The latter randomised trial was recently reported by the MRC Lung Cancer Working Party [20]. Patients were randomised to receive six cycles of a doxorubicincyclophosphamide-etoposide combination, repeated fortnightly (with G-CSF) or three-weekly. Short intervals were respected mainly during the first cycles in the dose-dense arm. No difference between response rates was detected. Patients receiving the dose-dense arm obtained a significantly higher overall survival rate (P = 0.04). One-year survival was 47% in the intensive arm compared to 39% in the control arm. The final report from this study is pending. Among these five randomised studies, the three which increased the doseintensity by reducing the interval showed an improvement in survival (statistically significant in two). There are continual debates as to whether dose-density, dose-intensity and survival of patients with SCLC might be increased by the use of hematopoietic growth factors. A survival benefit has been reported in some studies [18, 20, 22] but not in others [19, 21, 23, 24] when either G-CSF or GM-CSF were used to increase doseintensity. In the Japanese randomised trial that investigated whether the adjunction of G-CSF could improve the results of the CODE regimen, a survival benefit was reported [22]. However, this was a very small series and the survival of patients who received the standard CODE regimen was less than that originally reported with this schedule [25]. This signifies that the survival benefit reported with the CODE regimen combined with G-CSF by Fukuoka et al. may be artificial. Some authors claim that the beneficial effects of adding G-CSF to dose-dense regimens may be limited to regimens alternating hemato-toxic and non-hemato-toxic regimens [21, 22, 24]. Of note, a recent trial has shown that G-CSF administered up to 48 hours before the next chemotherapy course increases the risk of chemotherapyinduced neutropenia and thrombocytopenia [26]. The authors recommend increasing the window of time between the discontinuation of G-CSF and re-initiation of chemotherapy after the 48 hours in rapidly-repeated regimens. Some phase II studies have also tested short-interval regimens in SCLC. These studies mainly report the clinical development of experimental schedules that are then used as investigational intensive arms in the abovecited phase III randomised studies. Murray et al. designed the CODE regimen, a weekly alternation of myelosuppressive and non-myelosuppressive treatments [25]. They reported an encouraging 30% two-year survival rate in selected extensive SCLC with acceptable toxicity. The CODE regimen was recently used as salvage therapy in 17 patients with relapsed SCLC and a median time off first-line chemotherapy of 6-7 months. An objective response was achieved in 15 of 17 (88%), and 5 were complete. No treatment-related deaths were observed [27]. Similarly, the Southwest Oncology Group (SWOG) has developed a weekly short-duration regimen with interesting results in limited- and extensive-stage SCLC [28]. Toxicity was acceptable and included one toxicity-related death in 76 patients. Investigators from the EORTC have attempted reducing intervals of the CDE regimen (cyclophosphamide, doxorubicin, etoposide) from three to two weeks with GM-CSF. They were able to increase the dose-intensity with GM-CSF only during the first courses [29]. Dose-density in germ-cell tumors and trophoblastic gestational tumors Although germ-cell tumors (GCT) are the most chemosensitive solid neoplasms, only one randomized study on dose-density has been reported to date [30]. Randomized trials focusing on dose-intensity have failed to show any benefit of double-dose cisplatin [31] or intensive chemotherapy with autologous bone marrow transplantation [9] in poor-risk patients. Investigators from the Medical Research Council (MRC) and the EORTC very recently reported the final results of a phase III trial comparing a rapidly-repeated intensive regimen (BOP/ VIP-B) to the standard 4 BEP in 380 patients with poorprognosis non-seminomatous GCT (NSGCT) [30]. The BOP/VIP-B regimen is a complex combination of bleomycin, vincristine, cisplatin, etoposide and ifosfamide (Table 2). This schedule was initially reported to be highly efficient in 91 patients treated in an uncontrolled study reported in 1991 [32]. Unfortunately, the intensive regimen in this randomised study afforded no complete responses nor any survival benefit. The higher number of toxicity-related deaths in the intensive arm (19 8) indicates that this regimen is in fact not readily feasible in a multicentre study. A higher proportion (10 5) of primary mediastinal NSGCT, notorious for their dismal outcome, in the intensive arm, is another plausible explanation for these negative results. A final explanation is that shortening cycles of chemotherapy may not enhance its effects in patients with germ-cell tumours. However, non-randomised trials testing dose-dense chemotherapy have yielded very promising results. These trials are summarised in Table 2. One of the oldest dose-dense schedules was the British POMB/ACE regimen [33] consisting of a complex combination of seven drugs, repeated alternatively

5 137 Table 2. Dose-intense chemotherapy in germ-cell tumors. Author, year [reference] Selection criteria Study type Number of patients Regimens Survival Dearnaley, 1998 [36] Stage I NSGCT Vascular invasion 115 BOP biweekly (bleomycin. vincristine. cisplatin) Disease-free survival: 98% (median follow-up: 14 months) Wettlaufer [35] Advanced NSGCT No clear selection for prognosis 29 BOP weekly (bleomycin. vincristine. cisplatin) Disease-free survival: 83% (median follow-up: 31 months) Harstrick [39] Advanced GCT (including seminomas) German criteria for poor-risk patients 48 PEBOI biweekly (cisplatin, etoposide. vincristine, bleomycin/ifosfamide, vincristine, bleomycin) Three-year overall survival: 76% Three-year disease-free survival: 61% Lewis [32] Advanced NSGCT Criteria for poor-risk GCT closed to those of the MRC/ EORTC 91 BOP/VIP (10 days/3 weeks) (bleomycin. vincristine, cisplatin/etoposide, ifosfamide. cisplatin) Two-year overall survival: 69% Two-year progression-free survival: 66% Germa Lluch, 1992 [40] Advanced NSGCT Hospital de la Santa Creu l Sant Pau classification for poor-risk patients 20 BOMP/EPI (14 days/21 days) (methotrexate, vincristine, bleomycin, cisplatin/etoposide. ifosfamide, cisplatin) Three-year progression-free survival: 85% Bower, 1997 [33] Advanced NSGCT IGCCCG classification: - good risk - intermediate risk - poor risk POMB/ACE biweekly (cisplatin, vincristine, methotrexate, bleomycin/ d-actinomycin, cyclophosphamide, etoposide) Three-year overall survival 97% 88% 75% Horwich [38] Advanced NSGCT IGCCCG classification: - intermediate risk - poor risk C-BOP-BEP + G-CSF (carboplatin, bleomycin. vincristine, cisplatin, etoposide) Three-year overall survival 68% 91% Amato, 1996 [42] Advanced NSGCT - First-line therapy - Poor risk (M.D. Anderson classification) 22 BOP-CISCA-POMB-ACE + G-CSF No data on survival Kaye, 1998 [30] Advanced NSGCT Poor risk (MRC critenas) P 380 BOP/VIP-B ± G-CSF versus BEP/EP No difference in survival Motzer, 1997 [44] Poor-risk first relapse 24 Paclitaxel-ifosfamidecarboplatin-etoposide + G-CSF + peripheral blood stem cells Eleven of twenty-four (46%) patients Continuously disease-free after 11 months Culine, 1995 [45] Second or third relapse study 14 VB4-BOP-CISCA (vinblastine, bleomycin/bleomycin, vincristine, cisplatin/cisplatin, doxorubicin, cyclophosphamide) Three of fourteen (21%) patients continuously disease-free after 18+, 26+, 36+ months a Two different regimens compared. at fortnightly intervals. Results obtained in 339 patients show that the three-year overall survival of patients with a poor prognosis according to the International Germ- Cell Cancer Collaborative Group (IGCCCG) criteria [34] was 75%. This compares favorably with the 50% reported by the IGCCCG. However, the toxicity of this regimen has never been reported in detail and 1.4% of patients have died of its toxic effects. In 1984, investigators from the University of Colorado reported the results of a weekly cisplatin-vincristinebleomycin regimen (BOP). Although the activity of single-agent vincristine in GCT is unknown, it was substituted for vinblastine in order 'to be able to give chemotherapy at much closer intervals'. Twenty-nine patients with advanced NSGCT were included and obtained a disease-free survival rate of 83%. The authors concluded that the BOP regimen was less toxic and as effective as the PVB combination [35]. However, these results are difficult to interpret because the number of patients is low and they were not classified according to

6 138 prognosis. Subsequently, the BOP regimen was more often incorporated into alternating or sequential intensified regimens than used alone to treat advanced GCT. Dearnaley et al. reported the results of two cycles of a biweekly BOP regimen in a cohort of 116 patients with stage I NSGCT and at a high risk of relapse. The twoyear survival rate was 98%; toxicity was mild and compared favorably with that caused by the two cycles of BEP often used in this setting in Europe [36]. If these results are corroborated with a longer follow-up, two cycles of biweekly BOP should be considered a possible option in high-risk stage I NSGCT. Another accelerated induction regimen, the so-called C-BOP-BEP, was developed at the Royal Marsden Hospital [37, 38]. Courses are scheduled weekly (Table 2) but the results show that this interval could not be sustained. Toxicity was high with 7% of toxicity-related deaths. Again, survival results are very promising in the 41 poor-risk patients with a three-year OS of 91% (compared to 50% in the IGCCCG cohort). Surprisingly, results in the 27 intermediate-risk patients were more disappointing, as the three-year OS rate was only 68% (compared to 81% in the IGCCCG cohort). The excess number of toxicityrelated deaths may account in part for the mediocre results in this group. This regimen is now being evaluated in a phase II study by the EORTC. The German Cooperative Group for Testicular Cancer has reported its experience with a rapidly alternating combination (Table 2). The three-year disease-free survival and overall survival rates in poor-risk NSGCT according to this group's classification were, respectively, 61% and 76%. There were no deaths related to the intensive chemotherapy. However, two patients died of bleeding after resection of residual masses [39]. Similarly, investigators at the University of Barcelona have also reported results obtained with an intensive alternating multi-drug schedule, the so-called BOMP/EPI (Table 2). No hematopoietic growth factors were given during this period. Twenty patients with high-risk NSGCT according to this institution's classification were treated with this regimen, and 18 of them (85%) remained disease-free over the long term. Their results were compared retrospectively with those obtained by the same investigators with the PVB regimen: there was a trend towards a better survival rate. Caution should be exercised when interpreting these results because the PVB regimen is no longer considered the standard therapy for high-risk GCT [41], and secondly because a historical comparison may be biased. Investigators from the M.D. Anderson Cancer Center have also partially reported an interesting experience with dose-dense first-line chemotherapy in patients with poor-risk GCT. Combining the CISCA regimen developed at this institution, the British POMB/ACE and the BOP regimens, they used a dose-dense regimen in which courses were repeated every 7-14 days. Preliminary results of this phase II study show that this regimen is feasible without toxicity-related deaths. Strict recycling criteria were used : granulocyte count >1000/mm 3, platelet count ^ 50,000/mm 3 and grade 3-4 toxicity resolved. The preliminary results on response are promising (82% of complete responses after chemotherapy, with or without surgery) and the final results of this phase II study, including survival results, are now awaited [42]. These regimens ought to be investigated in phase III randomized studies in order to determine whether they are truly superior to the standard four cycles of BEP in poor-risk patients. Moreover, the wide variation between these different reports in their definition of patients with poor-risk NSGCT precludes an adequate analysis of data. Bajorin et al. demonstrated that this type of bias may have a major impact on the results of a non-randomised trial [43]. IGCCCG criteria [34] should now be used worldwide to design prospective trials in poor-risk GCT. The Memorial Group has developed a dose-dense and dose-intensive paclitaxel-based regimen for salvage strategies (Table 2) [44]. Preliminary results have been reported for 24 patients with a poor-risk relapse of GCT. Forty-six percent of them were alive and free of disease after a median follow-up of eleven months. A higher accrual is warranted but these results seem promising. Finally, a dose-dense intensive approach has proven capable of curing some patients who were highly pretreated with at least two previous lines of chemotherapy [45]. Using a complex combination of non-cross-resistant drugs recycled every week, Culine et al. obtained 8 of 14 complete responses; of these, 3 patients continue to be disease-free with an extended follow-up. This schedule is currently being investigated in a phase II multicenter French trial. Increasing the dose-intensity of cisplatin combined with sodium thiosulfate by shortening the interval between cycles has recently been validated [46]. Using a biweekly schedule, a dose of 180 mg/m 2 of cisplatin every two weeks was shown to be safe and feasible. Two dose-dense regimens have been reported to obtain an 86% complete response rate in high-risk metastatic trophoblastic gestational tumors [47, 48]. The EMA/CO regimen is a combination of etoposide, methotrexate and actinomycin D alternated with cyclophosphamide and vincristine initially developed by Bagshawe [47]. When data from other centers were reviewed, a 73% complete remission rate was demonstrated in 104 patients [49]. As the combined activity of cyclophosphamide and vincristine was questionable in the EMA/CO regimen, investigators at the Institut Gustave-Roussy developed a biweekly combination of cisplatin, actinomycin D and etoposide (APE). The rationale for this regimen was also based on the wellknown activity of cisplatin in male choriocarcinoma. Preliminary results obtained in 19 patients showed an 86% complete response rate without life-threatening toxicity [48]. Definitive results are imminent. To date, a dose-dense chemotherapy regimen (EMA/CO or APE) is the current standard for poor-prognosis metastatic trophoblastic gestational tumor.

7 Dose-density in early breast cancer Unlike those undertaken for advanced breast cancer, few randomized trials have evaluated the relevance of adjuvant or neo-adjuvant dose-dense regimens in early breast cancer. The feasibility of accelerated chemotherapy has been reported to be excellent in these patients [50], Only one randomized trial performed by the Eastern Cooperative Oncology Group (ECOG) has been reported to date in the adjuvant setting [51]. This study included 646 patients with node-positive and hormone receptor-negative breast cancer. Patients randomized in the experimental arm were to receive a weekly regimen combining cyclophosphamide, 100 mg/m 2 /day (d) (d 1-7), doxorubicin, 40 mg/m 2 dl, methotrexate, 100 mg/m" dl (+ leucovorin d 1 and 2), vincristine 1 mg dl and fluorouracil 600 mg/m 2 d2 (odd weeks), and continuous infusional fluorouracil, 300 mg/m 2 d 1 and 2 (even weeks), administered over 16 weeks. Women included in the control arm received six cycles of a CAF regimen (with oral cyclophosphamide, 100 mg/m 2 /d, dl 14). A total of 323 patients were included in each arm; 95 recurrences and 71 deaths occurred in the 16- week arm 105 and 86, respectively, in the control arm. A trend was seen indicating an improvement in four-year disease-free survival with densified treatment (67.5% 62.7%, P ) but overall survival was not significant (two-sided test). The median dose intensities were 16.9 mg/m 2 /w versus 12.5 mg/m 2 /w for doxorubicin, and 209 mg/m 2 /w versus 541 mg/m 2 /w for fluorouracil but were not increased for cyclophosphamide. It is still not clear whether such a modest increase in dose-intensity can contribute to these results, and endpoints other than dose-density, such as the effect of continuousinfusion fluorouracil or antimetabolite sequencing, have to be considered. No treatment-related deaths occurred in the 16-week arm versus three in the CAF arm (including one death due to congestive heart failure and two sepsis-related deaths). The quality of life of patients deteriorated significantly in the densified arm. Due to these facts, the authors do not recommend the 16- week regimen as a standard adjuvant treatment but their results warrant further investigations of densified schedules. The results of a randomised trial of neoadjuvant therapy in locally advanced breast cancer have been reported very recently in abstract form [52]. This large trial merits special attention for it is one of the few comparing an optimal control arm, based on the most recent criteria, to a substantially densified arm. The primary endpoint of this study was progression-free survival but data on response rates or breast preservation were not presented in this meeting abstract. This study included 448 patients who were randomized to receive either 6 cycles of CEF (cyclophosphamide, epirubicin, 5-FU) repeated every 4 weeks or a combination of epirubicin, 120 mg/m 2 and cyclophosphamide, 830 mg/m 2 + G-CSF, repeated every 2 weeks, allowing an increase of 119% and 79% in the median dose-intensity 139 for epirubicin and cyclophosphamide, respectively. Febrile neutropenia occurred in 14% of patients in the standard arm versus 8% in the densified arm, and a decrease in the left ventricular ejection fraction in 9% versus 12%. No toxicity-related deaths were reported. Unfortunately, with a median follow-up of 27 months, the progression-free survival is similar in the 2 treatment groups. Of note, the densified treatment results were tainted by a significantly lower quality of life score, but only during the first three months, and this score returned to baseline levels earlier in this group. Another smaller study (112 patients) comparing FAC repeated at 21-day versus 18-day intervals [53] showed significantly improved response rates at the expense of significantly increased toxicity (febrile neutropenia, mucositis, thrombocytopenia) without evidence of improved breast preservation rates. Moreover, the pathology-complete response rate was similar in both arms and there was no significant difference in disease-free survival at a very early stage of follow-up (Table 3). Results of biweekly adjuvant anthracycline-cyclophosphamide combinations have been reported in three prospective studies [54-56] which demonstrated excellent feasibility with a dose increment of 50%-100% [54] at the expense of significantly impaired quality of life [55]. With epirubicin at 75 mg/m 2 + cyclophosphamide 1200 mg/m 2 q 14 days, Cuvier et al. reported a 75%> twoyear disease-free survival rate in 59 patients with more than 5 involved nodes [56]. The demonstrated efficacy of taxanes in metastatic breast cancer has motivated exploration of new rapidlyrepeated adjuvant combinations containing these drugs. Most of the experiences reported concern paclitaxel. The largest experience was reported by the Memorial Sloan Kettering Cancer Center [57-60] which developed a sequential high-dose and biweekly-recycled combination of doxorubicin, paclitaxel and cyclophosphamide. The long-term distant disease-free survival yielded by these pilot studies published recently seems promising, with 81% after a median follow-up of 39 months in 42 patients with a median of 8 involved axillary nodes [60]. Data on doses and scheduling provided by these two trials [57, 60] were used to design a randomized study comparing the sequential biweekly-scheduled regimen to standard adjuvant chemotherapy which is ongoing. Several pilot studies of locally advanced breast cancer patients treated with neo-adjuvant dose-densified chemotherapy have been published to date [61-64] (Table 4); some of them included patients with advanced breast cancer and will be detailed further on. In summary, pathology-complete response rates were not enhanced in three series [61, 63, 64]. Moreover, evidence of better breast preservation rates is still lacking - 32% [64] to 80% [61]. In inflammatory breast cancer, a single study of 48 patients who received a biweekly combination of epirubicin at 75 mg/m 2 and cyclophosphamide, 1200 mg/m 2, without G-CSF [6] obtained a four-year overall survival rate of 66% which compares very favourably with literature data. The planned dose intensity was

8 140 Table 3. Early breast cancer: randomized studies with densified regimens. Author. year [reference] Study type Main selection criteria Number of patients included Regimens (mg/nr) Interval Received dose-intensity Response and survival Fetting. Adjuvant Node-positive, 1998(51] Different regimens ER-negative compared patients Therasse [52] Dhingra, 1996 [53] Gundersen, 1986(66] Richards, 1992 [67] Blomqvist [68] Ardizzoni. 1994(69] Lahsang, 1997(72] Neoadjuvant Different regimens compared Neoadjuvant Same regimens compared with higher doses in densified arm Single-agent polychemotherapy Same drug in both arms Same regimen with 1 /4 doses in weekly arm Same regimen in both arms Randomized study comparing intensified and dose-escalated chemotherapy Endpoint: dose-intensity Locally advanced breast cancer Locally advanced breast cancer Metastatic breast cancer Metastatic breast cancer Metastatic breast cancer Metastatic breast cancer Metastatic breast cancer Bonadonna. Randomized Metastatic breast 1997(70] study comparing cancer standard dose-escalated intensified regimen Fountzilas. Same single agent Metastatic breast (73] in both arms cancer Joensuu [74] Accelerated single agent standard polychemotherapy Metasatic breast cancer C 100 (d 1-7), A 40, M 1 week 100, V I, a F 600 alternated with F 300 d weeks C100(dl-14),A60, F600 C 500, E 60, F weeks E 120, C G-CSF 2 weeks F 1000, A 50, C weeks F 1200, A 60, C days G-CSF A 20 V 2," A 50, C 600 A 25 vs A 75 1 week 3 weeks 1 week 3 weeks F125, E15, C125 1 week vs F500, E60, C500 4 weeks F600. E 60, C days GM-CSF vs 3 weeks F600, E60, C600 E 75. C500 + G-CSF 1-2 weeks E120 to 180, C700 to 1000 vs + G-CSF 3 weeks F600, N10, C600 F800, N14, C800 + G-CSF F600. N10. C600 + G-CSF El 10 +G-CSF El 10 +G-CSF 303 E20 E60, C500, F500 x 3 3 weeks 3 weeks 2 weeks 4 weeks vs 2 weeks Increased for D and F, not for C Comparison made difficult because of the different regimens Higher in densified arm Higher in densified arm Impossible to assess accurately Identical in the two arms Identical in the two arms Higher in densified arm Significantly higher in densified arm Increased in both experimental arms Two-fold higher in densified arm 1 week Comparison made difficult because of the 3 weeks different regimens Trend towards improved four-year survival (67.5% 62 7%, P = 0.063) Impaired quality of life with weekly schedule No significant improvement in progression-free survival Significantly improved response rate (98% 76%) No difference in pathologic complete response rates and disease-free survival No difference in response rates and overall survival No difference in response rates and overall survival Significantly worse response rates and overall survival in intensified arm Trend towards improved response rates Non-significant preliminary results No significant improvement of response rates or survival Higher rate of complete responses in densified arm No significant difference in survival between two arms Less toxicity and better quality of life with singleagent therapy Trend towards a longer duration of response with single-agent therapy Abbreviations: A doxorubicin. E epirubicin: C - cyclophosphamide: F - fluorouracil: M - methotrexate; N - mitoxantrone; V - vincristine: all doses are expressed in mg/m 2 except. a total dose; b CR rate is significantly higher: - data not available: NS - non-significant OS - overall survival admirably attained with dose intensities of 102% and 97% at the fourth and sixth cycles, respectively. These particular subgroups of rapidly growing malignancies may represent a good indication for densified schedules. A randomized study is the sole means of demonstrating the superiority of dose-dense regimens but in view of the rarity of this entity, such a study would be difficult to complete. In summary, with three negative randomized trials published to date, evidence is still lacking as to whether

9 141 Table 4. Intensified regimens in breast cancer: non-randomized studies. Author, year Study Main selection Number of Regimens (mg/m 2 ) [reference] type criteria patients included Objective responses (n) Overall survival Culine [6] Phase 11 T4d 48 C E 75 q 14 d without CSF 26 (4 CR) 66% at 4 years Extra [79] Metastatic at diagnosis 81 C 1200, E 75 q 14 d without CSF 45(12CR) 37% at 4 years Piccart, 1995 [78] Locally advanced and metastatic LA: 11 M: 18 C83O, E 120 q 13 to 22 d +CSF LA:10(3CR) M: 15(5CR) Scinto, 1995 [77] Bernardo, 1995 [61] Breier [64] Graham [62] Locally advanced and metastatic Locally advanced 36 Locally advanced 32 Locally advanced 34 LA: 19 M:22 C600, E 120 q 14d + CSF M14(8CR) LA: I9(6CR) F600, E65, C600q 14d + CSF 36 C800, E 100 q 15d + CSF 28 (3 CR) A90q 17d + CSF Followed by CMF x 4 with dose escalation % at 1 year Zelek, 1997 [80] Metastatic relap- 70 ses C E 75 q 14 d without CSF 40 (12 CR) Median = 36 months Cazap [81] Locally advanced 30 and metastatic F600, E70, C600ql4d + CSF 24 (8 CR) Cottu [82] Metastatic 26 C 3000, E 100 q 14 d + CSFand PBSC 14(3CR) 85% OS at 2 years Lahsang, 1996 [83] Metastatic 24 P135, E75q lod + CSF McCaskill-Stevens, 1996 [84] Metastatic 25 P 90, CDDP 60 q 15 d without CSF 15 (0 CR) Luck, 1997 [88] Metastatic 31 P 60 to 90 q 7 d without CSF Sikov [89] Locally advanced and metastatic M:9 LA: 12 P175q7d + CSF M:8(2CR) LA:10(2CR) Blohmer, 1998 [63] Locally advanced 24 (LA) D75, A50q 14d + CSF 17(3CR) Cuvier, 1998 [90] Metastatic 22 (M) D60-66,C q 14d without CSF 6(1CR)/ 11 evaluable patients Seidman, 1998 [87] Metastatic 30 (M) P 100 q 7 d without CSF 16(3CR) Abbreviations: M - metastatic; LA - locally advanced; CSF - colony-stimulating factor; PBSC - peripheral blood stem-cell support; A doxorubicin; C - cyclophosphamide; CDDP - cisplatinum; D - docetaxel; E - epirubicin: F - fluorouracil; M - methotrexate; P - paclitaxel. adjuvant or neo-adjuvant densified regimens have an edge over optimal standard anthracycline-based regimens. Moreover, although feasibility is excellent, dosedensity does exert a detrimental effect on the quality of life reported in two of these trials. This is essentially why densified-regimens cannot be recommended as a standard adjuvant or neo-adjuvant therapy. Two points need, however, to be underlined and warrant further investigations:firstly,a trend towards improved progression-free survival is reported in the only randomized study on adjuvant therapy [51]. This must be appreciated in the light of the encouraging results reported in pilot studies including patients with massive axillary involvement [56, 60]. Secondly, breast preservation was the endpoint addressed for locally advanced breast cancer only in a single randomized study and the impressive local control rates achieved in some small pilot studies [61, 62, 64] have yet to be confirmed. Dose-density in metastatic breast cancer An entire body of randomised trials evaluating dosedensified chemotherapy are now officially available on metastatic breast cancer. Biganzoli and Piccart undertook a review of these results in a recent issue of Annals of Oncology [65]. Most of the studies, however, compared standard anthracycline-based regimens repeated every three weeks with a reduced dose per course and shortened cycle intervals [66-68, 74] and failed to demonstrate a significant benefit with accelerated chemotherapy (Table 3). Of special interest is the study by Blomqvist et al. in which the relative dose-intensity was similar in both arms [68]. The response and survival rates obtained in the accelerated chemotherapy group was significantly below those of the standard FEC arm. The population included in this trial was surprisingly low for the rate of measurable disease but it may indeed be an accurate reflection of the general population of patients with

10 142 metastatic breast cancer, (avoiding some of the selection biases commonly present in phase II studies). These results strongly support the hypothesis of an efficiency threshold per cycle, regardless of dose intensity. Three other trials [66, 67, 74] compared doxorubicin weekly versus every 21 days at a similar dose-intensity in each arm and failed to show any benefit with the weekly schedule. The aim of the most recent trial [74] was to compare a weekly single-agent versus combination chemotherapy repeated every three weeks: with standard FEC every 21 days, a trend was noted towards a longer duration of response (12 versus 10.5 months; P ) whereas overall survival was not significantly different. Only two studies have compared standard and reduced cycle intervals with a constant dose per course and conventional anthracycline-based regimens [69, 70]. Ardizzoni et al. [69] randomized FEC every 21 days versus FEC every 14 days: an increase of approximately 30% in dose intensity was achieved without life-threatening toxicity. Results, however, remained disappointing and the upgraded objective response and disease-free survival rates did not reach statistical significance. The second study was a three-arm randomisation between standard FNC (5-FU, mitoxantrone, cyclophosphamide) every 21 days versus every 14 days versus escalated FNC every 21 days. Results are far too preliminary for definitive conclusions [70]. Past experience with dose-escalation in metastatic breast cancer provides firm evidence that increasing dose intensity by a factor of does not necessarily improve treatment results [71]. The probability of obtaining a major benefit is very slim if cycle intervals are simply reduced without increasing the dose per course. Associating reduced intervals with moderate dose increments is, however, a more effective and safer way of increasing dose intensity than dose escalation alone. Incidentally, investigators from the Netherlands have shown that the interval between cycles can be reduced safely to 10 days, allowing better dose-intensity with less toxicity, compared to a three-week schedule with dose escalation [72], The only study that randomised biweekly versus monthly high-dose anthracyclines failed to show any survival benefit with dose-densified chemotherapy despite an improvement in the complete response rate [73]. The median dose intensity of epirubicin was doubled (27.2 versus 52.9 mg/m 2 /w) but there was no significant difference in the incidence of grade 3-4 toxicity between the two groups. Data from the literature show that complete response is highly correlated with prolonged survival [75]; the survival benefit in this study may therefore be restricted to this subgroup of patients and remains undetermined because of the limited number of patients. It is also noteworthy that this study used a single agent rather than a conventional combination. In summary, the randomised trials conducted in metastatic breast cancer do not report any major benefit from dose-densified chemotherapy (Table 3). Moreover, accelerated chemotherapy with reduced doses per cycle is probably detrimental and should no longer be recommended. However, the improved complete response rates and the trend toward better objective response rates and survival observed in two studies [69, 73] appears to support the hypothesis that dose-densified chemotherapy may be useful in a selected subgroup of patients. Further investigations are therefore warranted in selected subgroups. Many pilot studies of dose-dense chemotherapy in metastatic breast cancer have become available since the small series reported by Bronchud et al. in 1984 [76] and are summarized in Table 4 [6, 61-64, 77-90]. The safety and feasibility of densified anthracycline-cyclophosphamide combinations have been demonstrated and the toxicity reported in these studies, especially regarding neutropenic fever or congestive heart failure, is not greater than that observed with anthracycline-based regimens repeated every 21 days. Moreover, a large number of these studies [77-80] reported response or survival rates which compare very favourably with those in the literature. More recent studies have used weekly taxanes whose tolerance profile is acceptable, with no severe hematologic or neurologic toxicities [83-90]. Several of them, however, mixed locally advanced and metastatic disease and sometimes other cancers such as ovarian carcinomas [86] and thus obviate firm conclusions. Weekly administration appears to be a safe method of dose-intensity expansion with taxanes, whose efficiency must now be evaluated in larger trials. Dose-density in the Ewing's sarcoma family of tumors Ewing's sarcoma and peripheral primitive neuroectodermal tumors share common clinical and biological features and have been designated the Ewing family of tumors [106]. In unselected patient populations, fiveyear overall survival is about 55% when disease is localized and less than 10% in patients with metastasis at diagnosis. A few attempts have recently been made to increase these rates using dose-dense regimens. No randomized study has compared a dose-dense regimen to a standard regimen, because these neoplasms are rare. However, data reported to date with the use of dosedense regimens are encouraging. Kushner et al. have developed the P6 protocol at the Memorial Sloan- Kettering Cancer Center, a rapidly-repeated regimen combining doxorubicin, vincristine and high-dose cyclophosphamide (HD-DAV) and ifosfamide-etoposide regimen [107]. Courses were started once neutrophil and platelet counts had reached 500/ul and 100,000/ul, respectively. An excellent two-year event-free survival rate of 77% was obtained in patients with localized disease and adverse prognostic factors at presentation (tumor volume > 100 cm ). The regimen was shown to be feasible in a short period of time and most patients received the first three courses of chemotherapy within seven weeks. Toxicity was severe and included myelosuppression (one toxicity-related death), mucositis and peripheral neuropathy. Similar encouraging results were

11 obtained in patients with desmoplastic small round-cell tumors (DSRCT), a treatment-resistant cancer, with 5 of 10 disease-free survivors after a follow-up of about 1 year [108]. However, further follow-up is necessary for firm conclusions as to whether successful results of the P6 regimen will improve the cure rates of patients with Ewing's sarcoma and DSRCT. Another dose-dense strategy was developed at the Institut Gustave-Roussy. Patients with neoplasms of the Ewing's sarcoma family received two induction courses of doxorubicin, cisplatin and ifosfamide followed by the same regimen without cisplatin on day 15. Each course was repeated every four weeks with G-CSF support. Preliminary results have been reported and show that 15 of 16 patients, with localized disease and an axial primary or a bulky tumor in 14 of them, were rendered disease-free after a median follow-up of 24 months at this point in time. A mean interval of 15.7 days between cycles was reported [108]. These results compare favorably with the 77% two-year overall survival reported in a cohort of adults with localized Ewing's sarcoma, including those whose lesions were not axial or bulky primaries [110]. Dose-density in colorectal carcinoma In colorectal cancer, dose-densified chemotherapy has been used extensively. As hematological toxicity is limited, weekly or biweekly administration of 5-FU has been used since the 1980s. In 1994, Buroker et al. reported the results of a randomized study comparing weekly 5-FU plus high-dose leucovorin to the classic monthly Mayo clinic regimen. A non-significant trend towards a benefit was found in progression-free survival. However, the difference between treatments remained non-significant after adjustment with prognostic factors. The monthly regimen was reported to be associated with a lower financial cost and manageable toxicity with less need for hospitalization [111]. More recently, French investigators have compared a bimonthly high-dose leucovorin and fluorouracil bolus plus a continuous infusion regimen to the Mayo clinic regimen [112]. The De Gramont regimen (LV5-FU2) combines leucovorin 200 mg/m 2, bolus 5-FU 400 mg/m 2 and 5-FU 600 mg/m 2 as a 24-hour infusion given on two consecutive days and repeated every two weeks. This regimen has proven effective, yielding an objective response rate of 32.6%, a median progression-free survival duration of 28 weeks and a median survival of 62 weeks. The figures for the monthly regimen were, respectively, 14.4%, 22 weeks and 57 weeks. The difference was significant for the response rate (P = ) and progression-free survival (P ) and of borderline significance for survival (P ). Moreover, the bimonthly regimen was less toxic than the monthly regimen. Very recently, a randomized German trial suggested that patients with metastatic colorectal cancer receiving an intensive regimen could expect a significantly better survival. To date, the results of this trial have been reported only in 143 abstract form [113]. Patients were randomly assigned to receive either weekly folinic acid, 500 mg/m 2, followed by 5-FU, 2600 mg/m 2 as a 24-hour infusion, with a twoweek free interval after one month of therapy, or a standard Mayo Clinic regimen, repeated every month. This trial included 149 patients with untreated metastatic colorectal cancer. Arms were well-balanced with respect to prognostic factors. The median survival is 463 days in the intensive arm compared to 370 days in the standard arm. Of note, the median survival of the patients recruited in the standard arm is roughly the same as that usually reported in unpretreated colorectal cancer patients. The difference between curves was statistically significant. However the P-value was not reported. This difference was probably associated with a significantly smaller number of patients with early disease progression (19% and 48%, respectively). Many non-randomized trials using weekly or bi-weekly 5-FU have been reported, mostly in abstract form. To date, many European investigators consider dose-dense 5-FU-based regimens as standard therapy in advanced colorectal cancer. This is why such regimens have been combined with new active drugs, mainly oxaliplatin and irinotecan. The experience with regimens combining 5-FU, leucovorin and oxaliplatin has recently been reviewed extensively [114]. De Gramont has reported a small but significant benefit in time to progression with the adjunction of bi-weekly oxaliplatin to the LV5-FU2 regimen [115]. Currently, various new combinations of bi-weekly oxaliplatin, 5-FU or CPT11 are under study in phase I and II trials [114]. Whether the three-drug combination of thymidilate synthase inhibitors (mainly 5-FU and Tomudex), oxaliplatin and CPT11 recycled bi-weekly will be feasible and more effective than the two-drug combinations will be a new challenge for oncologists. Furthermore, we are now awaiting the results of dose-dense regimens in adjuvant colorectal cancer patients. Dose-density in gastric carcinoma At the end of the 1980s, the majority of investigators considered 5-FU alone or in combination with doxorubicin and mitomycin (FAM) to be standard therapy for advanced gastric cancer. The sequential combination of high-dose methotrexate and 5-FU alternated every two weeks with doxorubicin (FAMTX) was shown to yield encouraging results in phase II trials. Based on these data, a randomized study comparing FAMTX to FAM was performed by the EORTC and showed a significantly better response rate and overall survival (41% versus 22% at one year, respectively, P = 0.004) for the dose-dense FAMTX regimen [116]. Subsequently, FAMTX was compared to the etoposide, doxorubicin and cisplatin (EAP) combination given on days 1-8, repeated every four weeks [117]. No significant difference in survival was detected between the two regimens and the toxicities of FAMTX were found to be more manageable. Thus, FAMTX became standard therapy

12 144 for several years in advanced gastric cancer. Recently, however, investigators from the Royal Marsden Hospital reported a better response rate, quality of life and survival with continuous 5-FU, epirubicin and cisplatin, repeated every three weeks (ECF), compared to the FAMTX regimen: 36% and 21% at one year, respectively (P = ) in a randomized study [118]. The high level of dose-intensity obtained using protracted 5-FU in the ECF regimen might explain these better results. Finally, Italian investigators very recently reported encouraging results of a phase II study of a complex weekly regimen combining fluorouracil, cisplatin, epirubicin, leucovorin, glutathione and G-CSF [119]. In this trial, a response rate of 62% and a median survival of 11 months were obtained in 105 patients. The promising results of this dose-dense regimen will have to be confirmed in a randomized trial comparing it to the ECF regimen before the former can be considered the standard therapy. Dose-density in urothelial carcinoma The M-VAC regimen has been considered the gold standard for advanced transitional cell carcinoma of the urinary tract since the 1980s. Despite extensive use of hematopoietic growth factors, enhancing tolerance to chemotherapy in some trials, no significantly improved response rates or survival have been reported to date. In a study reported by a team from the M.D. Anderson, 48 patients were randomized to receive a modified M-VAC schedule with or without GM-CSF [120]. Chemotherapy was repeated every 23 days or when biological criteria were met. The results show no difference in the dose intensity, response rate or survival. However, this trial was not designed to detect a difference in response or survival. A second randomized trial conducted by the EORTC has compared the classic M-VAC regimen to a biweekly M-VAC + G-CSF. Accrual to this trial was recently discontinued. An interim toxicity analysis was presented in 1997: a lower incidence of febrile neutropenia, a greater dose-intensity for doxorubicin and cisplatin and better treatment compliance were reported for the dose-dense arm [121]. Data on response and survival are now awaited. Several non-randomized trials have tried to reduce the intervals in the M-VAC regimen [ ]. Investigators at the Memorial Sloan-Kettering Cancer Center initiated a phase I study in 23 patients, to evaluate the maximum tolerable dose of M-VAC with G-CSF support and to define the best interval between cycles (two or three weeks). They found that the dose-intensity of M-VAC could be escalated by 33% using a threeweekly schedule. However, this dose-intensity was maintained through the first three cycles only [122]. A Japanese trial has also demonstrated that the interval between cycles of the M-VAC regimen could be safely reduced from three weeks to four weeks with G-CSF support. In this study, the duration of the hospital stay was reduced by two weeks in patients receiving the dosedense M-VAC [123]. During roughly the same period, the Eastern Cooperative Oncology Group reported disappointing results and a significant toxicity with a fourweekly regimen of intensified M-VAC + G-CSF [124]. Thus, reducing the time interval of the M-VAC regimen from four to three weeks seems feasible and may slightly enhance the dose intensity. However, the toxicity of standard M-VAC and the poor tolerance documented in patients with advanced urothelial carcinomas makes it unlikely that patient outcome will be significantly improved. Dose-density in miscellaneous tumors To our knowledge, only anecdotal studies of dose-densified chemotherapy in other solid neoplasms have been published and these mainly concern head and neck carcinomas, gynecologic cancers and soft-tissue sarcomas. No definitive results are available from randomized studies for these neoplasms. The feasibility of dosedense chemotherapy was the focus of the series devoted to these neoplasms, but the limited size of the patient populations precludes any definitive conclusions. In head and neck carcinomas, a biweekly 5-FU + cisplatinum schedule yielded an encouraging 82% response rate in a heterogeneous series of 44 patients [125]. Similar results were reported with cisplatin, 100 mg/m 2 on days 1 and 8 followed by local irradiation in those patients with locally advanced disease [126]. Of particular interest is the study reported by Verweij et al., 59 patients with locally advanced head and neck carcinoma were given six cycles of weekly cisplatin (80 mg/m 2 ) before local therapy. Non-hematologic toxicities were manageable. However, the incidence of severe leukopenia and/or thrombocytopenia was high (17 patients with each toxicity) and only nine patients received the planned dose intensity. Moreover, auditive toxicity was noted in 13 patients. The authors concluded that dose-densified cisplatin as a single agent compares favourably with classic combinations in head and neck carcinomas [127]. A single report has been published on a weekly schedule of adjuvant and neo-adjuvant vincristine-bleomycin-methotrexate in 38 patients with nodal metastases from penile carcinomas. Toxicity was moderate but this regimen seems less effective than conventional 5-FU-cisplatin combinations in patients with unresectable metastases [128]. In metastatic adenocarcinoma of the cervix, weekly epirubicin is well-tolerated but of minimal activity [129]. A moderate increase in doseintensity has been achieved in adult soft-tissue sarcomas, while shortening cycle intervals with an EID regime, but its clinical implications remain unproven [130]. In mesothelioma, promising results had initially been reported with the combination of weekly cisplatin and oc-interferon or with weekly cisplatin alone [131,132] but were not confirmed later [133]. Moreover, the toxicity/efficacy ratio of this strategy compares poorly with that of more recent chemotherapy combinations [134].

13 145 Conclusions Considerable experience has been accumulated in rapidly-repeated regimens, as summarised in this review. Results clearly indicate that this strategy is feasible in many oncologic settings, with manageable toxicity and a limited number of toxicity-related deaths, when therapy is adequately monitored by well-trained oncologists. Response rates are impressive in a number of neoplasms, and particularly in the case of germ-cell tumours, breast cancer, ovarian cancer, Ewing's sarcoma and colorectal cancer. To date, only a few randomized trials have demonstrated a benefit in disease-free survival or overall survival in small-cell lung cancer and advanced colorectal cancer [18, 20, 112, 113]. Such a benefit is suggested in other uncontrolled studies in various neoplasms, and especially germ-cell tumors, gestational trophoblastic disease, Ewing's sarcoma and perhaps high-risk node-positive breast cancer [33, 38, 42, 47, 48, 57-60, 107, 109]. In parallel, a better efficacy/ toxicity ratio has been reported in some models with sequential chemotherapy as compared to alternated or concurrent schedules [5]. This indicates that investigators should now focus research on rapidly-repeated sequential schedules. However, 'one cycle every three or four weeks' is certainly not definitively obsolete, even if many biological and clinical arguments are strongly in favour of rapidly-repeated regimens. Although the activity of dosedense chemotherapy is potentially superior to that of standard treatments in some solid tumours, the clinical impact of this benefit is likely to be limited and should be weighed against quality of life impairment that may be associated with more toxic dose-dense regimens. Moreover, clinical and biological criteria earmarking candidate patient subgroups likely to benefit from this strategy are not available. Combining dose-dense chemotherapy regimens with new anti-cancer strategies in light of insights into the mechanisms of oncogenesis will be a challenge on the eve of the millenium. It is noteworthy that most randomized studies which suggested a benefit with dose-dense chemotherapy also reported increased dose-intensity in the experimental arm. If such is the case, shortening cycle intervals would probably be one of the most efficient ways of increasing dose intensity, as previously suggested [72]. This brings to the fore the concept of dose-intensity measurement which has been a subject of controversy since the measurement standard (mg/m /week) was recently criticised by Gurney et al. [135, 136]: the issue addressed by these authors, which is of major import in the practice of medical oncology, could be summarised thus: 'Prescribing anticancer agents in mg/m 2, is it obsolete?' They showed that there is no or only a limited correlation between the body surface area and pharmacokinetic parameters for most drugs. The body surface area proved to be an inadequate yardstick for standardising marked inter-patient variation in pharmacokinetics and this appeared relevant for most cytotoxic agents [ ]. This could partially account for some of the shortcomings of many current dose-intensity strategies. Acknowledgements We wish to thank Dr P. Cottu, Dr J.-M. Extra, Dr J.-P. Armand, Dr P. Pautier, Dr V. Boige, Dr M. Ducreux and Prof. M. Marty for fruitful discussions and pertinent advice. We thank L. Saint Ange for editing the manuscript. References 1. Bergsagel DE. An assessment of massive-dose chemotherapy of malignant disease. Can Med Assoc J 1971; 104: Goldie JH, Coldman AJ. The genetic origin of drug resistance in neoplasms: Implication for systemic therapy. Cancer Res 1984; 44: Skipper HE. Historic milestones in cancer biology: A few that are important in cancer treatment (revisited). Semin Oncol 1979; 6: Norton L, Simon R. The Norton-Simon hypothesis revisited. Cancer Treat Rep 1986; 70: Bonadonna G, Zambetti M,Velagussa P. Sequential or alternating doxorubicin and CMF regimens in breast cancer with more than three positive nodes. JAMA 1995; 273: Culine S, Extra JM, Espie M et al. Promising results of an induction regimen with high-dose cyclophosphamide and epirubicin for inflammatory breast cancer. Bull Cancer 1993; 80: Green M. Dose-intensive chemotherapy with cytokine support. Semin Oncol 1994, 21: Crilley P, Goldstein LJ. Peripheral blood stem cell transplant in breast cancer. Semin Oncol 1995; 22: Droz JP, Pico J, Biron P et al No evidence of a benefit of early intensified chemotherapy (HDCT) with autologous bone marrow transplantation in first-line therapy of poor-risk non seminomatous germ-cell tumors. Proc Am Soc Clin Oncol 1992; 11: Rahman ZU, Hortobagyi GN, Buzdar AU, Champlin R. Highdose chemotherapy with autologous stem-cell support in patients with breast cancer. Cancer Treat Rev 1998; 24: Raptis G, Tiersten A, Hamilton N et al. Outpatient based, rapidly administered multi-cycle high-dose chemotherapy (HDC) in patients with responding metastatic breast cancer (MBC). Proc Am Soc Clin Oncol 1995; 14: Arriagada R, Le Chevalier T, Pignon JP et al. Initial chemotherapeutic doses and survival in patients with limited small-cell lung cancer. N Engl J Med 1993; 329: Weckstein D, Hurley C. Karp D et al. Phase I trial of sequential high dose cyclophosphamide (Cy), carboplatin (Cb), and escalating etoposide (E) with G-CSF primed peripheral blood stem-cell (PBSC) rescue in patients with metastatic cancer. Proc Am Soc Clin Oncol 1995; 14: Bramwell VHC, Williams CJ. Do authors of review articles use systematic methods to identify, assess and synthesize information? Ann Oncol 1997; 8: Mulrow CD. The medical review article: State of the science. Ann Intern Med 1987; 106: Sculier JP, Paesmans M, Bureau G et al. Multiple-drug weekly chemotherapy versus standard combination regimen in smallcell lung cancer: A phase III randomized study conducted by the European Lung Cancer Working Party. J Clin Oncol 1993; 11: Souhami RL, Rudd R, Ruiz de Elvira MC et al. Randomized

14 146 trial comparing weekly versus three-week chemotherapy in small-cell lung cancer: A Cancer Research Campaign trial. J Clin Oncol 1994; 12: Steward WP, von Pawel J, Gatzemeier U et al. Effects of granulocyte-macrophage colony-stimulating factor and dose intensification of V-1CE chemotherapy in small-cell lung cancer: A prospective randomized study of 300 patients. J Clin Oncol 1998; 16: Furuse K, Fukuoka M, Nishiwaki Y et al. I study of intensive weekly chemotherapy with recombinant human granulocyte colony-stimulating factor versus standard chemotherapy in extensive-disease small-cell lung cancer. J Clin Oncol 1998; 16: Thatcher N, Sambrook RJ, Stephens RJ et al. Dose intensification (DI) with G-CSF improves survival in small-cell lung cancer (SCLC). Results of a randomized trial. Proc Am Soc Clin Oncol 1998; 17: 456a. 21. Woll PJ, Hodgetts J, Lomax L et al. Can cytotoxic dose-intensity be increased by using granulocyte colony-stimulating factor? A randomized controlled trial of lenograstim in small-cell lung cancer. J Clin Oncol 1995; 13: Fukuoka M, Masuda N, Negoro S et al. CODE chemotherapy with and without granulocyte colony-stimulating factor in smallcell lung cancer. Br J Cancer 1997: 75: Miles DW, Fogarty O, Ash AM et al. Received dose-intensity: A randomized trial of weekly chemotherapy with and without granulocyte colony-stimulating factor in small-cell lung cancer. J Clin Oncol 1994; 12: Pujol JL, Douillard JY, Riviere A et al. Dose-intensity of a fourdrug chemotherapy regimen with or without recombinant human granulocyte-macrophage colony-stimulating factor in extensive-stage small-cell lung cancer: A multicenter randomized phase III study. J Clin Oncol 1997, Murray N, Shah A, Osoba D et al. Intensive weekly chemotherapy for the treatment of extensive-stage small-cell lung cancer. J Clin Oncol 1991; 9: Tjan-Heijnen VCG, Biesma B, Festen J et al. Enhanced myelotoxicity due to granulocyte colony-stimulating factor administration until 48 hours before the next chemotherapy course in patients with small-cell lung carcinoma. J Clin Oncol 1998; 16: Kuboka K, Nishiwaki Y, Kakinuma R et al. Dose-intensive weekly chemotherapy for treatment of relapsed small-cell lung cancer. J Clin Oncol 1997; 15: Taylor CW, Crowley J, Williamson SK et al. Treatment of smallcell lung cancer with an alternating chemotherapy regimen given at weekly intervals' A Southwest Oncology Group pilot study. J Chn Oncol 1990; 8: Ardizzoni A, Venturini M, Crino L et al. High dose-intensity chemotherapy, with accelerated cyclophosphamide. doxorubicin, etoposide and granulocyte-macrophage colony stimulating factor, in the treatment of small-cell lung cancer. Eur J Cancer 1993; 29A: Kaye SB, Mead GM, Fossa S et al. Intensive induction-sequential chemotherapy with BOP/VIP-B compared with treatment with BEP/EP for poor-prognosis metastatic non-seminomatous germ-cell tumor: A randomized Medical Research Council/ European Organization for Research and Treatment of Cancer study. J Chn Oncol 1998: 16: Nichols CR, Williams SD, Loehrer PJ et al. Randomized study of cisplatin dose-intensity in poor-risk germ-cell tumors: A Southeastern Cancer Study Group and Southwest Oncology Group protocol. J Clin Oncol 1991: 9: Lewis CR. Fossa SD, Mead G et al. BOP/VIP- A new platinumintensive chemotherapy regimen for poor prognosis germ-cell tumours. Ann Oncol 1991; 2: Bower M, Newlands ES. Holden L et al. Treatment of men with metastatic non-seminomatous germ-cell tumours with cyclical POMB/ACE chemotherapy Ann Oncol 1997: 8: International Germ-Cell Cancer Collaborative Group. International germ-cell classification. A prognostic factor-based staging system for metastatic germ-cell cancers. J Clin Oncol 1997; 15: Wettlaufer JN. Feiner AS, Robinson WA. Vincristine, cisplatin, and bleomycin with surgery in the management of advanced metastatic nonseminomatous testis tumors. Cancer 1984; 53: Dearnaley DP. Fossa SD, Kaye SB et al. Adjuvant bleomycin, vincristine and cisplatin (BOP) for high risk clinical stage I (HRCS 1) non-seminomatous germ-cell tumours (NSGCT) - A Medical Research Council (MRC) pilot study. Proc Am Soc Clin Oncol 1998; 17: 309a 37. Horwich A. Dearnaley DP, Norman A et al. Accelerated chemotherapy for poor prognosis germ-cell tumours. Eur J Cancer 1994; 30A: Horwich A, Mason M, Fossa SD et al. Accelerated induction chemotherapy (C-BOP-BEP) for poor and intermediate prognosis metastatic germ-cell tumours (GCT). Proc Am Soc Chn Oncol 1997; 16' 319a. 39. Harstrick A, Schmoll HJ, Kohne-Wompner CH et al. Cisplatin, etoposide, ifosfamide. vincristine and bleomycin combination chemotherapy for far advanced testicular carcinoma. Ann Oncol 1991; Germa Lluch JR, Segui Palmer MA, Chment Duran MA et al. Intensive chemotherapy in poor-prognosis nonseminomatous germ-cell tumors of the testis. Eur Urol 1992; 21: Williams SD, Birch R, Einhorn LH et al. Treatment of disseminated germ-cell tumors with cisplatin, bleomycin and either vinblastine or etoposide. N Engl J Med 1987; 316: Amato RJ, Finn L, Logothetis CJ et al. Preliminary results of alternating sequential combination chemotherapy (CHT) for the treatment of high volume nonseminomatous germ-cell tumor (HVNSGCT). Proc Am Soc Clin Oncol 1996; 15: Bajorin D, Katz A, Chan E et al. Comparison of criteria for assigning germ-cell tumor patients to 'good risk' and 'p or risk' studies. J Clin Oncol 1988; 6: Motzer RJ. Bajorin DF, Bosl GJ et al. Paclitaxel (T) containing first-line salvage therapy selected by risk for patients (pts) with germ-cell tumors (GCT). Proc Am Soc Clin Oncol 1997; 16: 322a. 45. Culine S, Droz JP, Delva R et al. Rapidly recycled, intensive alternating chemotherapy in heavily pretreated progressive nonseminomatous germ-cell tumors. A feasibility study. Urol Oncol 1995; 1: Kim S, Howell B, Me Clay E et al. Dose-intensification of cisplatin chemotherapy through biweekly administration. Ann Oncol 1993; Bagshaw KD. Treatment of high-risk choriocarcinoma. J Reprod Med 1984; 29: Theodore C, Azab M, Droz JP et al. Treatment of high-risk gestational trophoblastic disease with chemotherapy combinations containing cisplatin and etoposide. Cancer 1989; 64: Berkowitz RS. Goldstein DP. Gestational trophoblastic disease. Cancer 1995; 76: Ellis G. Livingston RB. of dose-intensive continuous 5-fluorouracil, doxorubicin. and cyclophosphamide as adjuvant therapy for breast cancer. Cancer 1993; 71: Fetting J. Gray R. Fairclough D et al. Sixteen-week multidrug regimen versus cyclophosphamide, doxorubicin, and fluorouracile as adjuvant therapy for node-positive, receptor-negative breast cancer: An intergroup study. J Clin Oncol 1998: 16: Therasse P, Mauriac L, Welnicka M et al. An EORTC-NCIC- SAKK neoadjuvant randomized phase III study comparing CEF (5-FU, epirubicin. cyclophosphamide) dose intensified EC + G-CSF (filgrastim) in locally advanced breast cancer (LABC). Updated results including quality of life. Ann Oncol 1998: 9: Dhingra K. Singletary E. Strom E et al. Randomized trial of G-CSF (filgrastim)-supported does-intense neoadjuvant (neo) chemotherapy (CT) in locally advanced breast cancer (LABC). Proc Am Soc Clin Oncol 1995: 14: 94.

15 Findlay B. Tonkin K, Crump M et al. A dose-escalation trial of adjuvant CEF chemotherapy with G-CSF for premenopausal women with node-positive breast cancer Proc Am Soc Clin Oncol 1996; 15: Swain SM, Rowland J. Weinfurt K et al. Intensive outpatient adjuvant therapy for breast cancer: Results of dose escalation and quality of life. J Clin Oncol 1996; 14: Cuvier C. Extra JM, Espie M et al. Epirubicin 75 mg/m 2 and cyclophosphamide 1200 mg/m 2 every two weeks for stage II breast cancer patients with more than four involved nodes. Proc Am Soc Clin Oncol 1997; 16: 165a. 57. Hudis C. Seidman A. Raptis G et al. Sequential (Seq) dose-dense (DD) doxorubicin (A), paclitaxel (T). cyclophosphamide (C ) is less toxic than A. concurrent T + C as adjuvant therapy in resected node-positive breast cancer. Proc Am Soc Clin Oncol 1996: 15: Zujewski J, Danforth D, Nosne M et al. Short course, doseintensive, 14 day cycles of doxorubicin (A) and cyclophosphamide (C ) followed by infusionnal paclitaxel (T) in the treatment of high-risk primary breast cancer. Proc Am Soc Clin Oncol 1996, 15: Hudis C, Seidman A, Baselga J et al. Sequential adjuvant therapy with doxorubicin, paclitaxel, cyclophosphamide for resectable breast cancer involving four or more axillary nodes. Semin Oncol 1995; 22: Riccio L, Hudis C, Seidman A et al. Long-term distant diseasefree survival (DFS) from two pilot studies of dose-dense sequential adjuvant chemotherapy (CRX) in women (pts) with resected breast cancer (BC) and more than three positive lymph nodes. Proc Am Soc Clin Oncol 1997: a. 61. Bernardo C, Plastina M, Bernardo A et al. Neoadjuvant chemotherapy (cyclophosphamide, epirubicin, and fluorouracil) plus G-CSF for operable breast cancer Proc Am Soc Clin Oncol 1995; 14: Graham M, Gonzalez P, Cance Wet al. Dose-intensive sequential adriamycin (A) and i.v. CMF (A > CMF) with G-CSF support for stage II and III breast cancer (BrCa): A feasible and effective outpatient regimen. Proc Am Soc Clin Oncol 1996; 15: Blohmer J, Tulusan A, Jackisch et al. Dose intensified preoperative chemotherapy with doxorubicin/docetaxel in operable breast cancer - a phase II trial. Ann Oncol 1998; 9: Breier S, Lebedinsky C, Giangiacomo G et al. Intensive chemotherapy with 15 days interdose interval with G-CSF support in locally advanced breast cancer (LABC). Proc Am Soc Clin Oncol 1996; 15: Biganzoh L, Piccart M. The bigger the better?...or what do we know and what we still need to learn about anthracyclin dose per course, dose density and cumulative dose in the treatment of breast cancer. Ann Oncol 1997; 8: (Editorial). 66. Gundersen S, Kvinnsland S, Klepp O et al. Weekly adriamycin versus VAC in advanced breast cancer A randomized trial. Eur J Cancer 1986; 22: 1431^t. 67. Richards MA, Hopwood P, Ramirez AJ et al. Doxorubicin in advanced breast cancer: Influence of schedule on response, survival and quality of life. Eur J Cancer 1992: 28A: Blomqvist C, Elomaa I, Rissanen Pet al. Influence of treatment schedule on toxicity and efficacy of cyclophosphamide, epirubicin and fluorouracil in metastatic breast cancer: A randomized trial comparing weekly and every-four-week administration. J Clin Oncol 1993; 11: Ardizzoni A, Venturini M, Sertoli MR et al. Granulocyte-macrophage colony-stimulating factor (GM-CSF) allows acceleration and dose intensity increase of CEF chemotherapy: A randomized study in patients with advanced breast cancer. Br J Cancer 1994; 69: Bonadonna G, Pannuti F, Robustelli della Cuna G et al. I study comparing standard dose FNC (5-fluorouracil, mitoxantrone, cyclophosphamide) versus dose-intensive FNC + rhug-csf (lenograstim) versus time intensive FNC + lenograstim in advanced breast cancer patients: Preliminary results of the CSF-001 study. Proc Am Soc Clin Oncol 1997: 16: Hortobagyi G. Bodey G. Buzdar A et al. Evaluation of high-dose versus standard FAC chemotherapy for advanced breast cancer in protected environment units: A prospective randomized study. J Clin Oncol 1987: 5: Lalisang R, Wils J, Nortier H et al. Comparative study of dose escalation versus interval reduction to obtain dose-intensification of epiriubicin and cyclophosphamide with granulocytc colony-stimulating factor in advanced breast cancer. J Clin Oncol 1997: 15: Fountzilas G. Athanassiades A. Giannakakis Tet al. A randomized study of epirubicin monotherapy every four or every two weeks in advanced breast cancer. A Hellenic Cooperative Oncology Group study. Ann Oncol 1997: 8: Joensuu H, Holli K. Heikkinen M et al. Combination chemotherapy versus single-agent therapy as first- and second-line treatment in metastatic breast cancer: A prospective randomized trial. J Clin Oncol 1998; 16: Greenberg P, Hortobagyi G. Smith Tet al. Long-term follow-up of patients with complete remission following combination chemotherapy for metastatic breast cancer. J Clin Oncol 1996; 14: Bronchud M, Howell A, Crowther D. The use of granulocyte colony-stimulating factor to increase the intensity of treatment with doxorubicin in patients with advanced breast and ovarian carcinoma. Br J Cancer 1989; 60: Scinto AF, Ferraresi V, Campioni N et al. Accelerated chemotherapy with high-dose epirubicin and cyclophosphamide plus r-met-hug-csf in locally advanced and metastatic breast cancer. Ann Oncol 1995; 6: Piccart MJ, Brumng P,Wildiers J et al. An EORTC pilot study of filgrastim (recombinant human granulocyte colony-stimulating factor) as support to a high dose-intensive epiadriamycin-cyclophopshamide regimen in chemotherapy-naive patients with locally advanced or metastatic breast cancer. Ann Oncol 1995; 6: Extra JM, Rousseau F, Mignot L et al. High-dose induction chemotherapy (HDIC) in breast cancer with metastases at diagnosis (MBC). Proc Am Soc Clin Oncol 1995; 14: Zelek L, Cottu P, Espie M et al. Updated results of dose densified chemotherapy with epirubicin and cyclophosphamide in metastatic breast cancer. Breast Cancer Res Treat 1997; 46: Cazap E, Jovtis S, Torchinsky D et al. Fluorouracil (F), epirubicin (E) and cyclophosphamide (FEC) each 14 days with GM-CSF in locally advanced and metastatic breast cancer. Proc Am Soc Clin Oncol 1997; 16: 164a. 82. Cottu P, Extra JM, Cuvier C et al. Improved survival with highdose sequential (HDS) cyclophosphamide (C ) and epirubicin (E) with peripheral blood stem cell (PBSC) at first-line therapy for metastatic breast cancer (MBC). Proc Am Soc Clin Oncol 1997; 16: 112a. 83. Lalisang R. Nortier J, Wils J et al. Dose dense epirubicin (E) and paclitaxel (P) with G-CSF (filgrastim) in metastatic breast cancer (MBC): A weekly schedule? Proc Am Soc Clin Oncol 1997; 16: 178a. 84. McCaskill-Stevens W, Ansari R, Fisher W et al. Phase 11 study of biweekly cisplatin (C ) and paclitaxel (P) in the treatment of metastatic bresat cancer. Proc Am Soc Clin Oncol 1996; 15: Sparano JA, Neuberg D, Click JH et al. A phase II trial of biweekly paclitaxel and cisplatin in patients with advanced breast carcinoma: An ECOG trial. Proc Am Soc Clin Oncol 1996; 15: Breier S, Lebedinsky C, Pelayes L et al. Phase I II weekly paclitaxel (P) 80 mg/m 2 in pretreated patients with breast (BC) and ovarian cancer (OC). Proc Am Soc Clin Oncol 1997; 16: 163a. 87. Seidman AD. Hudis CA. Albanel J et al. Dose-dense therapy with weekly one-hour paclitaxel infusions in the treatment of metastatic breast cancer. J Clin Oncol 1998; 16:

16 Luck HJ, Marhenke D, Petry K.U et al Weekly paclitaxel monotherapy as salvage treatment in pretreated patients with metastatic breast cancer: Experience with an one-hour schedule. Breast Cancer Res Treat 1997; 46: Sikov WM, Akerley W, Cummings F et al. Weekly high-dose paclitaxel in locally advanced (LABC) and metastatic (MBC) breast cancer. Breast Cancer Res Treat 1997; 46: Cuvier C, Romieux G, Culine S et al. Dose-finding study of docetaxel (Taxotere 8 ') and cyclophosphamide (Endoxan ) every two weeks as first-line chemotherapy for metastatic breast cancer. Ann Oncol 1998; 9 (Suppl 4): Torn V, Korn EL, Simon R. Dose intensity analysis in advanced ovarian cancer patients. Br J Cancer 1993; 67: Kaye SB, Paul J, Cassidy J et al. Mature results of a randomized trial of two doses of cisplatin for the treatment of ovarian cancer. J Clin Oncol 1996; 14: Murphy D, Crowther D, Renninson J et al. A randomised dose intensity study in ovarian carcinoma comparing chemotherapy given at 4 weeks intervals for 6 cycles with half-dose chemotherapy given for 12 cycles. Ann Oncol 1993; 4: McGuire W, Hoskins WJ, Brady MF et al. Assessment of doseintensive therapy in suboptimally debulked ovarian cancer: A Gynecologic Oncology Group study. J Clin Oncol 1995; 13: Wrigley E, Weaver A, Jayson G et al. A randomized trial investigating the dose intensity of primary chemotherapy in patients with ovarian carcinoma: A comparison of chemotherapy given every four weeks with the same chemotherapy given at three weeks intervals. Ann Oncol 1996; 7: Bolis G, Favalli G, Danese S et al. Weekly cisplatin given for two months versus cisplatin plus cyclophosphamide given for five months after cytoreductive surgery for advanced ovarian cancer. J Clin Oncol 1997; 15: Advanced Ovarian Cancer Trialists Group. Chemotherapy in advanced ovarian carcinoma. An overview of randomised clinical trials. BMJ 1991; 303: Hidalgo M, Mendiola C, Lopez-Vega JM et al. Multicenter randomized trial of G-CSF supported platinum-based chemotherapy with mid-cycle platinum dose intensification in advanced ovarian cancer. Ann Oncol 1996; 7: Van der Burg MEL, Logmans A, de Wit R et al Six weeks of weekly high dose cisplatin (P) and daily oral vepesid (VP): A highly active regimen for ovarian cancer patients (pts) failing on or relapsing after conventional platinum containing combination chemotherapy. Proc Am Soc Clin Oncol 1996; 15: Swenerton K, Hoskins P, Stuart G et al. A phase I study of biweekly paclitaxel-cisplatin as initial therapy for advanced ovarian cancer. A study of the National Cancer Institute of Canada Clinical Trials Group. Ann Oncol 1996; 7: Fennelly D, Shapiro F, Aghajanian C et al. Applications of doseintensive paclitaxel (PTX) utilizing a weekly (w) one-hour (H) infusion schedule: Efficacy and feasibility in patients (pts) with recurrent ovarian cancer. Ann Oncol 1996: 7: Fennelly D, Shapiro F, Schneider J et al. A novel approach to delivery of dose-dense chemotherapy in patients (pts) with advanced ovarian cancer (AOC): A phase I trial. Proc Am Soc Clin Oncol 1995: 14: Weaver A, Wrigley E, Watson A et al. A study of ovarian cancer patients treated with dose-intensive chemotherapy supported with peripheral blood progenitor cells mobilised by filgastrim and cyclophosphamide. Br J Cancer 1996; 74: OConnell G, Shelley W. Carmichael J et al. High dose-intensity regimen of weekly doxorubicin and cisplatin in the treatment of patients with stage III and IV epithelial ovarian carcinoma. Cancer Treat Rep 1987: 71: Vermorken JB, Bolis G. van Rijswijk Ren et al. High-dose intensity regimens with epirubicin in ovarian cancer. Semin Oncol 1994: 21: Delattre O. Zucman J. Melot Tet al. The Ewing family of tumors. A subgroup of small-round cell tumors defined by a specific chimeric transcripts. N Engl J Med 1994; 331: Kushner BH, Meyers PA, Gerald WL et al. Very-high-dose shortterm chemotherapy for poor-risk peripheral primitive neuroectodermal tumors, including Ewing's sarcoma, in children and young adults. J Clin Oncol 1995; 13: Kushner BH, La Quaglia MP, Wollner N et al. Desmoplastic small round-cell tumor: Prolonged progression-free survival with aggressive multimodality therapy. J Clin Oncol 1996; 14: Fizazi K, Dohollou N, Missenard G et al. Rapidly recycled doxorubicin, ifosfamide and cisplatin + G-CSF (AP1-AI regimen) associated with surgery and radiotherapy: A highly effective therapy in adults with the Ewing's sarcoma family of tumors Proc Am Soc Clin Oncol 1997; 16: 499a Fizazi K, Dohollou N, Blay JYet al. Ewing's family of tumors in adults: Multivariate analysis of survival and long-term results of multimodality therapy in 182 patients. J Clin Oncol 1998; 16 (in press) Buroker TR, OConnell MJ, Wieand S et al. Randomized comparison of two schedules of fluorouracil and leucovorin in the treatment of advanced colorectal cancer. J Clin Oncol 1994; 12: De Gramont A, Bosset JF, Milan C et al. Randomized trial comparing monthly low-dose leucovorin and fluorouracil bolus with bimonthly continuous infusion for advanced colorectal cancer: A French Intergroup study. J Clin Oncol 1997; 15: Weh HJ, Zschaber R, Braumann D et al. A randomised phase- Ill study comparing weekly folinic acid (FA) and high dose 5-fluorouracil (5-FU) with monthly 5-FU/FA d 1-5 in untreated patients with metastatic colorectal carcinoma (CRC). Ann Oncol 1998; 9: Raymond E, Chaney SG,Taama A et al. Oxaliplatin: A review of preclinical and clinical studies. Ann Oncol 1998; 9: De Gramont A, Figer A, Seymour M et al. A randomized trial of leucovorin (LV) and 5-fluorouracil (5-FU) with or without oxaliplatin in advanced colorectal cancer. Proc Am Soc Clin Oncol 1998; 17: 257a Wils JA, Klein HO, Wagener DJT et al. Sequential high-dose methotrexate and fluorouracil combined with doxorubicin. A step ahead in the treatment of advanced gastric cancer: A trial of the European Organization for Research and Treatment of Cancer Gastrointestinal Tract Cooperative Group. J Clin Oncol 1991; 9: Kersen D. Atiq OT, Saltz L et al. FAMTX versus etoposide. doxorubicin and cisplatin: A random assignment trial in gastric cancer. J Clin Oncol 1992, 10: Webb A, Cunningham D, ScarfTe JH et al. Randomized trial comparing epirubicin. cisplatin. and fluorouracil versus fluorouracil, doxorubicin. and methotrexate in advanced esophagogastric cancer. J Clin Oncol 1997; 15: Cascinu S, Labianca R, Alessandroni Pet al. Intensive weekly chemotherapy for advanced gastric cancer using fluorouracil, cisplatin, epi-doxorubicin. 6S-leucovorin. gluthatione, and filgrastim: A report from the Italian Group for the Study of Digestive Tract Cancer. J Clin Oncol 1997; 15: Logothetis C. Finn LD. Smith Tet al. Escaladated MVAC with or without recombinant human granulocyte-macrophage colony-stimulating factor for the initial treatment of advanced malignant urothelial tumors: Results of a randomized trial. J Clin Oncol 1995: 13: Sternberg CN. De Mulder P. Fossa S et al. Interim toxicity analysis of a randomized trial in advanced urothelial tract tumors of high-dose intensity MVAC chemotherapy (HD-MVAC) and recombinant human granulocyte colony-stimulating factor (G-CSF) versus classic MVAC chemotherapy (EORTC 30924). Proc Am Soc Clin Oncol 1997: 16: 320a Seidman A. Scher H, Gavrilo\e J et al. Dose-intensification of M-VAC with recombinant granulocyte colony-stimulating factor

17 149 as initial therapy in advanced urolhelial cancer. J Clin Oncol 1993: 11: Nogushi S. Kubota Y. Shuin Tet al. Schedule intensified M-VAC chemotherapy for advanced urothelial cancer with recombinant human granulocyte colony-stimulating factor (rhg-csf). Int J Urol 1994: 1: Loehrer P. Elson P. Dreicer R et al. Escalated dosages of methotrexate, vinblastine, doxorubicin and cisplatin plus recombinant human granulocyte colony-stimulating factor in advanced urothelial carcinoma. An Eastern Cooperative Oncology Group Trial. J Clin Oncol 1994; 12: Zidan J, Kuten A. Robinson E. Intensive chemotherapy using cisplatin and 5-FU followed by radiotherapy in adanced head and neck cancer. Proc Am Soc Clin Oncol 1995: 14: Valleio C, Ferreyra R. Leone B et al. High-dose cisplatin as neo-adjuvant organ preserving chemotherapy for squamous-cell carcinoma of the larynx. Proc Am Soc Clin Oncol 1995; 14: Planting AST. de Mulder PHM, de Graeff A, Verweij J. study of weekly high-dose cisplatin for six cycles in patients with locally advanced squamous-cell carcinoma of the head and neck. EurJ Cancer 1997: 33: Pizzocaro G. Piva L, Nicolai N. Adjuvant and neoadjuvant chemotherapy for nodal metastases from squamous-cell carcinoma of the penis. Proc Am Soc Clin Oncol 1995: 14: van der Burg MEL, Bolis G, Bakker PJM et al. study of Weekly 4'-epidoxorubicin in patients with metastatic adenocarcinoma of the cervix: An EORTC gynecological cancer cooperative group study. Eur J Cancer 1997; 33: Michelotti A, Romanini A, Giannessi P et al. Accelerated epirubicin-ifosfamide-dacarbazine regimen in patients with adult soft-tissue sarcomas. Am J Clin Oncol (CCT) 1996; Soulie P, Ruffle P, Trandafir L et al. Combined systemic chemolmmunotherapy in advanced diffuse mesothehoma: Report of a phase I-I I study of weekly cisplatin-interferon-oc-2a. J Clin Oncol 1996; 14: Planting AST. Schellens JHM, Goey SH et al. Weekly high-dose cisplatin in malignant pleural mesothelioma Ann Oncol 1994; Trandafir L. Ruffle P. Borel C et al. Higher doses of a-interferon do not increase the activity of the cisplatin-interferon combination in advanced malignant mesothelioma. J Clin Oncol 1997; 33: Fizazi K, Daniel C, Calliandro R et al. Tomudex-oxaliplatin: A step ahead in the struggle against mesothelioma? The Institut Gustave Roussy experience with chemotherapy and chemoimmunotherapy in mesothelioma. Presented at the Malignant Pleural Mesothelioma meeting. Lignano Sabbiadoro, March Gurney HP. Dose calculation of anticancer drugs. A review of the current practice and introduction of an alternative. J Clin Oncol 1996; 14: Gurney HP, Ackland S, Gebski V et al. Factors affecting epirubicin pharmacokinetics and toxicity: Evidence against bodysurface area for dose calculation. J Clin Oncol 1998; 16: Ratain MJ. Body-surface area as a basis for dosing of anticancer agents: Science, myth, or habit? J Clin Oncol 1998; 16: Dobbs, Twelves CJ. What is the effect of adjusting epirubicin doses for body surface area? Br J Cancer 1998; 78: Sandstrom M, Freijs A, Larsson R et al. Lack of relationship between systemic exposure for the component drugs of the fluorouracil, epirubicin, and 4-hydroxycyclophosphamide regimen in breast cancer patients. J Clin Oncol 1996; 14: Received 28 July 1999; accepted 16 December Correspondence to: K. Fizazi, MD Department of Medicine Institut Gustave Roussy 39 rue Camille Desmoulins Villejuif France fizazi@igr.fr