Evaluation of Dose-Intense Ifosfamide, with and Without Edatrexate, in Adults with Sarcoma

Purpose. To define the maximally tolerated dose (MTD) of ifosfamide when given with G-CSF on an every other week schedule, and to define the MTD of edatrexate that can be given every two weeks with an intense schedule of ifosfamide. Patients and Methods. Forty-one patients with metastatic or unresectable, locally advanced sarcoma participated in this 2-step phase I trial.The starting dose of ifosfamide was 10 gm/m2 given by continuous intravenous infusion over 4 days every 2 weeks.When the MTD was defined, edatrexate, beginning at a dose of 40 mg/m2 intravenously every 2 weeks was added in subsequent cohorts of patients. Results. Myelosuppression was the most prominent toxicity. Fatigue, nausea, and vomiting were observed in the majority of patients. Ifosfamide 12 gm/m2 given every 2 weeks approached or exceeded the MTD. Edatrexate 100 mg/m2 could be given safety as an intravenous bolus with ifosfamide 10 gm/m2 every 2 weeks. Therapeutic responses were observed in patients with measurable disease. Conclusions. This study demonstrates the feasibility of administering a dose-intense schedule of ifosfamide alone or ifosfamide with edatrexate that might be applied in the adjuvant or neo-adjuvant setting.


Introduction
T he activity of ifosfam ide in soft tissue sarcom a has been recognized since 1973, but the developm ent of this drug was hindered by its urothelial toxicity. The introduction of the disul® de uroprotective agent, m esna, perm itted extensive clinical evaluation of ifosfam ide. In a large phase II trial, the response to ifosfam ide 8 gm /m 2 was 21% am ong 124 patients w ith sarcom a. 1 In a random ized phase II trial, the response to ifosfam ide 5 gm /m 2 was 24% am ong patients w ith soft tissue sarcom a who had not previously been treated w ith chemotherapy. 2 Response to its analog, cyclophosph am ide, at a dose of 1.5 gm /m 2 was only 9% , in spite of greater myelosuppression in the latter group of patients. A dose± response relationship has been claim ed for ifosfam ide in patients w ith soft tissue sarcom a. 3 In m ost studies of high dose ifosfam ide, the drug has been given at increased doses ever y 3 to 4 weeks. C olony stim ulating factors, such as G-C SF 3 and G M -CSF, 4,5 reduce the myelosuppressive toxicity of high dose ifosf am id e, but dose-depen den t renal and neu rolog ic toxicities lim it the dose of ifosfa m id e that can be g iven per cyc le. 6 M ost investigato rs h a ve a tte m p te d to i n c re a s e d o s e-in te n s ity b y in creasing the total dose o f ifosfa m ide given over several days ever y 3 weeks. In this study, we explored in tensi® cation of ifosf am ide therapy by in creasing the frequency of ifosf am ide adm inistration to ever y 2 weeks. As part of our phase II program for patients with soft tissue sarcom a, we observed 5 m ajor objective responses to edatrexate am ong 35 patients w ith soft tissue sarcom a. 7 The starting dose of edatrexate in that trial was 80 m g/m 2 weekly, but patient tolerance was variable, and weekly doses ranged from 50 to 120 m g/m 2 . E datrexate also has show n activity in patients w ith sarcom a in an E astern C ooperative O n co lo g y G ro u p p h a se I I tr ial. 8 In ad d itio n , edatrexate has dem onstrated antineoplastic activity in hum an sarco m a cell lines in ou r laborator y. 9 Preclin ical data have suggested synergy between edatrexate and alkylating agents in several m urine tu m o r m o d els, in clu d in g T 2 41 sa rco m a . 10 We reasoned that the com bination of edatrexate and ifosfam ide m ight have useful clinical efficacy in patients with sarcom a. Furtherm ore, an active, dose-intense ifosf am ide-b ased com bin atio n m ight be a usefu l component of neoadjuvant or adjuvant chem otherapy regim en for use in sequence or alternating w ith doxorubicin. A two-stage phase I trial of high dose ifosfam ide, therefore, was conducted. F irst, the m axim um tolerated dose (M TD ) of ifosfam ide given by continuous intravenous infusion over 4 days ever y other week with G-C SF support was determined.We then sought the M T D of edatrexate that could be given safely every two weeks w ith an intense schedule of ifosfamide. We now report the results of this investigation.

Patients, m aterials and m ethods
The protocol was reviewed and approved by the In stitu tion al R eview B oard of M em or ial S lo an -Kettering C ancer Center.

Patients
This study was designed as a disease-speci® c phase I trial for patients w ith m etastatic or unresectable, locally advanced sarcom a. Histologic con® rm ation of sarcoma by the D epartment of Pathology of M em orial H ospital was necessar y. T he trial was lim ited to patients who had not previously been treated with ifosfam ide, and those with not more than one previous chem o th erapy re gim en. Patien ts co uld n ot have received chem otherapy or therapeutic irrad iation during the 4 weeks before their ® rst dose of ifosfamide. A Karnofsky perform ance status of 60% or greater was required, as w as a W BC 4000/m m 3 , pla telet co u n t 1 5 0,0 00 /m m 3 , s er u m bilirubin<1.5 m g/dl, and serum creatinine<1.5 or creatinine clearance 60 m g/m in/1.73m 2 . Patients with clinically evident third space¯uid (ascites or pleural effusions) were excluded from this trial.
All patients had a physic al exam ination, a chest rad io g rap h, an d appro priate im ag in g stud ies to establish extent of disease and tumor m easurem ents when possible. M easurable or evaluable disease was not required, however. h u m an ® lg ra stim (G -C S F, N eupogen) for this study was provided by Am gen, Inc. (Thousand O aks, C A) under an ag reement with M em orial Sloan-K ettering Cancer C enter.

Treatment plan
S ign ed in fo rm ed con sen t w as obtained from all patients who participated in this trial. Treatment was given in the hospital. O ral or intravenous hydration was adm inistered to m aintain a urine output of at least 2000 m l/24 h during ifosfam ide infusion. T he ifo sfam id e a n d m esn a w er e ad m in istered b y contin u ous 24-h o ur in traven ou s in fusion over 4 con secutive d ays ever y 2 weeks. M esn a was not continued beyond the com pletion of the ifosfam ide in fu sion . Ifosf am id e therapy w as interr upted fo r hem atu ria (R B C g re ater than 11 R BC /h pf), an increase in serum creatinine>1.0 m g/dl, or for clinically signi® cant neurological side effects. T he initial dose of ifosfam ide was 10 gm /m 2 . In patients who received edatrexate, this agent was given on the ® rst day of each cycle. The initial dose of edatrexate was 40 m g/m 2 . G -C S F ( 5m /k g/d ay) w a s self-a d m in istered subcutaneously daily beginning 48 hours after the com pletion of th e ifosfa m id e infu sion , an d w as adm inistered until the recovering AN C was 1500/ m m 3 . S ubsequ en t do ses of ifosfa m id e were n ot adm inistered within 24 hours of G -C SF injection. T he 4-day course of therapy followed by G -CSF was de® ned as a cycle of therapy.
If on day 15 the W BC was <4000/m m 3 the AN C was <1500/m m 3 or the platelet count was <150,000/ m m 3 , or if grade 2 or greater m ucositis was present, chemotherapy was held one week. If toxicity had not resolved by day 21, treatment was held one additional week. Patients were taken off study if hematologic recover y was delayed beyond this time. At least 3 patients were to be treated at each dose level. D ose esca latio n w ith in in d ivid u al p atien ts w as n o t perm itted. If dose-lim iting toxicity was seen in the initial 3 patients at a given dose level, additional patients were added to further evaluate the toxicity.
In the original plan, the m axim um tolerated dose (MT D) was de® ned on the basis of duration of myelosuppression, w ith m ore than 5 days of grade 4 neutropenia constituting dose-lim iting toxicity. D uring the stu dy it b ecam e clear that the no n-h em atolog ic toxicity likely to be associated with that degree of myelosuppression would be unacceptable.T hus, since our objective was to treat patients every 2 weeks, any toxicity that prevented recycling of treatm ent on day 15 (e.g. AN C <1500/m m 3 or platelet count<150,000/ m m 3 ) was accepted as dose-lim iting, and the M TD was de® ned as the highest dose level at which the m ajor ity of patients could be recycled on day 15. This was a two-part study. First, the M T D and schedule for the adm inistration of ifosfam ide every 2 weeks was de® ned. Additional cohorts of patients were then treated w ith escalating doses of edatrexate in com bination with ifosfam ide to determ ine the M T D of edatrexate for use in com bination. One dose below the M TD of ifosfam ide de® ned in the ® rst part of the study was to have been used as the constant in determ ining the M TD for edatrexate in the second part of the study.

Evaluation dur ing study
Patients were followed until dose-lim iting toxicity precluded further therapy, or until prog ression of disease was observed. An automated C BC and platelet count was perform ed twice weekly. Physical exam ination and assessm ent of toxicity was done before each cycle of ifosfam ide. A biochemical pro® le including electrolytes, BUN , creatinine, calcium, phospho rus, and hepatic enzym es was obtained ever y 2 weeks. M easurem ent of indicator lesion(s) was perform ed at least ever y eight weeks for patients with m easurable disease.

Criteria for therapeutic response and toxicity
T he N CI Com m on Toxicity Criteria were used in this trial. Therapeutic response was not the principal endpoint of this trial. H owever, for those patients who had m easurable disease, the criteria of the Adult Intergroup Soft-Tissue Sarcom a C om m ittee were applied. 11 T he duration of response was measured from the ® rst day of therapy. U nequivocal clinical deterioration as evident from increasing pain, prog ressive weight loss and falling perform ance status was accepted as an indication of disease progression in the absence of signi® cant change in m easured lesions.
In patients with evaluable disease, unequivocal tumor shrinkage was recorded as im provem ent.

Results
The characteristics of the 41 patients who participated in the trial are presented in Table 1. T he prim ar y sites for the 34 patients w ith non-osseous sarcom a are listed in Table 2. All patients had a Karnofsky performance status of 70% or greater, and 78% had no previous chemotherapy. W ith the exception of one patient whose prior therapy was only paclitaxel, the previously treated patients had all been treated with doxorubicin or a doxorubicin-based com bination. Tw o p atien ts w ere in evalu a ble fo r h em ato log ic toxicity. O ne had acute central nervous system toxicity during his infosfam ide infusion, and therapy was discontinued. Another was ineligible because he was found to have a pleural effu sion shortly after receiving his ® rst doses of therapy. All but 3 evaluable patients had m easurable disease. An additional patient was lo st to fo llow -u p b efo re p o st-treatm en t tu m o r m easurem ents could be m ade. Table 3 demonstrates the W BC and platelet toxicity of the ® rst cycle of chem otherapy for each dose level. Table 4 o utlines the W BC and platelet toxicity encountered when all cycles adm inistered at each dose level are included. T he ® rst two levels involved infosfam ide without edatrexate. The starting dose of ifosfam ide, 10 gm / m 2 , was well-tolerated. At 12 gm /m 2 , one episode of septicemia was encountered, and sufficient myelosuppression was observed in the other 2 patients that a decision was m ade to interrupt accrual of patients at that level. Based on the plan to begin escalation of edatrexate in com bination with a dose of infosfam ide one level below the M T D, the dose of infosfam ide It is difficult to draw ® rm conclusions regarding th e re lation ship b etw een d o se an d h em ato lo g ic toxicity given the wide range of W BC , AN C, platelet nadirs, and the lim ited range of ifosfam ide doses stud ied. O verall, however, there seem ed to be a threshold effect for edatrexate, w ith doses up to 80 m g/m 2 having little im pact on myelosuppression. W ith edatrexate 100± 120 m g/m 2 , the add ition of edatrexate to a given dose of ifosfam ide app eared to be asso ciated with greater myelotoxicity.

Toxicity
In addition, 15 patients had grade 3 or 4 anem ia. Also anem ia was not clearly related to dose level, although it did appear related to duration of therapy. All patients experienced alopecia. T he incidence of the other com mon non-hem atologic toxicities, nausea and vom iting, m ucositis, and fatigue is shown in Table 5. Paralleling the anem ia obser ved, fatigue app eared to be asso ciated with prolonged duration of therapy.
In com bination with ifosfam ide 8 gm/m 2 , it was possib le to escalate the dose of edatrexate to 100 m g/ m 2 , the phase II do se o f that agen t for weekly adm inistration. Therefore, it was elected to treat a cohort of patients w ith infosfam ide 10 gm /m 2 , while keeping the dose of edatrexate at 100 m g/m 2 . W hen

Therapeutic responses
T he en d po in t of th is trial w as n ot th erap eu tic response, and phase II conclusions should not be d raw n fro m a ph a se I tr ial. N o n eth eless, th e overwhelming m ajority of patients with soft tissue sarcom a had received on previous chemotherapy, and m o st w ere evalu a b le fo r th era p eu tic resp o n se ( Table 6). It is im possible to evaluate duration of re sp on se , sin ce the len gth o f tre atm en t on this protocol was generally brief, and therapy after completion of the protocol treatm ent was not u niform . Am ong the responders, however, one patient w ith an angiosarcom a of the pelvis is recorded as having experienced a clinical PR. He had no viable tum or at the time of resection after treatment w ith ifosfam ide plus edatrexate. Although that patient had previously undergone resection of a m etastatic lym ph node in the axilla, he rem ains free of active sarcom a 48+ m onths after initiation of ifosfam ide therapy. The response rate was 14% am ong the 29 patients w ith m easurable soft tissue sarcom a. In addition, 3 patients with evaluable, but not strictly m easurable disease experienced unequivocal antitum or responses. This 18% response rate observed among the 32 patients with evaluable soft tissue sarcom a, and the 24% response rate overall is sim ilar to that reported for less intense schedules of ifosfam ide alone.

Discussion
The num ber of drugs with reproducible activity in patients w ith soft tissue sarcom a is lim ited, and the activity of the m ost active agents is modest at best. In la rg e, ran d o m iz ed tr ials, even co m b in atio n chem otherapy yields responses in fewer than 35% of patients. 12± 14 In the absence of novel drugs w ith m eaningful clinical activity, it is essential to m axim ize the therapeutic bene® t of the available agents. The antineoplastic activity of ifosfam ide in patients with sarcoma is well-established. In virtually all studies of ifosfam ide or ifosfam ide based combinations, the drug has been given every 3± 4 weeks. Phase II trials of ifosfam ide 5± 8 gm/m 2 yielded responses in 18± 24% of patients w ith sarco m a, 1,2 an d ifo sf am ide-b ase d regimens have become a standard in the treatment of patients with sarcoma. D rawing phase III conclusions from phase I or phase II data is unreliable. Taken together, however, several lines of evidence from phase I and phase II trials suggest a dose± response relationship for ifosfam ide in patients with sarcom a. High response rates have been reported in trials of higher doses of ifosfam ide. In a sm all cohort of patients with synovial cell sarcoma, a 100% CR+PR to infosfamide 14 g/m 2 was reported. 15 In a phase I trial in which the dose of ifosfam ide was escalated from 8 to 18 g/m 2 in sequential cohorts of patients, the m aximal tolerated dose was estimated to be 16 g/m 2 . 6 Am ong 20 patients with sarcom a in that trial, 7 (35%) had major responses. More recently a 43% CR+PR rate was reported among 34 evaluable patients who received ifosfam ide 14 gm/m 2 as ® rst-line therapy. 16 In a phase II trial, ifosfam ide 14 g/m 2 given over three days by continuous infusion yielded responses in 29% of 37 patients with soft tissue sarcom a, and 40% of patients w ith bone sarcom a. 3 Also within that report was a sm all cohort of patients in which the sam e total dose of ifosfam ide was given by intermittent bolus infusion. Five of 11 patients with soft tissue sarcom a as well as 3 of 3 patients with bone sarcom a responded, leading the authors to suggest that bolus therapy is m ore efficacious than continuous infusion.  10  0  6  24  0  0  1  0  4  0  12  0  3  7  0  0  1  2  2  1  8  40  5  22  2  0  2  0  2  0  8  60  7  28  2  0  4  0  2  0  8  80  6  30  2  1  1  0  2  0  8  100  5  17  0  2  1  0  3  1  10  100  5  22  2  0  3  0  0  5  10  120  5  19  4  0  2  1 3 2

Dose-intense ifosfamide and edatrexate in sarcoma
Pharm acokinetic studies, however, have not shown a difference in area under the curve for serum ifosfamide or its m etabolites, or in ifosfam ide m etabolites in urine for 1 hour or bolus infusions of ifosfam ide. 27 Further support for a dose± response relationship comes from studies in which responses to higher doses of ifosfam ide yield responses after progression on lower doses. For exam ple, ifosfam ide 12 g/m 2 yielded responses in 33% of 36 patients who had progressed after prior doxorubicin and/or ifosfam ide therapy with less than 8 g/m 2 per cycle. 18 It should be noted that not all patients in such studies have been shown to be truly refractory to the lower dosesÐ not all clearly progressed under treatment. F in ally, high respon se rates have been seen in groups of patients treated with intensive anthracycline/ ifo sf am id e reg im en s. 19,20 A lth o u gh the relative contributions of the high doses of ifosfam ide and the high doses anthracycline in the favo rable results in such trials is unclear, the seeming lack of bene® t for dose-escalation of doxorubicin in com bination with modest doses of ifosfam ide. 21 suggests that it is the intensity of ifosfam ide therapy that is responsible. Studies of high-dose therapy are likely to be biased toward younger patients of good perform ance status. Furtherm ore, many investigators have observed that gastro in testin al leiom yo sarcom a is re fractor y to ifosfam ide-based regim ens, and that synovial sarcoma tends to be particularly responsive to such treatm ents. Leiom yosarcom a accounts for 35± 50% of patients in m any large phase II or phase III trials. 12,13 Thus, selection of younger patients with extrem ity sarcoma for phase I and pilot phase II studies m ay be respo nsible, at least in p ar t, fo r the d ifferen ces obser ved in response rate am ong various studies.
A conclusion regarding the relative efficacy of highdose ifosf am id e requires prospective rand om iz ed trials. In sequential phase II trials, increasing the dose of doxo ru b icin fro m 50 m g/m 2 to 75 m g/m 2 in combination w ith infosfam ide 5 gm/m 2 appeared to result in im proved efficac y. 22,23 Yet, in a random ized trial by the sam e group, there was no difference in response or sur vival. 21 In the only random ized trial in which the dose of infosfam ide was the m ajor variable, the EO RT C reported a response rate of 3% am ong patients treated with ifosfam ide 5 g/m 2 by 24-h our infu sion, but 17% am ong patien ts w ho received 9 g/m 2 given by 4-h our infusion daily for 3 days. 24 T he explanation of the low observed response rate observed in the`standard dose' arm rem ains a matter of speculation, but m ay be related to the high proportion of patients w ith leiomyosarcom a am ong the study population. This E O RT C trial did not em p loy th e extrem ely high d oses o f ifo sfa m id e described in other studies, although the results do support the concept of a dose± response relationship. T he strategy generally em ployed to increase the dose-intensity of infosfam ide has been to increase the am ount of drug given ever y three to four weeks.
In our study, the strategy was to increase the frequency of ifosfam ide adm inistration to every two weeks. In so doing, the total dose adm inistered, expressed in gm /m 2 /week was as high as any previously reported regim en. Although acute toxicity, including myelosuppression, was not imm ediately dose-lim iting, in the population treated m ost patients did not receive m ore than four cycles of therapy. D uring the treatment the m ajor ity of patients were able to accept treatm ent on schedu le. N on etheless, these patien ts w ith their advanced disease and w ith a lim ited life-expectancy often stated that continuation of this regim en was inconsistent w ith their goal of m aintaining a good quality of life. Based on current practice at the time this study was planned, the treatm ent was given entirely on an inpatient basis. This alm ost certainly was a confounding factor in patient`burnout' . In addition, the use of erythropoietin might have reduced the fatigue in patients who becam e anem ic during the course of treatment.Today, the regimen we report could be given on an outpatient basis to most individuals. Indeed, after the prelim inary report of the present trial, 25 another group also demonstrated the feasibility o f repeatin g cycles o f ifo sf am id e-con tain ing chemotherapy to patients with sarcoma every 2 weeks. 26 In a phase II trial at M emorial Sloan-K ettering C ancer Center, edatrexate induced responses in 13% of patients w ith advanced sarcom a. 7 Although the activity of edatrexate in a N ational Cancer institute of C anada trial was m arginal, 27 respon ses have been seen in an E C O G trial. 8 A high respon se rate in the latter tr ia l w o u ld p r ov id e im p etu s fo r fu r th e r exploratio n of the com bin ation of ed atre xate an d ifosfa m id e. Perhaps the greatest potential for any active drug o r reg im en lies in ad ju van t th era p y. S eve r al uncontrolled trials 19,20 and one prospective random ized trial 29 suggest that pre-operative or posto p er ative a n th rac yc lin e/ifo sf am id e co m bin atio n chemotherapy regim en results in im proved sur vival in patients with operable soft tissue sarcom a. W hereas d o se-in ten se regim en s are d iffi cu lt to app ly in patients w ith advanced disease, experience in other d i s ea s es d o c u m e n ts t h a t v i g o ro u s , r ep e ti tiv e chem otherapy program s can be tolerated by patients fo r ® n ite period s of tim e. 30 T he p re sen t stu d y dem onstrates the feasibility of adm inistering a d osein ten se schedu le of ifosf am ide alone or ifosf am ide w ith edatrexate.

A cknow ledgem ents
T his work was supported by N IH G rant CA47179.