A Phase II Study of Docetaxel in Chemotherapy-Naïve Patients With Recurrent or Metastatic Adult Soft Tissue Sarcoma

Purpose: To determine the efficacy and toxicity of docetaxel as first-line chemotherapy in adult patients with locally advanced and/or metastatic soft tissue sarcoma (STS). Patients/methods. Thirty eligible patients, with histologically proven STS, Eastern Cooperative Oncology Group (ECOG) performance status 0–2 and bidimensionally measurable disease, entered this study. None had received previous chemotherapy. Docetaxel 100 mg m-2 was given as a 1-h intravenous infusion every 3 weeks. Patients were evaluable for response, evaluated by WHO criteria, after one cycle of chemotherapy and toxicity was graded by NCIC-CTG common toxicity criteria. Results. One hundred and thirty two cycles were aldministered, with a range per patient of 1–9. The median delivered dose intensity was 32.2 mg m-2  weekm-1 (planned 33.3 mg m-2  weekm-1 ) and 67% of patients received ≥90% planned dose intensity. There were three partial responses (10.7%; 95% confidence interval 2.3–28.2) with a median duration of 7 months (range 6.4–8.3 months). Thirty patients were evaluable for non-haematological toxicity and 28 for haematological toxicity (repeat counts were not available in two patients). Haematological toxicity was moderately severe, with 18 (64%) patients experiencing at least one episode of grade 4 neutropenia, and 7 (25%) patients experiencing febrile neutropenia. Conclusions. In this study, activity of docetaxel in adult chemotherapy-naïve patients with advanced STS was modest


Introduction
In the past 10 years, there has been little progress in the m anagement of m etastasized soft tissue sarcom a (ST S). A m inority of patients have long-term bene® t from resection of lung m etastases. 1 N o new active agents have been identi® ed since doxorubicin and dacarbazine in the 1970s, and ifosfam ide in the m id-1980s. 2 D espite encouraging data from singlearm studies, random ized trials have failed to demonstrate convincingly the superiority of com bination chem otherapy 3 com pared with single-agent doxorubicin. Sim ilarly, dose intensi® cation of standard agents or regim ens, which have produced prom ising response rates in pilot studies, have not been proven m ore active in random ized trials. 4 The need for active new agents is clear.
To optim ize the possibility of identifying new active drugs for this disease, the Canadian Sarcom a G roup working with the N C IC C linical Trials G roup has pursued for several years, a policy of offering investigational drug treatment to chem otherapy-naõ È ve patients, with the option to proceed to standard drugs if initial treatment fails. D ocetaxel (Taxotere) is a sem i-syn thetic analogue of paclitaxel, prepared using a precursor extracted from the needles of the E uropean yew, Taxus baccata. 5 By enhancing m icrotubule assem bly and inhibiting depolym erization of tubulin, docetaxel causes accum ulation of m icrotubules in the cell and, by blocking cells in the M -phase of the cell cycle, prevents cell division. 6 Based on m ore potent in vivo antitum our activity com pared with paclitaxel, in B16 m elanom a and a variety of colon carcinom as, 7 as well as a favou rable toxicity pro® le in anim als, docetaxel w as selected for hum an studies. Com paring the toxicities observed in the ® ve hum an phase I trials, the highest m axim um tolerated dose (MT D ) and dose intensity were achieved with the 1-h 3-w eekly infusion schedule. 8  . It was given at a starting dose of 100 mg m 2 2 , diluted in 250 ml of 0.9% sodium chloride or 5% dextrose solution, infused over 1 h, repeated every 3 weeks. All patients were premedicated for possible hypersensitivity with 20 mg of dexamethasone orally 12 and 6 h prior to docetaxel, and 50 mg of diphenhydramine and 50 mg of ranitidine (or 300 mg of cimetidine) given slowly intravenously prior to docetaxel. D oses w ere adjusted for m yelosuppression based on nadir blood counts. For an abso lute granulocyte count of , 0.5 3 10 9 /l for 7 days or a platelet count , 25 3 10 9 /l, febrile neutropenia, $ grade 3 infection or bleeding requiring transfusion, the dose in the next cycle w as reduced by 25% . If the treatm ent day count show ed an absolute granulocyte count of , 1.5 3 10 9 /l or platelet count of , 100 3 10 9 /l, treatment was delayed until recovery. Haem atopoietic colony stim ulating factors were not used. If patients developed grade 3 peripheral neuropathy, they were taken off protocol therapy, and a dose reduction of 25% was applied for grade 2 neurotoxicity. For other toxic effects $ grade 3, drug administration was delayed until resolution to # grade 1, and then reinstituted with a dose reduction of 25%. O nce doses had been reduced for toxicity, they were not re-escalated.

Patient evaluation
Patients were evaluable for response after receiving one cycle of therapy (3 weeks on the study). Responding patients continued on the study until evidence of disease progression, or the occurrence of unacceptable toxicity. Patients with com plete (C R) or partial (PR) response continued on therapy for a m axim um of four cycles after docum entation of response. Patients with stable disease (SD ) continued on therapy for a m axim um of six cycles. W H O criteria for response were em ployed and toxic effects w ere graded according to the N C IC -CT G comm on toxicity criteria.

Results
Between April 1994 and June 1995, 32 patients were enrolled from eight Canadian centres. T wo patients were ineligible: one had no measurable lesions and in the other the histology was not ST S. T wo further patients were not evaluable for response and haem atological toxicity. T he condition of one patient deteriorated suddenly w ith onset of congestive heart failure, renal failure and sepsis, and she died 6 days after the ® rst dose of docetaxel. T he second patient did not com plete the ® rst dose of docetaxel because of a severe hypersensitivity reaction and went off study. Table 1 describes the characteristics of eligible patients. A m ajority of patients (18) had lung m etastases, other com m on sites of disease being nodes, soft tissue and within the abdom en/pelvis/retroperitoneum .
One hundred cycles were adm inistered at the starting dose of 100 m g m 2 2 every 3 weeks, and 26 cycles at the reduced dose level of 75 m g m 2 2 every 3 weeks (a 25% dose reduction). One and ® ve cycles were adm inistered at dose levels 55 and , 50 m g m 2 2 respectively. T he num ber of cycles per patient ranged from one to nine. T he planned dose intensity was 33.3 m g m 2 2 week 2 1 and the achieved m edian dose intensity was 32.24 m g m 2 2 week 2 1 (range   Table 2). F ebrile neutropenia was reported in seven (25% ) patients. Two other patients had a docum ented grade 4 infection not thought to be related to therapy. In 30 evaluable patients, the m ost frequently observed non-haem atological toxicities thought to be drug related were alopecia (77% ), lethargy (60% ), sensory neuropathy (43% ), oedema (27%), nausea (33%), vom iting (27% ), anorexia (23%) and diarrhoea (40% ), but m ost of these were m ild (Table 3).
In 28 evaluable patients, there were three PRs (10.7% ; 95% con® dence interval (CI) 2.3± 28.2% ) w ith a m edian duration of 7 months (range 6.4± 8.3 m onths). Thirteen patients had lasting 5.4 m onths (range 2.1± 20.9 m onths). T he m edian num ber of cycles received by these patients was six (range 3± 14). Two of the PRs occurred in patients with leiom yosarcom a (one uterine, one retroperitoneal), the m ost com m on histological type entered into the study. T he third response occurred in a patient with m alignant ® brous histiocytom a. T wo of the responses occurred in lung m etastases, and the third in a retroperitoneal m ass.

D iscussion
The taxanes, paclitaxel and docetaxel, are a new class of cytotoxic drugs with activity in a variety of solid tumours. Paclitaxel was the ® rst analogue to be tested, and was shown to be active in platinum-resistant ovarian cancer 9 and also in breast cancer. 10 A dif® cult extraction process, from the bark of the Paci® c yew, initially limited supplies and accelerated the search for semi-synthetic analogues, such as docetaxel.
In phase II trials, docetaxel has shown signi® cant activity, w ith reproducible response rates of greater than 20% , in breast cancer, non-sm all cell lung, ovarian , gastric, squam ous head and neck and bladder cancers. 11 Both analogues have been evaluated in advanced/ m etastatic ST S. T here have been two studies of paclitaxel at a dose of 250 m g m 2 2 every 3 weeks. In the larger South W est O ncology Group (SW O G) study 12 , there was one C R and 5 PRs for an overall response rate of 12.5% (95% CI 4.7± 25.3%), in 48 patients none of whom had received prior chemotherapy. A lower response rate of one PR (3.6% ) in 28 patients, 11 of w hom had received prior chemotherapy, w as reported by W altzman et al. 13 G ian et al.
14 did not observe any responses in 10 patients w ith STS (four had received prior chemotherapy) treated with 200 m g m 2 2 every 3 weeks of paclitaxel.
In contrast, Van Hoesel et al., 15 for the European O rganization for Research and Treatm ent of Cancer (EO RTC ), observed ® ve PRs in 29 patients (17.2%; 95% C I 6± 36%) treated with docetaxel 100 m g m 2 2 every 3 weeks, all of whom had received previous chem otherapy. These encouraging results prom pted the EO RTC to undertake a random ized phase III study com paring 100 m g m 2 2 every 3 weeks of docetaxel with 75 m g m 2 2 every 3 weeks of doxorubicin. T his study has just been reported in abstract form . 16 W ith 86 patients on study and 73 available for analysis, the prelim inary response rates were 0% for docetaxel and 27% for doxorubicin, the latter being very sim ilar to previous EO RT C experience w ith single-agent doxorubicin.
The current study was started shortly after publication in abstract form 17 of the preliminary results of the EORT C phase II study describing, at that time, ® ve PRs (21.7% ) in 23 patients. W e hoped to con® rm , and perhaps im prove on, these results in chem otherapy-n aõ È ve patients. The low response rate seen in our study w as surprising, but is less so now that it is seen in the context of the m ore recent EO RTC phase III results. 16 It is intriguing that in a different tum our type, sm all-cell lung cancer, a higher response rate of 25% 18 was seen in previously treated patients receiving docetaxel, com pared with 8% 19 in a second study including only chemotherapy-naõ È ve patients. As the EO RTC study 17 includes a cross-ove r com ponent this m ay provide further data on this issue.
Thirteen patients achieved SD . Rapidly progressive disease at entry was not a protocol requirement, but m ost Canadian oncologists use chemotherapy in the setting of progressive sym ptom atic m etastatic disease, and this disease stabilization m ay be an additional indicator of drug activity.
It has been suggested that leiom yosarcom as, particularly those of bowel m etastatic to liver, are less responsive to chem otherapy. However, leiom yosarcom as (including those of gastrointestinal origin now known as strom al tum ours) m ay com prise onethird to one-half of tum ours in clinical trials of m etastatic STS. T his is true of our study (39% ), the phase II 15 and III 17 studies of the EO RTC (41%; 36% ) and for the SW OG 12 and W altzm an studies 13 of paclitaxel, 54% and 53% respectively. This sim ilar distribution does not provide a reason for the discrepant response rates. T he m ost likely explanation is chance occurrence in sm all patient populations.

Conclusions
D espite initial prom ise, data from this study, and m ore recent phase II and III studies, do not suggest that taxanes have m ajor activity in adult ST S.

A cknowledgem ents
A study of the N ational C ancer Institute of Canada C linical T rials Group and the C anadian Sarcoma G roup. W e would like to thank W . Bogues, D . C harp entier, J. Latrielle, F. Letendre, J. Jensen, M . Knowling, D . Stein, M . Thirlwell and L. Yelle w ho also contributed patients to this study. Supported by a grant from Rho Ã ne-Poulenc-R orer Ltd.