Gemcitabine and docetaxel as front-line chemotherapy in patients with carcinoma of an unknown primary site




The current study was performed to evaluate the efficacy and toxicity of a noncisplatin-based chemotherapy regimen combining gemcitabine and docetaxel as front-line chemotherapy for patients with carcinoma of an unknown primary site.


Patients were to receive intravenous gemcitabine at a dose of 1000 mg/m2 over 30 minutes on Days 1 and 8 and docetaxel at a dose of 75 mg/m2 over 1 hour on Day 8 in an outpatient setting. The schedule was repeated every 3 weeks for a maximum of 6 cycles.


Thirty-five patients were assessable for response and survival. One complete and 13 partial responses were observed. The overall response rate was 40% (95% confidence interval, 28–52%). The median time to disease progression was 2 months (range, 1–4 months). The median overall survival time was 10 months (range, 0–32 months). Toxicity was reported to be manageable.


The combination of gemcitabine and docetaxel was found to be active in patients with carcinomas of an unknown primary site. However, the overall outcome of these patients remains poor and novel treatment approaches are required. Cancer 2004. © 2004 American Cancer Society.

Carcinomas of unknown primary site (CUP) represent a group of heterogeneous tumors that share the unique clinical characteristic of metastatic disease with no identifiable origin at the time of therapy. During the last decade, the identification of patient subsets with clinical and pathologic features requiring specific guidelines that may translate into prolonged survival, the standardization of diagnostic procedures, and the identification of reliable prognostic factors have improved the daily management of patients with CUP.1–4 However, overall outcome remains poor, with a median survival ranging from 6–12 months. Therefore, new treatment drugs are required. The recent introduction of several cytotoxic agents with broad spectra of clinical activity such as gemcitabine, irinotecan, and taxanes has created opportunities for the empiric treatment of CUP.5–9 In the current study, we report on the results of a Phase II study assessing the efficacy and toxicity of a chemotherapy regimen combining gemcitabine with docetaxel.



The current study was conducted at the Montpellier Cancer Center between December 2000 and April 2003. Eligibility criteria included age > 18 years; histologically confirmed CUP (well or poorly differentiated adenocarcinoma, poorly differentiated carcinoma); no previous chemotherapy; an Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1, or 2; measurable disease; an absolute neutrophil count (ANC) ≥ 1500/μL, a platelet count ≥ 100,000/μL, a serum creatinine concentration ≤ 1.25 times the upper limit of normal or creatinine clearance > 50 mL/mn, a serum bilirubin level ≤ 1.25 times the upper limit of normal; a cardiac ejection fraction > 50% by echocardiographic or radionuclide cineangiography; and the provision of informed consent. Patients were excluded if they had any of the following features: inclusion in subsets with specific well defined treatment (i.e., women with adenocarcinoma that involved only axillary lymph nodes, women with papillary serous carcinoma of the peritoneum, patients with squamous carcinoma that involved only cervical or inguinal lymph nodes, patients with carcinoma with neuroendocrine features, and patients with carcinoma that involved a single potentially resectable tumor site), symptomatic brain metastases, a history of previous malignancy with the exception of skin cancer or cervical carcinoma in situ, pregnant or lactating women, and women with severe coexistent medical illnesses.

Pretreatment Evaluation

Patients were required to undergo at least the following procedures: thorough history and physical examination; chemistry profile including in men the serum tumor markers prostate specific antigen, α-fetoprotein, and human chorionic gonadotrophin; chest roentgenograms; computed tomography scan of the abdomen and pelvis; mammography in women; and directed radiologic work-up of any symptomatic areas. In addition, specific pathologic evaluation was required for patients with a light microscopic diagnosis of poorly differentiated carcinoma to exclude, if possible, other malignancies. Immunoperoxidase staining with antibodies directed against leukocyte common antigen, cytokeratins, neuroendocrine markers (chromogranin and synaptophysin), and melanoma markers (S-100 protein and HMB-45) was recommended.

Treatment and Patient Monitoring

Patients were to receive intravenous (i.v.) gemcitabine at a dose of 1000 mg/m2 over 30 minutes on Days 1 and 8, and i.v. docetaxel at a dose of 75 mg/m2 over 1 hour on Day 8. The schedule was repeated every 3 weeks for a maximum of 6 cycles. Premedication with oral corticosteroids before and after the administration of docetaxel from Day 7 to Day 9 was mandatory. A treatment delay on Day 8 was allowed if necessary until the ANC reached a minimum of 1500/μL and the platelet count reached a minimum of 100,000/μL. Routine laboratory tests including electrolytes, creatinine, total protein, albumin, calcium, glucose, alkaline phosphatase, total and direct bilirubin, aspartate aminotransferase, alanine aminotransferase, and prothrombin time were evaluated on the first day of each course of chemotherapy. Additional complete blood cell counts were obtained weekly. Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria scale (Version 2). A complete reassessment of all metastatic sites was planned every three cycles of chemotherapy. Patients were assigned a response category based on World Health Organization standard definitions.10 Responding patients or those with nonprogressive lesions could receive additional courses of chemotherapy according to the discretion of the treating physician. After the completion of therapy, patients were monitored at 3-month intervals until progression of disease. Follow-up evaluation was performed until the time of death.

Statistical Analysis

The primary endpoint of the trial was the objective (complete plus partial) response (OR) rate. Patients were accrued using a two-stage, Phase II design. An OR rate of at ≥ 40% was required to conclude that chemotherapy was effective and < 25% was required to conclude that it was inactive. Seventeen patients had to be enrolled during the first stage of the study. If < 5 ORs were observed, no further patients would be accrued; if ≥ 5 ORs were observed, 18 additional patients were to be included. Based on 35 patients, chemotherapy would be considered as ineffective if < 14 ORs were observed and effective if ≥ 14 ORs were observed. Treatment results were expressed as the percentage with 95% confidence intervals (95% CI) or as the median and range. Survival was calculated using the Kaplan–Meier method.


Patient Population

Thirty-six patients, including 23 men and 13 women, were entered onto the study. The median age of the patients was 56 years (range, 31–78 years). Other patient characteristics are shown in Table 1. Adenocarcinoma was the most common histologic diagnosis. Ten patients (28%) were found to have only 1 metastatic site. According to the prognostic model developed by the French Study Group on Carcinomas of Unknown Primary (GEFCAPI),4 25 patients (69%) had good-risk characteristics (an ECOG performance status ≤ 1 and a normal serum lactate dehydrogenase level).

Table 1. Patient Characteristics
CharacteristicsNo. of patients
  1. ECOG: Eastern Cooperative Oncology Group; LDH: lactate dehydrogenase; GEFCAPI: French Study Group on Carcinomas of Unknown Primary.

 Well differentiated adenocarcinoma9
 Poorly differentiated adenocarcinoma8
 Poorly differentiated carcinoma12
No. of metastatic sites 
 > 212
Metastatic sites 
 Mediastinal lymph nodes16
 Retroperitoneal lymph nodes7
 Adrenal glands4
 Supraclavicular lymph nodes2
 Cutaneous lymph nodes2
Elevated serum tumor markers 
 Carcinoembryonic antigen11
 CA 19-910
 CA 15-35
 CA 1255
Performance status (ECOG scale) 
Serum LDH 
GEFCAPI prognostic subgroups 
 Good risk25
 Poor risk11

Treatment Delivery and Toxicity

A total of 149 courses of chemotherapy was administered, with a median of 5 cycles per patient (range, 1–6 cycles). The median doses delivered are depicted in Table 2. The median interval time between cycles was 21 days (range, 21–28 days). Toxicity data were available for 138 courses. Grade 3-4 toxicities are shown in Table 3. Neutropenia was predominant (occurring in 35% of the courses). Grade 3-4 thrombocytopenia and anemia were quite uncommon. Only one patient required a red blood cell transfusion during therapy. Other Grade 3-4 toxicities included dermatitis, stomatitis, edema, diarrhea, and asthenia, which were observed in one, one, one, two, and three patients, respectively. Six patients had to be withdrawn from therapy because of toxicity: one patient because of fatal septic shock after the first cycle, one patient for extensive edema (involving the lower and upper limbs, abdomen, pleura, and brain) after the fourth cycle, one patient for widespread dermatitis after three courses, and three patients for asthenia after Cycle 3, Cycle 4, and Cycle 5, respectively. With the exception of the only patient who died of toxicity (septic shock), Grade 3-4 toxicities resolved after the cessation of chemotherapy. Overall seven patients required hospitalization for the management of treatment-related side effects (five patients because of febrile neutropenia and two patients because of edema and dermatitis, respectively).

Table 2. Treatment Delivery
Cycle no.Patient no.Median doses (mg) (range)
Day 1Day 8Day 8
1361830 (1450–2000)1800 (0–2000)135 (0–150)
2321790 (1440–2000)1790 (1440–2000)135 (110–150)
3281800 (1440–2000)1800 (1120–2000)135 (85–150)
4251790 (1110–2000)1790 (1110–2000)135 (85–150)
5181780 (1130–2000)1780 (1130–2000)130 (85–150)
Table 3. Grade 3 and 4 Toxicitiesa
ParametersGrade 3Grade 4
  • a

    Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria.


Responses and Survival

Thirty-five patients were assessable for response and survival. One complete and 13 partial responses were observed (Table 4). Therefore, the OR rate was 40% (95% CI, 28–52%). A decrease in the serum tumor marker levels of ≥ 50% was observed in 16 patients (62%). At the end of chemotherapy, three patients who achieved a partial response received additional radiotherapy to their mediastinal lymph nodes. At the time of analysis, the median follow-up time was 22 months, and the median time to disease progression was 2 months (range, 1–4 months). The median overall survival time was 10 months (range, 0–32 months). The actuarial 1-year and 2-year survival rates were 43% and 7%, respectively (Fig. 1). Four patients remained free of progressive disease at 6 months, 12 months, 21 months, and 32 months, respectively, from the time of diagnosis.

Table 4. Efficacy
Complete response1
Partial response13
Stable disease4
Progressive disease17
Not assessable1
Figure 1.

Actuarial survival time.


The design of a treatment plan for patients with CUP remains a daily challenge for physicians. With the exception of the subsets of patients with clinical and pathologic features requiring specific guidelines that may translate into prolonged survival, to our knowledge no chemotherapy has been firmly established as the gold standard, first-line regimen.1, 3 Response rates ranging from 6–30%, along with median survival times ranging from 4–11 months, have been reported in series of patients treated with noncisplatin-based regimens (Table 5).11–19 The role of cisplatin remains a matter of debate. To our knowledge, only one of the three randomized trials published to date that compared a cisplatin-based versus a noncisplatin-based regimen showed a survival advantage for patients treated with the cisplatin-containing chemotherapy.15, 19, 20 However, the very low number of patients included in these randomized trials prevented us from drawing firm conclusions for daily practice.

Table 5. Results with Noncisplatin-Based Chemotherapy in Patients with CUP
RegimenPatient no.Objective response rate (%)Median survival (mos)AuthorsYear of publication
  1. F: 5-fluorouracil; NR: not recorded; A: doxorubicin; Vi: vindesine; M: mitomycin-C; D: dacarbazine; V: vincristine; B: carmustine; C: cyclophosphamide; G: gemcitabine; T: docetaxel.

F656NRJohnson et al.111964
F88164Moertel et al.121972
AVi38137Fiore et al.131985
FAM433011Goldberg et al.141986
AM51354.5Milliken et al.151987
DVFB61194Alberts et al.161989
AMV57307Kambhu et al.171990
VAC4084de Campos et al.181994
M39175Falkson and Cohen191998
GT364010Pouessel et al.Current series

During the last decade, new cytotoxic agents with a wide spectrum of activity have emerged from preclinical studies. Particularly promising results with gemcitabine alone or combined with docetaxel have been reported in the second-line management of patients with CUP.21, 22 In the current series, a 40% response rate was observed among 36 patients who were treated with a combination of gemcitabine and docetaxel as first-line therapy. The median survival was 10 months. These results are in global accordance with those observed in previous series combining the new drugs developed in recent years and a platinum salt (Table 6). In addition, overall toxicity was manageable and allowed outpatient treatment of the majority of patients.

Table 6. Results with New Cytotoxic Agents in Patients with CUP
RegimenPatient no.Objective response rate (%)Median survival (mos)AuthorsYear of publication
  1. Pa: paclitaxel; Ca: carboplatin; E: etoposide; T: docetaxel; P: cisplatin; G: gemcitabine; I: irinotecan.

PaCaE716311Greco et al.52000
PaCa773913Briasoulis et al.72000
TCa47228Greco et al.62000
PaCaG113259Greco et al.82002
IP40386Culine et al.92003
GT36408Pouessel et al.Current series

Clearly, whatever the chemotherapy regimen used, the cytotoxic agents developed during the last decade do not appear to have improved the prognosis of CUP patients significantly. Further prospective trials led by the French National group will have to define optimal standard treatments according to those prognostic groups defined by the GEFCAPI model.4 In good-risk patients, a trial will compare the impact on survival of single-agent cisplatin versus the combination of gemcitabine and cisplatin. In patients with poor-risk disease, low-toxicity chemotherapy will be compared with best supportive care only. Another promising new direction could be a better understanding of the biology of the disease and the development of targeted therapies. Hopefully, these approaches will be the next steps to contribute to the improved management of disease in patients with CUP.