| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
a The Sarah Cannon Cancer Center and Tennessee Oncology, PLLC, Nashville, Tennessee, USA; b South Texas Oncology and Hematology, PA, San Antonio, Texas, USA; c Northwest Georgia Oncology Centers, PC, Marietta, Georgia, USA
Correspondence: F. Anthony Greco, M.D., Sarah Cannon Cancer Center, 250 25th Avenue North, Suite 110, Nashville, Tennessee 37203, USA. Telephone: 615-986-4300; Fax: 615-986-0029; e-mail: fgreco{at}tnonc.com
 |
ABSTRACT |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Patients and Methods. One hundred thirty-two patients were treated with sequential combination chemotherapy for a maximum of six cycles. All patients had relatively poor prognostic features. Fifty-nine patients had well-differentiated adenocarcinoma, 73 patients had poorly differentiated carcinoma, and 121 patients had performance status scores of 0 or 1.
Results. Thirty-three (30%) of 111 assessable patients (95% confidence interval 27%–33%) had objective responses to treatment (26 partial responses, seven complete responses). The combination of gemcitabine and irinotecan was associated with significantly less toxicity than the triple-drug regimen and improved the responses in several patients (10%). The response rates were similar in the two major histologic tumor types, but were lower for patients with liver-dominant tumors (13%) and higher for patients with lymph-node-dominant tumors (50%). The median progression-free survival time, median survival time, and actuarial survival rates at 1 and 2 years were 5.7 months, 9.1 months, 35%, and 16%, respectively.
Conclusions. Sequential combination chemotherapy with paclitaxel/carboplatin/oral etoposide and gemcitabine/irinotecan is an active treatment for patients with carcinoma of unknown primary site, but overall toxicities are greater than those seen with other combinations of new drugs and survival appears similar to that observed in 264 other patients treated in our four previous phase II trials. A better understanding of the biology of these heterogeneous tumors will likely lead to improved therapy for these patients.
Key Words. Carcinoma • Unknown • Chemotherapy • Sequential • Metastasis • Survival
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Based upon these observations, we evaluated the feasibility and efficacy of first-line treatment with paclitaxel/carboplatin/etoposide followed by gemcitabine/irinotecan in patients with carcinoma of unknown primary site. We report here the results of this multicenter phase II study, conducted in the MPCRN.
|
PATIENTS AND METHODS |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
-fetoprotein, in those with PDC, and prostate-specific antigen (in men).
Patient subsets known to have relatively good prognostic features were excluded from this study. These subsets include women with adenocarcinoma involving only axillary lymph nodes or the peritoneal cavity; patients with squamous cell carcinomas involving only cervical or inguinal lymph nodes; patients with poorly differentiated carcinomas consistent with a germ cell tumor (i.e., isolated midline adenopathy/masses and/or multiple pulmonary nodules and/or elevated levels of ß-human chorionic gonadotropin or -fetoprotein), men with prostate-specific antigen elevated in their plasma or stained in their tumors, patients with single potentially resectable tumors, patients with neuroendocrine carcinomas, and patients with poorly differentiated neoplasms, not otherwise classified.
A specific pathologic study to exclude other neoplasms was required for each patient with the initial diagnosis of PDC. Immunoperoxidase staining for leukocyte common antigen was required in all patients to exclude lymphoma. Additional immunoperoxidase stains, including epithelial markers (cytokeratins), neuroendocrine markers (synaptophysin, chromogranin), and melanoma markers (S-100, Hmb-45) were also performed routinely. For men with adenocarcinoma, staining for prostate-specific antigen was performed when clinical features suggested prostate carcinoma. Patients with specific neoplasms identified by special stains were excluded. The pathology reports were reviewed on every patient by two of the authors (F.A.G. and J.D.H.), but no central pathology review was done.
Previous chemotherapy was not allowed. Other entry criteria included an Eastern Cooperative Oncology Group (ECOG) performance status score of 0, 1, or 2, adequate bone marrow function (white blood cells 
4,000/µl; platelets 100,000/µl), normal liver function (bilirubin 
1.5 mg/dl), and normal renal function (creatinine 1.5 mg/dl). Patients were excluded if they had any features as follows: recent myocardial infarction, congestive heart failure, other severe coexistent medical illnesses, a history of previous malignancy within 5 years with the exception of skin carcinoma or cervical carcinoma in situ, pregnant or lactating, and brain or meningeal metastases. The current study was approved by the appropriate Institutional Review Boards, and patients were required to give written informed consent before they participated in the study.
Treatment
The treatment received by patients and evaluation schedule in this trial are outlined in Table 1
. All patients began treatment with two courses of paclitaxel/carboplatin/etoposide (regimen A) as follows: paclitaxel, 200 mg/m2, 1-hour i.v. infusion, day 1; carboplatin, at area under the concentration-time curve (AUC) of 6.0, i.v. infusion, day 1; and etoposide, 50 mg (total dose) alternating with 100 mg (total dose) orally daily, days 1–10. Treatment courses were repeated at 21-day intervals. All patients received premedication with i.v. dexamethasone (Decadron®; Merck and Company, Inc.; West Point, PA), 20 mg, diphenhydramine (Benadryl®; Pfizer Pharmaceuticals; New York, NY), 50 mg, and cimetidine (Tagamet®; GlaxoSmithKline; Research Triangle Park, NC), 300 mg, given 30 minutes before paclitaxel administration. The carboplatin dose was calculated by the Calvert formula as follows: dose = ([glomerular filtration rate + 25] x 5). The glomerular filtration rate (GFR) was calculated by the Cockroft and Gault method using the following serum creatinine measurement: GFR = ([140 – age] x weight in kg) ÷ (72 x serum creatinine) x either 0.85 (females) or 1.00 (males).
|
After completing two cycles of regimen B, patients were once again reevaluated for response (week 12). At that time, patients with objective responses or stable disease received two additional courses of treatment. Selection of the treatment regimen for the fifth and sixth courses was at the discretion of the treating physician, based on the relative efficacy and toxicity of regimen A versus regimen B during the first four courses of treatment. In general, if one regimen appeared more efficacious or better tolerated, it was selected for the fifth and sixth courses of treatment. If the two regimens were similar, one additional course of each regimen was recommended. Patients with progressive disease after four courses were removed from study and treated at the discretion of their physician. All chemotherapy was stopped after a total of six cycles. Patients were reevaluated for a final time at week 15 and then observed off therapy and reevaluated as clinically indicated.
During therapy, blood counts were monitored weekly. Dose modifications for myelosuppression were based on blood counts measured on the day of scheduled treatment. The day 1 dose modifications for regimen A (paclitaxel/carboplatin/oral etoposide) were as follows: absolute neutrophil count (ANC) 1,500/µl and platelets 100,000/µl, 100% doses; ANC 1,000–1,500/µl or platelets 75,000–100,000/µl, 75% doses; ANC <1,000/µl or platelets <75,000/µl, treatment delay for 1 week or until ANC and platelets rose above these levels, then 75% doses. If there was a 3-week delay, the patient was taken off study. On day 8, if the ANC was 1,500/µl and the platelet count was 75,000/µl, the 10-day course of oral etoposide was completed. If the respective ANC and platelet counts were less than as stated, the remaining etoposide (i.e., days 8–10) was omitted. Any hospitalization for neutropenia and fever dictated a 25% dose reduction of paclitaxel and carboplatin and a reduction in the dose of oral etoposide to 50 mg total dose daily. Reversible grade 3/4 nonhematologic toxicities, with the exception of nausea and vomiting and alopecia, required a 25% dose reduction of the responsible agent or agents. If grade 3/4 nonhematologic toxicity was irreversible or had not resolved to grade 1 after a treatment delay of 3 weeks, the patient was removed from study. Patients who experienced severe (grade 3 or 4) acute hypersensitivity reactions related to either paclitaxel or carboplatin were removed from the study.
For regimen B (irinotecan/gemcitabine), dose modifications on days of scheduled treatment were as follows: ANC >1,500/µl and platelets >100,000/µl, 100% doses; ANC 1,000–1,500/µl or platelets 75,000–100,000/µl, 75% doses; ANC <1,000/µl or platelets <75,000/µl, dose delay for 1 week or until ANC >1,000/µl and platelets >100,000/µl, then further doses at 75% for both drugs. If more than a 3-week delay was required, the patient was taken off study and treated at the discretion of their physicians. Hospitalization for neutropenia and fever required a 25% dose reduction of both drugs for all remaining doses. Patients with reversible grade 3/4 nonhematologic toxicities had 25% dose reductions of the offending agent(s). Prolonged grade 3/4 nonhematologic toxicities for 3 weeks required that the patient be taken off study. For diarrhea, standard prophylactic loperamide (Imodium®; McNeil Consumer and Specialty Pharmaceuticals; Fort Washington, PA) was administered for grade 1 or 2. For grade 3 or 4 diarrhea, irinotecan was withheld and intensive therapy for diarrhea was initiated; irinotecan was resumed at the 75% dose when diarrhea improved to grade 1 or less.
Cytokines were not used prophylactically during the first course of treatment for any patient. Subsequent use of cytokines was at the discretion of the treating physician. In particular, if it was deemed important to continue treatment on schedule in a patient who had demonstrated an early response to therapy, cytokines could be used prophylactically to maintain dose and schedule.
Assessment of Treatment Efficacy
Patients were evaluated for tumor response three times during treatment (Table 1
). These evaluations occurred after completing two courses of regimen A (paclitaxel/carboplatin/oral etoposide), after completing two courses of regimen B (irinotecan/gemcitabine), and at the completion of all treatment (six courses). At each reevaluation, a response category was assigned based on standard World Health Organization (WHO) definitions. The complete response category required the total disappearance of all clinically and radiologically detectable disease for at least 4 weeks. The partial response category required at least a 50% reduction in the size of all measurable lesions as measured by the product of the greatest length and the greatest width, with no new lesions appearing. Patients in the stable disease category had reductions of <50% or increases of <25% in the size of measurable lesions, with no new lesions appearing. Patients assigned to the progressive disease category had development of any new lesions or an increase of 25% or more in the size of any existing measurable lesion.
All patients who received at least one dose of treatment were included in the evaluation of treatment-related toxicity. Toxicities were graded according to the National Cancer Institute Common Toxicity Criteria, Version 2.0.
Statistical Considerations
The primary goals of the present study were to determine the toxicity, objective tumor response rate, progression-free survival (PFS) duration, and overall survival duration in patients receiving the sequential administration of paclitaxel/carboplatin/etoposide and irinotecan/gemcitabine. It remains difficult to determine the number of patients necessary for a phase II trial involving carcinoma of unknown primary site. Most accrual designs for phase II trials evaluate patients with relatively homogeneous neoplasms. In contrast, carcinoma of unknown primary site demonstrates extreme histologic and clinical heterogeneity, often with variable responsiveness to cytotoxic chemotherapy. Therefore, we took the liberty of arbitrarily adding a total of 80 patients to the two-staged (20 patients in stage 1 and 60 patients in stage 2) accrual design method of Simon [6]. We estimated this number would dilute the problem of extreme heterogeneity and variable responsiveness to cytotoxic therapy in this patient group to some degree and provide a better estimation of the toxicity, objective response rate, and PFS time. We expected at least a 20% response rate, and if fewer than eight of the first 40 patients entered in this trial (stage 1) failed to obtain objective responses, the study would end. If eight or more responses were seen, an additional 90 patients (stage 2) would be accrued. A secondary objective was to determine survival following therapy.
All patients were followed up until the time of progression or death. PFS was defined as the interval from the first day of therapy until tumor progression was documented. Actuarial PFS and survival curves were constructed using the Kaplan and Meier method [7].
All patient data (including copies of all primary source documents) were reviewed at the data center of the MPCRN at the Sarah Cannon Cancer Center. All response and toxicity validations were done by reviewing primary source documents.
|
RESULTS |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
. The majority of patients had good performance status scores; 68% had either adenocarcinoma or PDA. Most patients (76%) had metastases in one or more visceral sites; however, 24% had dominant tumor sites in lymph nodes or soft tissue. Patients were treated at the Sarah Cannon Cancer Center or by participating affiliate members of the MPCRN (Appendix 1).
|
Efficacy
In the first stage of this trial, in 40 assessable patients, nine attained objective responses (22%) and the second stage of this study was initiated. Twenty-one of 111 patients (19%) had objective tumor responses after two cycles of regimen A (19 partial responses, two complete responses). Sixty-eight patients remained stable and 22 patients had progressive tumors.
Regimen B was delivered to 93 patients. Ten (10%) additional objective responses were seen, including eight partial responses (all stable after regimen A) and two complete responses (both stable after regimen A). Forty-two patients remained stable, and 25 patients had progressive tumors. Of the 11 patients with progressive tumors after regimen A who received regimen B, none had an objective response.
A total of 65 patients began cycles 5 and 6 with an option of two cycles of regimen A, two cycles of regimen B, or one cycle of each regimen. There were four patients showing additional improvements in tumor responses, including one partial response (stable after regimen B) and three complete responses (one stable, two partial responses after regimen B). Thirty-nine patients remained stable and one patient had progressive tumor.
The best overall responses during therapy for the 111 evaluable patients were as follows: 33 major responses (30%), including 26 (23%) partial responses and seven (7%) complete responses. Thirty-five patients (31%) had stable disease as their best response, and 43 (39%) had progressive disease (Table 3). The response duration ranged from 3–33+ months (median = 7 months).
|
). Patients with predominantly lymph node tumors had a 50% response rate, while those with liver metastases had a response rate of only 13%. There were no differences in response rate based on tumor histology (i.e., adenocarcinoma versus PDC).
All 132 patients were included in the survival analysis. The PFS analysis is illustrated in Figure 1. The median PFS duration was 5.2 months (95% confidence interval [CI] = 4.4–6.0), with a 1-year PFS rate of 16% and a 2-year PFS rate of 6%. The actuarial survival curve for all patients is shown in Figure 2. The minimum follow-up was 15 months (range 15–36 months). The median survival duration was 9.1 months (95% CI = 8.1–10.2), with a 1-year survival rate of 35% and 2-year survival rate of 16%. Figure 3 compares the survival analysis for patients treated in this study with the updated survival analysis for the 264 patients reported in our previous four sequential phase II trials of taxane- or gemcitabine-based chemotherapy [1–4]. There is no significant difference (log rank p = 0.2361) in these curves. For the previous 264 patients in four trials, the minimum follow-up was 30 months (range 30–96 months) and the median survival duration was 9.2 months, with survival rates at 1, 2, 3, 5, and 8 years of 38%, 19%, 16%, 10%, and 8%, respectively. The PFS analysis from those four trials is compared with the analysis from this study in Figure 4, and the results of these two analyses also do not appear to differ (log rank p = 0.6836).
|
|
|
|
outlines the grade 3 and 4 toxicities seen with regimen A (paclitaxel/carboplatin/etoposide) and regimen B (gemcitabine/irinotecan). A total of 583 courses of chemotherapy was administered, including 307 courses of regimen A and 276 courses of regimen B. Regimen A was more toxic, and myelosuppression represented the major toxicity. Grade 3/4 leukopenia was seen in 71% of patients receiving regimen A versus 28% of those receiving regimen B (p = 0.001). Fever and neutropenia requiring hospitalization were seen in 17 patients on regimen A (13%) and in only one patient (1%) on regimen B (p = 0.001). Five treatment-related deaths were seen on regimen A and none were seen on Regimen B. Documented infections were seen in 19 patients (14%) on regimen A and in five patients (5%) on regimen B (p = 0.01). Other nonhematologic toxicities were fairly similar between the two regimens, with the exceptions of a significantly greater incidence of diarrhea on regimen B (p = 0.01) and less arthralgia/myalgia on regimen B (p = 0.01).
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The overall response rate seen in this sequential phase II trial was 30%, with 33 responses, including 26 partial responses and seven complete responses. Responses varied according to the dominant site of metastatic disease (Table 3), and about half the evaluable patients had liver-dominant disease, a site known to be associated with lower response rates and a relatively poor prognosis [8]. The overall PFS duration and survival duration of the 132 patients entered on this large phase II trial were very similar to those found in our previous four phase II trials in a total of 264 patients [1–4] treated with taxanes combined with either carboplatin or cisplatin and with etoposide or gemcitabine (Fig. 3 and Fig. 4). The median survival time (9 months), 1-year survival rate (38%), and 2-year survival rate (19%) were similar to those seen in the 132 patients reported here. Furthermore, when considering the previous 264 patients, long-term survival was seen at 3, 4, 5, 6, and 8 years (Fig. 3). A smaller number of patients were alive at these time frames free of any known progressive tumor (Fig. 4). Follow-up for more than 1 year is not usually reported in prospective studies of unknown primary carcinoma patients; therefore, the frequency of PFS for more than 1 year with other regimens cannot be assessed.
These survival statistics illustrate the problems of comparing response rates and median survival times as measures of efficacy in various studies. Median survival times have not changed very much (a few months) from those resulting from chemotherapy given more than a decade ago. However, newer chemotherapy regimens are associated with a minority of the patients surviving for several years. Survival beyond 1 year was very rarely reported in prospective studies in the past. The survival rates at 1, 2, and 3 years and beyond are essential in comparing therapies for these patients, and in making judgments concerning the efficacy of therapy.
Several other phase II studies of the newer cytotoxic drugs, including paclitaxel, gemcitabine, and irinotecan in combination with a platinum agent, have also been reported [9–14]. Those studies showed similar response rates and median survival durations as reported here. However, long-term follow-up has generally not been reported. The definitive comparison of various therapies requires phase III prospective randomized trials.
Even though the phase II trial reported here was large, there are still many subgroups of patients represented with varying clinical and pathologic features and responsiveness to cytotoxic therapy. Prognostic factors in these patients are slowly coming to light [11, 14], and it will be necessary in the future to further stratify, or separate, patients with varying prognoses. Several subsets of patients with known good prognostic features were excluded from this trial and our previous four trials. However, other subgroups of patients with different prognoses were included (i.e., those with lymph-node-dominant and liver-dominant tumors).
The overall response rate, PFS time, and survival duration resulting from the sequential approach reported here do not appear to be any better than those previously reported with other combinations of new cytotoxic drugs [9–14]. Furthermore, the overall toxicity of the sequential administration of these combinations was greater than that seen in our previous four studies as well as in other reports [9–14]. Based upon the long-term survival rates observed in the 396 patients in our five phase II trials, a very compelling argument can be made that incorporating newer cytotoxic drugs (paclitaxel, docetaxel, gemcitabine, irinotecan) into therapy results in superior survival. We are now conducting a randomized phase III prospective trial comparing paclitaxel/carboplatin/etoposide with gemcitabine/irinotecan in previously untreated patients.
We expect the treatment of patients with carcinoma of unknown primary site to be less empiric and more effective once the lineage and biologic features of these tumors are better understood. Gene-expression profiling has particular promise as a means to determine the primary site, and perhaps the biology of individual tumors, and may also provide clues to more specific targeted therapies.
|
|
ACKNOWLEDGMENT |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G. R. Varadhachary, D. Talantov, M. N. Raber, C. Meng, K. R. Hess, T. Jatkoe, R. Lenzi, D. R. Spigel, Y. Wang, F. A. Greco, et al. Molecular Profiling of Carcinoma of Unknown Primary and Correlation With Clinical Evaluation J. Clin. Oncol., September 20, 2008; 26(27): 4442 - 4448. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Talantov, J. Baden, T. Jatkoe, K. Hahn, J. Yu, Y. Rajpurohit, Y. Jiang, C. Choi, J. S. Ross, D. Atkins, et al. A Quantitative Reverse Transcriptase-Polymerase Chain Reaction Assay to Identify Metastatic Carcinoma Tissue of Origin J. Mol. Diagn., July 1, 2006; 8(3): 320 - 329. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| THE ONCOLOGIST | STEM CELLS | CME | ALPHAMED PRESS JOURNALS |
