Metastastic Malignancy of Unknown Primary

Background

This condition rarely occurs in patients who have a biopsy proven cancer but no known primary lesion. The oncologist must rely upon the pathologist to use their diagnostic skills to render a likely source which may guide the therapy.

OUTLINE

Epidemiology

Disease Associations

Pathogenesis

Laboratory/Radiologic/
Other Diagnostic Testing

Gross Appearance and Clinical Variants

Histopathological Features and Variants

Special Stains/
Immunohistochemistry/
Electron Microscopy

Differential Diagnosis

Prognosis

Treatment

Commonly Used Terms

Internet Links

EPIDEMIOLOGY CHARACTERIZATION

INCIDENCE
Ann Surg 1977;186:625-630

0.5-9% of all malignancies in the United States

HISTOLOGICAL TYPES CHARACTERIZATION

General Pathol Case Rev 2001;6:137-145

VARIANTS

Malignant small round cell tumors Ewing's sarcoma
PNET
Small cell synovial sarcoma
Lymphoblastic lymphoma
Rhabdomyosarcoma solid-alveolar type
Neuroblastoma
Small cell neuroendocrine carcinomas

Malignant oncocytoid tumors Low grade neuroendocrine carcinomas of the lung and gut
Medullary carcinoma of the thyroid
Exocrine and endocrine carcinomas of the pancreas
Hepatocellular carcinomas

Malignant neoplasms with granular cytoplasmic eosinophilia Melanoma
Malignant rhabdoid tumors

Malignant neoplasms with globular cytoplasmic eosinophilia Neuroendocrine carcinomas of the thyroid, GI
Hepatocellular carcinomas
Melanomas

Malignant neoplasms with diffuse cytoplasmic eosinophilia Anaplastic large cell lymphomas
Sarcomas with epithelioid features
Renal cell carcinomas
Adrenocortical carcinoma
Large cell undifferentiated carcinomas
Amelanotic oncocytoid melanomas

Malignant clear cell tumors with nested architecture Paragangliomas
Renal cell carcinoma
Gastric carcinoma
Clear cell hepatocellular carcinoma
Clear cell thyroid carcinoma, papillary and medullary

Sclerosing clear cell malignancies
Limited mainly to salivary gland tumors

Hyalinizing clear cell carcinoma of the salivary gland
Acinic cell carcinoma
Some large cell non-Hodgkin's lymphomas
Clear cell melanomas

Clear cell tumors lacking consistent architectural patterns Mucoepidermoid carcinoma
Metastatic melanoma with balloon cell changes
Seminoma/germinoma
Primary clear cell carcinomas of the prostate, lung, kidney, skin

Malignant spindle and pleomorphic tumors Sarcomatoid renal cell carcinoma
Common also in lung, urinary tract, female genital tract, GI tract and pancreas, thyroid, upper airway

SPECIAL STAINS/ IMMUNOHISTO-CHEMISTRY/OTHER CHARACTERIZATION

The Role of Desmin and N-Cadherin in Effusion Cytology A Comparative Study Using Established Markers of Mesothelial and Epithelial Cells

Ben Davidson, M.D. , Ph.D. ; S�ren Nielsen, C.T. ; Jette Christensen, C.T. ; Pia Asschenfeldt, M.D. ; Aasmund Berner, M.D. , Ph.D. ; Bj�rn Risberg, M.D. , Ph.D. ; Preben Johansen, M.D.

From the Department of Pathology (Division of Cytology) (B.D., A.B., B.R.), Norwegian Radium Hospital, Oslo, Norway; and the Department of Pathology (S.N., J.C., P.A., P.J.), Aalborg Hospital, Aalborg, Denmark.

Am J Surg Pathol 2001;25:1405-1412 Abstract quote

The objective of the present study was to analyze the role of the mesothelial markers desmin and N-cadherin in the diagnostic panel of serous effusions.

A total of 181 pleural and peritoneal effusions consisted of 101 cases cytologically diagnosed as malignant (89 carcinomas, 12 mesotheliomas), 78 benign, and 2 inconclusive specimens. All specimens were immunostained using 11 antibodies, against epithelial membrane antigen, Ber-EP4, carcinoembryonic antigen, E-cadherin, CA 125, N-cadherin, desmin, calretinin, p53, vimentin, and CD45. After evaluation of immunocytochemistry results, 110 specimens were diagnosed as malignant (98 carcinomas, 12 mesotheliomas) and 71 as benign (56 cellular, 15 paucicellular). The presence of desmin was detected in benign mesothelial cells in 47 of 56 (84%) reactive cellular specimens compared with 1 of 12 (8%) malignant mesotheliomas and 2 of 98 (2%) carcinomas. N-cadherin was expressed in 48 of 56 (86%) reactive cases, 12 of 12 (100%) mesotheliomas, and 47 of 98 (48%) carcinomas.

In carcinomas, N-cadherin expression was most often seen in ovarian carcinoma but was also found in other carcinomas. Calretinin, an established marker of mesothelial cells, was detected in 52 of 56 (93%) reactive specimens, 11 of 12 (93%) mesotheliomas, and 3 of 98 (3%) carcinomas. Evaluation of staining results led to reclassification of six malignant specimens as benign, whereas 17 cases diagnosed as benign and the two diagnosed as inconclusive were classified as malignant.

In conclusion, desmin appears to be a promising marker for the distinction between reactive mesothelium and malignant epithelial cells in terms of both specificity and sensitivity, and its complementary use with calretinin is recommended. Unlike calretinin, it may also prove valuable for the distinction between benign and malignant mesothelial cells. N-cadherin does not have a role in the distinction between mesothelial and epithelial cells. However, it may prove useful in the characterization of carcinomas of unknown origin. As has previously been shown, a significant number of diagnoses that are based on morphologic examination alone are modified after the use of a broad antibody panel.

PROGNOSIS AND TREATMENT CHARACTERIZATION

PROGNOSIS
Pathol Case Rev 2001;6:178-184

Oncologists may classify these tumors by clinicopathologic categories which helps to guide therapy and prognosis

FAVORABLE CLINICAL SUBGROUPS

Women with peritoneal carcinomatosis

Axillary nodal metastases

Men with skeletal metastases and/or increased serum PSA

Poorly differentiated adenocarcinoma and poorly differentiated carcinoma

Poorly differentiated neuroendocrine tumors

Isolated cervical or inguinal squamous lymphadenopathy

TREATMENT

Etoposide/cisplatin-based chemotherapy for patients with metastatic poorly differentiated carcinoma of unknown primary site.

Greco FA, Johnson DH, Hainsworth JD.

Department of Medicine, Vanderbilt University, Nashville, TN.

Semin Oncol 1992 Dec;19(6 Suppl 13):14-8 Abstract quote

Patients with poorly differentiated carcinoma (PDC) or poorly differentiated adenocarcinoma (PDA) of unknown primary site comprise a sizable minority (25% to 35%) of patients with carcinoma of unknown primary site. Some of these neoplasms are highly responsive to combination chemotherapy, and a minority of patients are curable.

Between 1978 and 1982 we treated 67 patients with combination chemotherapy, most of whom received PVB (cisplatin/vinblastine/bleomycin) with or without doxorubicin. Thirty-eight patients (56%) responded to treatment, with 15 (22%) attaining complete responses (CRs). Nine patients (13%) are long-term disease-free survivors. Since that time, we have incorporated etoposide into our treatment program because of its synergism with cisplatin and its marked activity against several other neoplasms including germ cell tumors. Seventeen patients with PDC or PDA of unknown primary site received second-line therapy with etoposide/cisplatin after failing to respond to PVB. Ten of these patients had partial responses, with a median response duration of 5 months (range, 2 to 12). Eighty-five previously untreated patients with PDC or PDA received etoposide/cisplatin combinations as initial treatment; 57 of 78 evaluable patients (73%) responded to therapy, and 24 (31%) achieved CRs. Sixteen patients (19% of entire group) remain disease-free a median of 28 months (range, 9 to 66) after therapy.

Etoposide is active against poorly differentiated carcinoma of unknown primary site. These results indicate that initial treatment with etoposide/cisplatin combinations is equivalent or superior to our previous results with PVB.

Multicentric phase II study of cisplatin and etoposide in patients with metastatic carcinoma of unknown primary.

Voog E, Merrouche Y, Trillet-Lenoir V, Lasset C, Peaud PY, Rebattu P, Negrier S.

Medical Oncology Department, Centre L. Berard, Lyon, France.

Am J Clin Oncol 2000 Dec;23(6):614-6 Abstract quote

The aim of this study was to determine the efficacy and toxicity of combination cisplatin and etoposide chemotherapy in patients with metastatic carcinoma of unknown primary. Patients were treated with cisplatin (100 mg/m2 iv day 1) followed by etoposide (100 mg/m2 iv days 1-3) every 3 weeks for a maximum of 6 cycles.

Patients with progressive disease after two or four courses could receive FAC (fluorouracil, doxorubicin, and cyclophosphamide) until progression. Twenty-five patients were entered and were assessable for response and toxicity. Fifteen (60%) patients had adenocarcinomas. Patients received a median of four courses. Toxicity was mainly hematologic including grade III/IV neutropenia. The overall response rate was 32%. There was no complete response, 32% partial responses, 32% stable disease, and 36% disease progression. Median response duration was 4 months (range: 2-5 months). The median overall survival of the 25 patients was 8 months. No objective response could be obtained with FAC, but 33% of patients achieved stabilization of the disease for at least 3 months.

This cisplatin-etoposide combination demonstrated some activity against an usually resistant disease.

Carcinoma of unknown primary site: phase II trials with docetaxel plus cisplatin or carboplatin.

Greco FA, Erland JB, Morrissey LH, Burris HA 3rd, Hermann RC, Steis R, Thompson D, Gray J, Hainsworth JD.

Sarah Cannon-Minnie Pearl Cancer Center and Tennessee Oncology, Nashville 37203, USA.

Ann Oncol 2000 Feb;11(2):211-5 Abstract quote

PURPOSE: To evaluate the toxicity, response rate and short-term survival associated with the chemotherapy combinations of docetaxel plus cisplatin or carboplatin when used for the treatment of patients with metastatic carcinoma of unknown primary site.

PATIENTS AND METHODS: Twenty-six patients were treated with docetaxel 75 mg/m2 i.v. and cisplatin 75 mg/m2 i.v. given every three weeks (study A) and subsequently, 47 patients were treated with docetaxel 65 mg/m2 and carboplatin (AUC dose = 6) every three weeks (study B). Stable or responding patients received a maximum of eight courses of therapy. Patients who were known to be in treatable subset groups were excluded from these trials. The majority of patients had two or more sites of metastasis; about 45% had adenocarcinoma and 50% poorly differentiated carcinoma.

RESULTS: In study A, 6 of 23 (26%) assessable patients had a major response to therapy. The median survival was eight months and one-year survival 42%. Seven patients were removed from the study early for grade 3 or 4 nausea and vomiting. In study B, 9 of 40 assessable patients (22%) had a major response to therapy. Median survival was eight months and one-year survival 29%. Toxicity associated with this regimen was predominantly myelosuppression. Comparisons of the two sequential trials showed no differences in response rates or survivals (P = 0.75).

CONCLUSIONS: Docetaxel and cisplatin (study A) is an active combination in carcinoma of unknown primary site, but associated with substantial gastrointestinal toxicity. A combination of docetaxel plus carboplatin (study B) is better tolerated and produced a similar response rate, median survival and one-year survival. Comparative phase III trials will be necessary to unequivically prove a survival advantage for any form of therapy in these patients. However, the survival for patients with carcinoma of unknown primary site receiving docetaxel-based chemotherapy is comparable to the survivals for several other groups of advanced cancer patients, such as non-small cell lung cancer, receiving various types of chemotherapy.

Combination of cisplatin-doxorubicin-cyclophosphamide in adenocarcinoma of unknown primary site: a phase II trial.

Guardiola E, Pivot X, Tchicknavorian X, Magne N, Otto J, Thyss A, Schneider M.

Centre Antoine Lacassagne, Nice, France.

Am J Clin Oncol 2001 Aug;24(4):372-5 Abstract quote

The purpose of this report is to evaluate toxicity, response, and survival of the cyclophosphamide-doxorubicin-cisplatin (CAP) chemotherapy regimen in patients with adenocarcinoma of unknown primary site (ACUP).

Twenty-two patients with ACUP were eligible for this study between June 1992 and April 1999. There were 13 men (59%) and 9 women (41%) with a median age of 53.5 years (range: 29--78 years). Lung (seven), liver (six), vertebral bone site (six), and abdominal nodes (six) were the most common metastatic sites. Treatment consisted of doxorubicin 50 mg/m(2), cyclophosphamide 1,000 mg/m(2), and cisplatin 100 mg/m(2) (CAP), administered every 3 weeks; a total of six courses were planned. Twenty-two patients were assessable for toxicity and 20 patients were assessable for response. Grade III to IV neutropenia was observed in 14 patients (64%); febrile neutropenia occurred in 6 patients (27%) and in 10 cycles (12.5%). Grade III to IV anemia and thrombocytopenia were found in 12 (54.5%) and 9 patients (41%), respectively. Grade III to IV nausea and vomiting was observed in 9 patients (41%). Ten patients, 50% of the assessable population, obtained an objective response, including 3 complete (15%) and 7 partial (35%) responses. The median response duration was 3.9 months (range: 0.5--13.3 months). One patient (5%) had stable disease and 5 patients (25%) had progressive disease. The median overall survival and the median time to progression were 10.7 months (range: 0.4--56.9 months) and 8.8 months (range: 6.6--16.5 months), respectively.

The CAP regimen in patients with ACUP had significant activity. This chemotherapy regimen induced a high level of grade III to IV toxicities and could not be considered as a treatment of reference. However, the emergence of long-term survivors among responder patients highlighted the need to search for an active treatment for patients with ACUP.

Pathol Case Rev 2001;6(4).

Henry JB. Clinical Diagnosis and Management by Laboratory Methods. Twentieth Edition. WB Saunders. 2001.
Rosai J. Ackerman's Surgical Pathology. Ninth Edition. Mosby 2004.
Sternberg S. Diagnostic Surgical Pathology. Fourth Edition. Lipincott Williams and Wilkins 2004.
Robbins Pathologic Basis of Disease. Seventh Edition. WB Saunders 2005.
DeMay RM. The Art and Science of Cytopathology. Volume 1 and 2. ASCP Press. 1996.
Weedon D. Weedon's Skin Pathology Second Edition. Churchill Livingstone. 2002
Fitzpatrick's Dermatology in General Medicine. 5th Edition. McGraw-Hill. 1999.
Weiss SW and Goldblum JR. Enzinger and Weiss's Soft Tissue Tumors. Fourth Edition. Mosby 2001.

Commonly Used Terms

Basic Principles of Disease
Learn the basic disease classifications of cancers, infections, and inflammation

Commonly Used Terms
This is a glossary of terms often found in a pathology report.

Diagnostic Process
Learn how a pathologist makes a diagnosis using a microscope

Surgical Pathology Report
Examine an actual biopsy report to understand what each section means

Special Stains
Understand the tools the pathologist utilizes to aid in the diagnosis

How Accurate is My Report?
Pathologists actively oversee every area of the laboratory to ensure your report is accurate

Got Path?
Recent teaching cases and lectures presented in conferences

Internet Links

Pathologists Who Make A Difference
Search for a Physician Specialist

Last Updated July 15, 2005

Send mail to The Doctor's Doctor with questions or comments about this web site.
Read the Medical Disclaimer.

Copyright © The Doctor's Doctor

Epidemiology
Disease Associations
Pathogenesis
Laboratory/Radiologic/ Other Diagnostic Testing
Gross Appearance and Clinical Variants
Histopathological Features and Variants
Special Stains/ Immunohistochemistry/ Electron Microscopy
Differential Diagnosis
Prognosis
Treatment
Commonly Used Terms
Internet Links

HISTOLOGICAL TYPES	CHARACTERIZATION
General	Pathol Case Rev 2001;6:137-145
VARIANTS
Malignant small round cell tumors	Ewing's sarcoma PNET Small cell synovial sarcoma Lymphoblastic lymphoma Rhabdomyosarcoma solid-alveolar type Neuroblastoma Small cell neuroendocrine carcinomas
Malignant oncocytoid tumors	Low grade neuroendocrine carcinomas of the lung and gut Medullary carcinoma of the thyroid Exocrine and endocrine carcinomas of the pancreas Hepatocellular carcinomas
Malignant neoplasms with granular cytoplasmic eosinophilia	Melanoma Malignant rhabdoid tumors
Malignant neoplasms with globular cytoplasmic eosinophilia	Neuroendocrine carcinomas of the thyroid, GI Hepatocellular carcinomas Melanomas
Malignant neoplasms with diffuse cytoplasmic eosinophilia	Anaplastic large cell lymphomas Sarcomas with epithelioid features Renal cell carcinomas Adrenocortical carcinoma Large cell undifferentiated carcinomas Amelanotic oncocytoid melanomas
Malignant clear cell tumors with nested architecture	Paragangliomas Renal cell carcinoma Gastric carcinoma Clear cell hepatocellular carcinoma Clear cell thyroid carcinoma, papillary and medullary
Sclerosing clear cell malignancies	Limited mainly to salivary gland tumors Hyalinizing clear cell carcinoma of the salivary gland Acinic cell carcinoma Some large cell non-Hodgkin's lymphomas Clear cell melanomas
Clear cell tumors lacking consistent architectural patterns	Mucoepidermoid carcinoma Metastatic melanoma with balloon cell changes Seminoma/germinoma Primary clear cell carcinomas of the prostate, lung, kidney, skin
Malignant spindle and pleomorphic tumors	Sarcomatoid renal cell carcinoma Common also in lung, urinary tract, female genital tract, GI tract and pancreas, thyroid, upper airway

PROGNOSIS AND TREATMENT	CHARACTERIZATION
PROGNOSIS	Pathol Case Rev 2001;6:178-184 Oncologists may classify these tumors by clinicopathologic categories which helps to guide therapy and prognosis
FAVORABLE CLINICAL SUBGROUPS
Women with peritoneal carcinomatosis
Axillary nodal metastases
Men with skeletal metastases and/or increased serum PSA
Poorly differentiated adenocarcinoma and poorly differentiated carcinoma
Poorly differentiated neuroendocrine tumors
Isolated cervical or inguinal squamous lymphadenopathy
TREATMENT
Etoposide/cisplatin-based chemotherapy for patients with metastatic poorly differentiated carcinoma of unknown primary site. Greco FA, Johnson DH, Hainsworth JD. Department of Medicine, Vanderbilt University, Nashville, TN.	Semin Oncol 1992 Dec;19(6 Suppl 13):14-8 Abstract quote Patients with poorly differentiated carcinoma (PDC) or poorly differentiated adenocarcinoma (PDA) of unknown primary site comprise a sizable minority (25% to 35%) of patients with carcinoma of unknown primary site. Some of these neoplasms are highly responsive to combination chemotherapy, and a minority of patients are curable. Between 1978 and 1982 we treated 67 patients with combination chemotherapy, most of whom received PVB (cisplatin/vinblastine/bleomycin) with or without doxorubicin. Thirty-eight patients (56%) responded to treatment, with 15 (22%) attaining complete responses (CRs). Nine patients (13%) are long-term disease-free survivors. Since that time, we have incorporated etoposide into our treatment program because of its synergism with cisplatin and its marked activity against several other neoplasms including germ cell tumors. Seventeen patients with PDC or PDA of unknown primary site received second-line therapy with etoposide/cisplatin after failing to respond to PVB. Ten of these patients had partial responses, with a median response duration of 5 months (range, 2 to 12). Eighty-five previously untreated patients with PDC or PDA received etoposide/cisplatin combinations as initial treatment; 57 of 78 evaluable patients (73%) responded to therapy, and 24 (31%) achieved CRs. Sixteen patients (19% of entire group) remain disease-free a median of 28 months (range, 9 to 66) after therapy. Etoposide is active against poorly differentiated carcinoma of unknown primary site. These results indicate that initial treatment with etoposide/cisplatin combinations is equivalent or superior to our previous results with PVB.
Multicentric phase II study of cisplatin and etoposide in patients with metastatic carcinoma of unknown primary. Voog E, Merrouche Y, Trillet-Lenoir V, Lasset C, Peaud PY, Rebattu P, Negrier S. Medical Oncology Department, Centre L. Berard, Lyon, France.	Am J Clin Oncol 2000 Dec;23(6):614-6 Abstract quote The aim of this study was to determine the efficacy and toxicity of combination cisplatin and etoposide chemotherapy in patients with metastatic carcinoma of unknown primary. Patients were treated with cisplatin (100 mg/m2 iv day 1) followed by etoposide (100 mg/m2 iv days 1-3) every 3 weeks for a maximum of 6 cycles. Patients with progressive disease after two or four courses could receive FAC (fluorouracil, doxorubicin, and cyclophosphamide) until progression. Twenty-five patients were entered and were assessable for response and toxicity. Fifteen (60%) patients had adenocarcinomas. Patients received a median of four courses. Toxicity was mainly hematologic including grade III/IV neutropenia. The overall response rate was 32%. There was no complete response, 32% partial responses, 32% stable disease, and 36% disease progression. Median response duration was 4 months (range: 2-5 months). The median overall survival of the 25 patients was 8 months. No objective response could be obtained with FAC, but 33% of patients achieved stabilization of the disease for at least 3 months. This cisplatin-etoposide combination demonstrated some activity against an usually resistant disease.
Carcinoma of unknown primary site: phase II trials with docetaxel plus cisplatin or carboplatin. Greco FA, Erland JB, Morrissey LH, Burris HA 3rd, Hermann RC, Steis R, Thompson D, Gray J, Hainsworth JD. Sarah Cannon-Minnie Pearl Cancer Center and Tennessee Oncology, Nashville 37203, USA.	Ann Oncol 2000 Feb;11(2):211-5 Abstract quote PURPOSE: To evaluate the toxicity, response rate and short-term survival associated with the chemotherapy combinations of docetaxel plus cisplatin or carboplatin when used for the treatment of patients with metastatic carcinoma of unknown primary site. PATIENTS AND METHODS: Twenty-six patients were treated with docetaxel 75 mg/m2 i.v. and cisplatin 75 mg/m2 i.v. given every three weeks (study A) and subsequently, 47 patients were treated with docetaxel 65 mg/m2 and carboplatin (AUC dose = 6) every three weeks (study B). Stable or responding patients received a maximum of eight courses of therapy. Patients who were known to be in treatable subset groups were excluded from these trials. The majority of patients had two or more sites of metastasis; about 45% had adenocarcinoma and 50% poorly differentiated carcinoma. RESULTS: In study A, 6 of 23 (26%) assessable patients had a major response to therapy. The median survival was eight months and one-year survival 42%. Seven patients were removed from the study early for grade 3 or 4 nausea and vomiting. In study B, 9 of 40 assessable patients (22%) had a major response to therapy. Median survival was eight months and one-year survival 29%. Toxicity associated with this regimen was predominantly myelosuppression. Comparisons of the two sequential trials showed no differences in response rates or survivals (P = 0.75). CONCLUSIONS: Docetaxel and cisplatin (study A) is an active combination in carcinoma of unknown primary site, but associated with substantial gastrointestinal toxicity. A combination of docetaxel plus carboplatin (study B) is better tolerated and produced a similar response rate, median survival and one-year survival. Comparative phase III trials will be necessary to unequivically prove a survival advantage for any form of therapy in these patients. However, the survival for patients with carcinoma of unknown primary site receiving docetaxel-based chemotherapy is comparable to the survivals for several other groups of advanced cancer patients, such as non-small cell lung cancer, receiving various types of chemotherapy.
Combination of cisplatin-doxorubicin-cyclophosphamide in adenocarcinoma of unknown primary site: a phase II trial. Guardiola E, Pivot X, Tchicknavorian X, Magne N, Otto J, Thyss A, Schneider M. Centre Antoine Lacassagne, Nice, France.	Am J Clin Oncol 2001 Aug;24(4):372-5 Abstract quote The purpose of this report is to evaluate toxicity, response, and survival of the cyclophosphamide-doxorubicin-cisplatin (CAP) chemotherapy regimen in patients with adenocarcinoma of unknown primary site (ACUP). Twenty-two patients with ACUP were eligible for this study between June 1992 and April 1999. There were 13 men (59%) and 9 women (41%) with a median age of 53.5 years (range: 29--78 years). Lung (seven), liver (six), vertebral bone site (six), and abdominal nodes (six) were the most common metastatic sites. Treatment consisted of doxorubicin 50 mg/m(2), cyclophosphamide 1,000 mg/m(2), and cisplatin 100 mg/m(2) (CAP), administered every 3 weeks; a total of six courses were planned. Twenty-two patients were assessable for toxicity and 20 patients were assessable for response. Grade III to IV neutropenia was observed in 14 patients (64%); febrile neutropenia occurred in 6 patients (27%) and in 10 cycles (12.5%). Grade III to IV anemia and thrombocytopenia were found in 12 (54.5%) and 9 patients (41%), respectively. Grade III to IV nausea and vomiting was observed in 9 patients (41%). Ten patients, 50% of the assessable population, obtained an objective response, including 3 complete (15%) and 7 partial (35%) responses. The median response duration was 3.9 months (range: 0.5--13.3 months). One patient (5%) had stable disease and 5 patients (25%) had progressive disease. The median overall survival and the median time to progression were 10.7 months (range: 0.4--56.9 months) and 8.8 months (range: 6.6--16.5 months), respectively. The CAP regimen in patients with ACUP had significant activity. This chemotherapy regimen induced a high level of grade III to IV toxicities and could not be considered as a treatment of reference. However, the emergence of long-term survivors among responder patients highlighted the need to search for an active treatment for patients with ACUP.