Chemotherapy for Unknown Primary: A Phase II Trial

Abstract & Commentary

By William B. Ershler, MD, INOVA Fairfax Hospital Cancer Center, Fairfax, VA; Director, Institute for Advanced Studies in Aging, Washington, DC.

Synopsis: Optimal treatment for carcinoma of unknown primary has not been established. In a multicenter, randomized Phase II trial, 66 patients were treated with cisplatin, gemcitabine and either paclitaxel or vinorelbine. The regimens were well tolerated and the observed response rates approached 50%. Response duration and median time to progression were comparable between the two treatments, but the overall survival of 13.6 months for the vinorelbine-containing regimen compared to 9.6 months for the paclitaxel-containing regimen) and its slightly better tolerability led these investigators to favor that regimen for promotion to large scale trial.

Source: Palmeri S, et al. Cisplatin and gemcitabine with either vinorelbine or paclitaxel in the treatment of carcinomas of unknown primary site. Results of an Italian multicenter, randomized, Phase II Study. Cancer. 2006;107:2898-2905.

Despite advances in imaging and pathology techniques, slightly less than 5% of all cancers present as metastatic disease from clinically unrecognized primaries.1 This obviously heterogeneous group represents a challenge for oncologists and there currently is no established standard of care. In the current report, two somewhat different treatment regimens were compared in a randomized, multicenter trial conducted by the Unknown Primary Italian Study Group. Sixty-six previously untreated patients with carcinomas of unknown primaries (CUP) were randomized to receive Cisplatin (35 mg/m2), gemcitabine (1000 mg/m2) and either paclitaxel (70 mg/m2) (CGT) or vinorelbine (25 mg/m2) (CGV) on days 1 and 8 of 21-day cycles. Twenty-nine (44%) presented with 2 or more involved sites, and histopathology was adenocarcinoma in 73% and squamous in 11%. For all patients, the search for a primary site included routine but extensive imaging studies and endoscopy, as indicated.

For the 33 patients who received the paclitaxel-containing regimen (CGT), 16 (48.5%) experienced an objective response and 9 (27.2%) had stable disease. In comparison, for those 33 patients who received the vinorelbine-containing regimen (CGV), 14 patients (42.3%) experienced an objective response and 8 (24.2%) patients had disease stabilization. The median response duration and the median time to progression were similar in both treatment groups. The median overall survival (OS) was 9.6 months (95% confidence interval [CI], 7.11-12.09 months) for patients who received CGT and 13.6 months (95% CI, 6.61-20.59 months) for those who received CGV. Grade 3 and 4 toxicities were more frequent in the paclitaxel-containing (CGT) arm. The authors, based upon the marginal differences described in overall survival and lesser toxicity, favor the CGV regimen for future examination in Phase III studies.


It is clear that we have not reached a satisfactory regimen for the effective treatment of metastatic carcinoma of unknown primary. Several trials have been published (nicely reviewed in the current report) and most report response rates in the 25-40% range with median overall survivals of approximately 6-10 months. CGT falls right in this range, whereas CGV might be a little better, at least on the basis the observed median overall survival of 13.6 months is among the best ever reported for carcinomas of unknown primary. In general, conclusions regarding overall survival are not typically made from phase II trial inasmuch as patient selection in relatively small studies can easily exaggerate such measures. This is particularly dangerous for a heterogeneous group, such as those with carcinoma of unknown primary. Thus, despite appropriate randomization methodology, potentially having non-comparable groups renders a comparison of CGT vs CGV (or either, with published reports in the literature), risky. For example, it is known that metastatic disease involving the liver is more ominous than other sites (eg, lymph nodes or bone). 1, 2 Other factors such as sex, gender, performance status, number of sites involved, and histology are also predictive of outcome and it would not be possible, outside of a large-scale study, to provide adequate trial design to definitively establish the superiority of CGV or other regimen when compared to those others with comparable (but slightly lesser) results in phase II investigation.

That stated, the data from this trial and from the literature would support the inclusion of each of the agents in future studies. A number of published trials have focused on platinum analogues3,4 and on taxanes5,6, both of which, it would seem, improve response rates but not necessarily overall survival. Gemcitabine, because of its broad spectrum of activity and favorable toxicity profile, is a logical drug for investigation in this setting.7 Similarly, vinorelbine has demonstrable activity in patients with lung, breast, head/neck, ovarian and uterine carcinomas, and also an acceptable toxicity profile.8 With the data presented in this report, outside of a clinical trial, the CGV regimen would seem as logical a choice as any for the treatment of newly diagnosed patients with metastatic carcinoma from unknown primary.


1. Pavlidis N, et al. Cancer of unknown primary (CUP). Crit Rev Oncol Hematol. 2005;54:243-250.

2. Hess KR, et al. Classification and regression tree analysis of 1000 consecutive patients with unknown primary carcinoma. Clin Cancer Res. 1999;5:3403-3410.

3. Milliken ST, et al. Metastatic adenocarcinoma of unknown primary site. A randomized study of two combination chemotherapy regimens. Eur J Cancer Clin Oncol. 1987;23:1645-1650.

4. Eagean RT, et al. Lack of value for cisplatin added to mitomycin-doxorubicin combination chemotherapy for carcinoma of unknown primary site. A randomized trial. Am J Clin Oncol. 1987;10:82-85.

5. Greco FA, et al. Carcinoma of unknown primary site. Cancer. 2000;89:2655-2660.

6. Greco FA, et al. Carcinoma of unknown primary site: sequential treatment with paclitaxel/carboplatin/etoposide and gemcitabine/irinotecan: a Minnie Pearl Cancer Research Network phase II trial. Oncologist. 2004;9:644-652.

7. Hainsworth JD, et al. Combination chemotherapy with gemcitabine and irinotecan in patients with previously treated carcinoma of an unknown primary site: a Minnie Pearl Cancer Research Network Phase II trial. Cancer. 2005;104:1992-1997.

8. Gridelli C, Shepherd FA. Chemotherapy for elderly patients with non-small cell lung cancer: a review of the evidence. Chest. 2005;128:947-957.