ABSTRACT
Published and ongoing randomized trials in patients with carcinoma of an unknown primary site (CUP) were identified and criticized. To our knowledge, no randomized trial comparing chemotherapy versus best supportive care has been conducted in CUP to date. Seven randomized clinical trials have been published, including two randomized phase II trials. The shortcoming in all of these trials is the limited number of recruited patients, which signifies that the trials were inadequately powered to detect a small survival difference. Of the three trials that compared a cisplatin-containing regimen to another chemotherapy regimen, one reported a benefit both in terms of response and overall survival. The second trial showed a nonstatistically significant trend in overall survival in favor of the cisplatin arm. In the third trial, no difference was found, although this might be related to the low dose intensity and the short duration of cisplatin administration. Anthracyclinecontaining regimens apparently increased response rates in two randomized trials although the difference might also be related to another drug, given
the multiplicity of agents in the different chemotherapy regimens. On the basis of the results of one randomized trial, closed early due to poor accrual, it is unlikely that mitomycin-c increases survival in CUP patients. Lastly, a phase II randomized trial showed promising activity with both the cisplatin/ gemcitabine and the cisplatin/irinotecan regimens, with different patterns of toxicity, and this was the rationale for an ongoing phase III trial.
