Comparison of a gene expression profiling strategy to standard clinical work-up for determination of tumour origin in cancer of unknown primary (CUP)

J Chemother. 2013 Aug;25(4):239-46. doi: 10.1179/1973947813Y.0000000085.

Abstract

CupPrint® is a genomic signature able to identify 47 different cancer types. The aim of our study was to compare the accuracy of this genomic signature to that of a full clinical work-up in diagnosing the primary tumour site. Patients with newly diagnosed, untreated metastatic tumours were eligible for this trial. The clinical work-up and gene expression profiling on a biopsy from a metastatic site were started at the same time. The study was planned using a one-stage Fleming design. Patients in whom no primary site was diagnosed by the clinical work-up were excluded. Out of the 67 patients registered, the primary site was identified by clinical work-up in 36 patients, and diagnosis with CupPrint was obtained in 53. There were 31 evaluable patients with both clinical and CupPrint diagnoses, and out of these a similar diagnosis was obtained in 11 patients, i.e. the concordance rate was 35% (95% confidence interval: 19-55%). The median time to diagnosis through the clinical work-up was 48 days, and 10 days with CupPrint (P<0·001). We concluded that in patients with newly diagnosed metastatic tumours, CupPrint has low accuracy in diagnosing the primary cancer site.

Publication types

  • Comparative Study
  • Evaluation Study
  • Multicenter Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Female
  • Gene Expression Profiling / methods*
  • Humans
  • Intention to Treat Analysis
  • Male
  • Middle Aged
  • Neoplasms, Unknown Primary / genetics
  • Neoplasms, Unknown Primary / pathology*
  • Prospective Studies
  • Sensitivity and Specificity