PET/CT in cancer: moderate sample sizes may suffice to justify replacement of a regional gold standard

Mol Imaging Biol. 2009 Nov-Dec;11(6):381-5. doi: 10.1007/s11307-009-0224-1. Epub 2009 May 8.

Abstract

Purpose: For certain cancer indications, the current patient evaluation strategy is a perfect but locally restricted gold standard procedure. If positron emission tomography/computed tomography (PET/CT) can be shown to be reliable within the gold standard region and if it can be argued that PET/CT also performs well in adjacent areas, then sample sizes in accuracy studies can be reduced.

Procedures: Traditional standard power calculations for demonstrating sensitivities of both 80% and 90% are shown. The argument is then described in general terms and demonstrated by an ongoing study of metastasized prostate cancer.

Results: An added value in accuracy of PET/CT in adjacent areas can outweigh a downsized target level of accuracy in the gold standard region, justifying smaller sample sizes.

Conclusions: If PET/CT provides an accuracy benefit in adjacent regions, then sample sizes can be reduced and the conduct of trials accelerated, leading to earlier decisions on the use of PET/CT while exposing fewer patients and reducing overall costs.

MeSH terms

  • Clinical Trials as Topic
  • Humans
  • Male
  • Positron-Emission Tomography / methods*
  • Prostatic Neoplasms / diagnosis*
  • Sample Size
  • Sensitivity and Specificity
  • Tomography, Emission-Computed / methods*