• Optimal Two-Stage Phase II Design

  • Biomarker Targeted Randomized Design
  • Targeted design randomizes only marker positive patients to treatment or control arm. Untargeted design does not measure marker and randomizes all who otherwise are eligible.


  • Biomarker Stratified Randomized Design
  • Stratified design randomizes both marker positive and negative patients.


  • Sample Size Planning for Developing Classifiers Using High Dimensional Data

  • © NIH, 2007