Table 1

Summary characteristics of included trials

CharacteristicsTrial(s) (n=76)Patients (n=45 688)
Fund source
 Pharmaceutical7343 557
 Open label3721 455
 Single blind1185
 Double blind3824 048
 Global6239 766
 Superiority7142 213
Number of enrolled patients
Tumour type
 Breast cancer1711 132
 Renal cell carcinoma116720
 Ovarian cancer64536
 Other†2312 298
Response assessment criteria
 RECIST7142 756
 Not given1545
Primary endpoint
 Central assessed‡4326 344
 Local assessed¶2313 167
Primary endpoint outcome
 Positive5129 982
 Negative2314 600
  • *Four non-inferiority, one hybrid design combing superiority and non-inferiority.

  • †Four gastrointestinal stromal tumour, three pancreatic tumour, three sarcoma, three medullary thyroid cancer, two glioblastoma, two prostate cancer, two neuroendocrine tumour, one colorectal adenocarcinoma, one gastric cancer, one head and neck cancer and one hepatocellular carcinoma.

  • ‡Forty central-assessed PFS, two central-assessed time to progression and one central-assessed ORR.

  • §Nine overall survival and one unknown-assessed ORR.

  • ¶Twenty-three local-assessed PFS.

  • **One study used ORR as the primary endpoint, but we were unable to recognise which assessment (central or local assessment) for the ORR was considered as the primary endpoint (central-assessed ORR or local-assessed ORR?). Because a significant difference was found in central review (p=0.03) but not found in local assessment (p=0.05), we considered the outcome of the primary endpoint as indeterminate.48

  • ††Another study considered local-assessed PFS and OS as coprimary endpoints: a significant difference was found in PFS (p<0.01), but was not found in OS (p=0.10). We considered the outcome of the primary endpoint as indeterminate as well.83

  • NSCLC, non-small-cell lung cancer; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; RECIST, Response Evaluation Criteria in Solid Tumors.