ReviewPhase-II trials in osteosarcoma recurrences: A systematic review of past experience

- A systematic review of phase-II clinical trials evaluating new therapies for osteosarcoma relapses in the last 13 years.
- Significant heterogeneity in phase-II trial definitions of response, end-point, evaluation time point, statistical design.
- Disappointing results obtained from past Phase-II trials (5% CR + PR).
- No stable statistical hypothesis can be derived from the past experience.
- Urgent need to develop international randomised phase-II trials across all age ranges with homogenous primary end-points.

BackgroundThe most appropriate design of Phase-II trials evaluating new therapies in osteosarcoma remains poorly defined.ObjectiveTo study consistency in phase-II clinical trials evaluating new therapies for osteosarcoma recurrences with respect to eligibility criteria, response assessment, end-points, statistical design and reported results.MethodsSystematic review of clinical trials registered on clinicaltrials.gov, clinicaltrialsregister.eu and French National Cancer Institute website or referenced in PubMed and American Society of Clinical Oncology websites, between 2003 and 2016, using the following criteria: (osteosarcoma OR bone sarcoma) AND (Phase-II).ResultsAmong the 99 trials identified, 80 were Phase-II, 17 I/II and 2 II/III, evaluating mostly targeted therapy (n = 40), and chemotherapy alone (n = 26). Results were fully (n = 28) or partially (abstract, n = 6) published. Twenty-four trials were dedicated to osteosarcoma, 22 had an osteosarcoma stratum. Twenty-eight out of 99 trials refer to the age range observed at recurrence (28%). Overall, 65 trials were run in multicentre settings, including 17 international trials. Only 9 trials were randomised. The primary end-point was tumour response in 71 trials (response rate, n = 40 or best response, n = 31), with various definitions (complete + partial ± minor response and stable disease), mainly evaluated with RECIST criteria (n = 69); it was progression-free survival in 24 trials and OS in 3. In single-arm trials evaluating response rate, the null hypothesis tested (when available, n = 12) varied from 5% to 25%.ConclusionNo robust historical data can currently be derived from past efficacy Phase-II trials. There is an urgent need to develop international randomised Phase-II trials across all age ranges with standardised primary end-point.