Is a comparative clinical trial for breast cancer tumor markers to monitor disease recurrence warranted? A value of information analysis.

BACKGROUND: Breast cancer tumor markers are used by some clinicians to screen for disease recurrence risk. Since there is limited evidence of benefit, additional research may be warranted. AIM: To assess the potential value of a randomized clinical trial of breast tumor marker testing in routine follow-up of high-risk, stage II-III breast cancer survivors. MATERIALS &
METHODS: We developed a decision-analytic model of tumor marker testing plus standard surveillance every 3-6 months for 5 years. The expected value of sample information was calculated using probabilistic simulations and was a function of: the probability of selecting the optimal monitoring strategy with current versus future information; the impact of choosing the nonoptimal strategy; and the size of the population affected.
RESULTS: The value of information for a randomized clinical trial involving 9000 women was US$214 million compared with a cost of US$30-60 million to conduct such a trial. The probability of making an alternate, nonoptimal decision and choosing testing versus no testing was 32% with current versus future information from the trial. The impact of a nonoptimal decision was US$2150 and size of population impacted over 10 years was 308,000. The value of improved information on overall survival was US$105 million, quality of life US$37 million and test performance US$71 million.
CONCLUSION: Conducting a randomized clinical trial of breast cancer tumor markers appears to offer a good societal return on investment. Retrospective analyses to assess test performance and evaluation of patient quality of life using tumor markers may also offer valuable areas of research. However, alternative investments may offer even better returns in investments and, as such, the trial concept deserves further study as part of an overall research-portfolio evaluation.

RCT Entities:   Population Interventions Outcomes