In our @JCO_ASCO piece out today, @VPplenarysesh and I argue that recent drug approvals based on metastasis-free survival (MFS) in #ProstateCancer illustrate the trade-off between speed and certainty in cancer drug approvals. Long thread to follow... ascopubs.org/doi/pdf/10.120… 
Good treatments have dramatically improved life expectancy for patients with advanced #ProstateCancer. This is a good thing. But as life expectancy has increased, overall survival becomes a difficult endpoint to achieve in clinical trials, requiring lots of patients, time, & $$$
Hence the need for a surrogate endpoint like MFS. Two drugs - apalutamide and enzalutamide - just received @US_FDA approval for non-metastatic castrate-resistant #ProstateCancer (nmCRPC) based on MFS improvement. MFS will likely be a std endpoint in future prostate cancer trials
But there are problems with this…
1) MFS was validated in a pre-modern era of advanced prostate cancer. Since the trials used to validate MFS's surrogacy, >5 drugs (e.g. abiraterone, enzalutamide, Provenge...) were approved, greatly improving life expectancy. Thus, the surrogate relationship may have changed.
2) MFS does not capture important prognostic variables (PSA kinetics, site of metastasis, visceral mets) that could be stronger substitutes of OS
3) MFS does not capture important prognostic variables (PSA kinetics, site of metastasis, visceral mets) that could be stronger surrogates of OS
4) Patients may not care about an asymptomatic metastasis compared to more important patient-centered outcomes, e.g. symptomatic metastasis, time to chemotherapy. Unclear what the effect of earlier treatment is on these important outcomes
5) Most importantly, patients with non-metastatic #ProstateCancer will probably live years. As treatments for prostate cancer move earlier in the disease course based on an MFS benefit, we need to pay attention to long-term side effects of these therapies, which are significant
This will be an issue for other advanced cancers (e.g. lung, melanoma), where life expectancy has also dramatically increased, and treatments are moving earlier in the disease course. Surrogate endpoints have and will be used to justify approvals for these earlier treatments.
But surrogate endpoints are surrogates for a reason – because overall survival and quality of life are what we strive for in cancer treatment. We ought to make sure surrogate endpoints like MFS are speeding efficacious drugs to market, and not justifying mediocre/costly drugs
While this was my first time writing about drug approval policy, @VPplenarysesh has been on the forefront highlighting similar issues with surrogate endpoints in cancer...it was really enlightening to work with him on this.
