Is pCR a Reliable Survival Surrogate in Breast Cancer Trials?
-
Pathologic complete response (pCR) is not a reliable surrogate endpoint in trials evaluating neoadjuvant therapies for early breast cancer, according to a critical analysis of the literature. A team of researchers pointed to their 2021 meta-analysis in BMJ, which found that the association between pCR and overall survival was weak — with a coefficient of determination, or R2, of 0.08. In other words, only 8% of the variability among treatment effects on overall survival could be explained by pCR — far from the optimal R2 cutoff value of 0.70 for a candidate surrogate endpoint. This weak pCR surrogacy at the randomized controlled trial level "limits the possibility to use pCR as a surrogate endpoint to predict long-term outcomes of the patient populations enrolled in [these] trials," Fabio Conforti, MD, with the European Institute of Oncology, Milan, Italy, and colleagues wrote in JAMA Oncology.
-
@madhumita This decision may be partly based on the strong correlation observed between pCR and overall survival at the patient level; however, as Conforti and colleagues explained, this correlation does not exist at the trial level. In other words, patients who achieve pCR have significantly better long-term survival compared with those who do not, but this benefit is not observed among patients enrolled in trials, which have consistently demonstrated a poor association between the degree of improvement in pCR rate and survival. Further data from the team's 2021 meta-analysis revealed that the association between pCR and disease-free survival was "similarly weak."