Additivity predicts efficacy of therapies for advanced cancer

A new study recently published on Nature Cancer, by applying a model for clinical drug additivity to progression free survival results from all combination therapy trials in advanced cancers that led to US FDA approval between 1995 and 2020, showed that most clinically effective drug combinations owe their success to having effective ingredients, rather than to being ‘more than additive’.

As authors say, «The most effective known treatments for many types of cancer involve combination therapy. Because of the vast number of possible combinations, prioritizing drug combinations that are most likely to succeed in the clinic is a critical need». In this study, they propose a model of drug additivity for progression-free survival to assess whether clinical efficacies of approved drug combinations are additive or synergistic: it’s not a semantic difference because it concerns whether a mechanism of positive drug interaction is needed to develop clinically effective combination therapies. The proposed model of additivity was applied to progression-free survival results from all combination therapy trials in advanced cancers that led to US FDA approval between 1995 and 2020, for which matched combination and monotherapy data were available; results show that it accurately matches the clinical efficacy of most approved drug combinations, predicting the success of every positive trial analyzed and the failure of most negative trials. Only 5% of combinations were significantly more than additive: the accuracy of the additivity model and scarcity of synergy suggest that tumor heterogeneity, not drug–drug interaction, is the major source of benefit of approved combinations of cancer therapies. Tumor heterogeneity may explain the seeming inconsistency between preclinical synergies and clinical additivity: synergy may arise at certain concentrations in a fraction of patients, as it does in cell cultures, without noticeably affecting survival in populations. As authors conclude, «The consistency between 25 years of practice-changing trial results and the additivity model suggests that it could be useful for the prospective design of phase 3 trials of combination therapies. By estimating survival distributions, the additivity model can predict the likelihood of success of new drug combinations in different cancer types and inform trial designs and statistical analyses. Thus, the model of drug additivity has the potential to improve the rate of success of phase 3 trials and accelerate progress in cancer treatment».