Reply to A. Levy et al

Abstract

In their letter, Levy et al provide thoughtful comments with respect to some of RECIST’s limitations and highlight recent developments regarding techniques toevaluatenovel targetedcompounds.LikeLevyetal, we too celebrate adaptations of response methodologies in the evaluation of anticancer therapies. Examples include incorporation of tumor morphology by computed tomography (CT) in colorectal cancer, tumoral enhancement in arterial CT phases for hepatocellular carcinoma, metabolic response by [F]fluorodeoxyglucose positron emission tomography (PET) for GI stromal tumors, and even modifying criteria to allow for agents that may have a slower time course to action than traditional anticancerdrugs(eg, immune-relatedresponsecriteria). Additionally, as Levy et al point out, adaptations that incorporate information regarding tumor dynamics such as tumor growth rate or the nonprogression index can also add valuable information when assessing tumor response to treatment. Despite the fact that these and other modifications are becoming available, our data suggest that the traditional RECIST assessment of change in tumor size is still generally meaningful when assessing tumor response, even in the current era of targeted agents. As we reported, more than 90% of patients included in our study had treatment regimens that included only targeted agents without traditional cytotoxic drugs. Additionally, our data were generated by analyzing a wide variety of tumor types across trials of agents with diverse mechanisms of action, but modified response methodologies have often been studied in limited settings only. Although specialized tumor response evaluation methodologies can be difficult to implement in a uniform fashion across global cancer treatment sites, traditional methods of evaluating change in tumor size by CT or magnetic resonance imaging are relatively straightforward to implement and have been in use by the global community for more than 30 years. Certainly no single tumor response methodology is better than others in all cases. There are trade-offs between using a generally applicable methodology versus measures that are more specific to individual settings. RECIST’s benefits are that it is straightforward and supported by many years of global experience. Our data suggest that incorporating RECIST evaluation of change in tumor size into early-phase clinical trials of novel agents as a key end point is reasonable and can provide critical information in assessing a novel drug’s potential efficacy. As we report, this is particularly true when tumor size changes are evaluated on a continuous scale rather than by RECIST’s complete response/partial response/stable disease/progressive disease response categorizations. In many cases, RECIST data may be well supplemented by a second, concurrently employed methodology that is appropriate for the situation such as [F]fluorodeoxyglucose PET, fluoro-l-thymidine PET, dynamic contrast-enhanced magnetic resonance imaging, magnetic resonance spectroscopy, modified RECIST criteria as referenced by Levy et al, and others. We look forward to the further refinement of RECIST and other tumor response methodologies, which will ultimately help our patients with cancer.