RESUMO
BACKGROUND: We explored whether the effectiveness of immune checkpoint inhibitors (ICIs) can be characterized by incorporating a composite of duration of response (DOR) to complement traditional Response Evaluation Criteria in Solid Tumors (RECIST) criteria for objective response rate (ORR) in an intent-to-treat (ITT) population. Furthermore, the correlation of this novel endpoint, characterized by the restricted mean time in response (RMTR), with overall survival (OS) will be examined. METHODS: We analyzed ORR alone or in combination with DOR (RMTR) in available phase I, II, and III trials evaluating nivolumab monotherapy or in combination with ipilimumab across solid tumor types. ORR was evaluated per RECIST V.1.1. DOR was estimated using individual patient data in ITT populations regardless of RECIST response, with non-responders imputed as zero. Associations between ORR alone or RMTR and OS were evaluated in the ITT population. DOR curves were generated using the Kaplan-Meier product limit method, and 6-month RMTR, a measure of response durability, was derived from the area under the curves. For ORR and RMTR in the ITT population, the strength of association with OS was analyzed using Pearson correlation coefficients (r). RESULTS: Nivolumab treatment was associated with longer response durations than active control in responder and ITT populations. Similarly, ORR and RMTR were both significantly correlated with OS (ORR vs OS: r=0.684, p=0.02; RMTR vs OS: r=0.695, p=0.018). CONCLUSIONS: Combining ORR and DOR (RMTR) to objectively characterize tumor shrinkage in an ITT patient population is a novel approach that appears to correlate well with OS in patients treated with nivolumab monotherapy or in combination with ipilimumab. This endpoint may provide a more complete characterization of tumor shrinkage to incorporate into the design of future ICI clinical trials. However, confirmation of this approach will require further research.