Response to Comment on “Population Pharmacokinetics of Valemetostat and Exposure–Response Analyses of Efficacy and Safety in Patients with Relapsed/Refractory Peripheral T‐Cell Lymphoma”

Dear Dr. Bhise and colleagues, We sincerely appreciate your thoughtful and valuable comments on our manuscript. We acknowledge the limitations of this study, and several of the concerns you pointed out have been thoroughly discussed in the manuscript. In particular, the evaluation of the exposure–efficacy relationship based on a single-dose level restricts the statistical power to detect significant exposure–response (ER) relationships. Therefore, we think it is difficult to draw definitive conclusions about efficacy from this analysis. However, we conducted an evaluation of the ER relationship using a full modeling approach incorporating spike and slab priors, which helps reduce confounding risk by appropriately accounting for covariate effects. We consider that we have made efforts to detect the ER relationship more accurately, even when the exposure range is narrow. Additionally, given the observed overall response rate (ORR) of 44% and the median duration of response of 11.9 months, a 200 mg dose once daily (QD) of valemetostat provides a meaningful therapeutic benefit to patients with peripheral T-cell lymphoma (PTCL) compared with existing treatments.1 We also provide the dosing recommendations for patients with hepatic impairment or those taking a cytochrome P450 3A (CYP3A) and/or P-glycoprotein (P-gp) modulator based on the results from dedicated clinical pharmacology studies.2-5 Regarding exposure–safety analyses, we used the valemetostat unbound average concentration up to event (Cavgtte) as the primary exposure metric. While this approach allowed us to account for dose modifications during the study period, we acknowledge that it might introduce a bias. To address this concern, we also conducted sensitivity analyses using the unbound average concentration during Cycle 1 (CavgC1), which assumes no dose modifications. This metric is not influenced by dose modifications and is independent of the timing of the event. We confirmed that the conclusions derived from the CavgC1-based models were consistent with those obtained using Cavgtte. For the handling of covariates, we fully acknowledge the importance of carefully considering the factors you highlighted. In our analyses, we employed a full covariate modeling approach that included both main effects and interaction effects. Notably, in addition to the statistically significant main effects of hepatic function (according to the National Cancer Institute–Organ Dysfunction Working Group [NCI–ODWG]) and baseline level of lactate dehydrogenase (LDH) on Grade ≥3 treatment-emergent adverse events (TEAEs), baseline LDH also demonstrated a statistically significant interaction with thrombocytopenia risk, indicating a steeper ER relationship in patients with elevated LDH levels. While these patients may require careful attention, simulations conducted using resampled data from the analysis population showed that no subpopulation exhibited an obviously different TEAE profile (Figure S9 of the original article6). As more data become available in the future, we will continue assessing the toxicity risk in these patient populations. Finally, due to data limitations, we were only able to evaluate the ER relationship for efficacy within a narrow exposure range, and stratified analyses or external validation were not performed. However, we would like to emphasize that the dose for PTCL was determined based on the totality of evidence. Our dose selection has not been based solely on this ER analyses; from the early stages of development, biomarkers have been used to guide dose selection.1 Furthermore, favorable efficacy and safety have already been confirmed in patients with adult T-cell leukemia/lymphoma (ATLL), which has received approval.7 In addition, the VALENTINE-PTCL01 study demonstrated high efficacy in patients with PTCL, and we believe that 200 mg provides meaningful therapeutic benefit.1 Valemetostat 200 mg has now been approved for multiple hematological malignancies in Japan, and real-world data continue to accumulate. Further analyses may be warranted in the future to optimize the benefit–risk balance of valemetostat, if needed, and we remain committed to ensuring optimal therapeutic outcomes for patients treated with valemetostat. Once again, thank you for your constructive and thoughtful review. We highly value this dialogue and remain committed to advancing knowledge for the benefit of patients. Sincerely, Hiroyuki Inoue and co-authors HI, MT, YL, and YY are employed by and may hold stock in Daiichi Sankyo. YC was employed by Daiichi Sankyo at the time of the study. RG is an employee of the Metrum Research Group. BR and XW were employees of the Metrum Research Group at the time of the study. Work performed by the Metrum Research Group for this study was paid by Daiichi Sankyo. The data that support the findings of this study are openly available in The Journal of Clinical Pharmacology at https://accp1.onlinelibrary.wiley.com/doi/10.1002/jcph.70100, reference number 10.1002/jcph.70100.