Exposure-safety Markov modeling of ocular adverse events in patient populations treated with tisotumab vedotin
Yan Summer Feng (1), Rudy Gunawan (2), Chaitali Passey (3), Jenna Voellinger (2), Daniel Polhamus (4), Arnout Gerritsen (5), Christine O’Day (2), Leonardo Nicacio (2), Ibrahima Soumaoro (3), Manish Gupta (1), William D Hanley (2)
(1) Genmab, Princeton, NJ, USA, (2) Seagen, Inc., Bothell, WA, USA, (3) Genmab, Plainsboro, NJ, USA, (4) Metrum Research Group, Tariffville, CT, USA, (5) Genmab, Utrecht, The Netherlands
Objectives: Tisotumab vedotin (TV), a tissue factor-directed antibody-drug conjugate (ADC), is approved in the United States at a dose of 2.0 mg/kg every 21 days (Q3W) for adult patients with recurrent or metastatic cervical cancer who have progressed on or after chemotherapy. This regimen demonstrated clinical efficacy and there are ongoing efforts to optimize the dose regimen of TV in other solid tumors. Previous modeling work suggests that a more frequent administration can achieve higher dose intensity and may enhance clinical activity [1]. The current analysis aimed to characterize exposure-response relationships of TV exposure on the incidence, severity, and longitudinal time course of grade ≥2 ocular adverse events (AEs).
Methods: This retrospective analysis was conducted with pooled data from 757 patients with advanced solid tumors who participated in 7 clinical studies. The evaluated TV monotherapy or combination (TV+carboplatin, TV+bevacizumab, and TV+pembrolizumab [Pembro]) dosing included 3 regimens of TV: day 1 of a 21-day cycle; days 1, 8, and 15 of a 28-day cycle; and days 1 and 15 of a 28-day cycle (Q2W). Model-based simulations were conducted to assess the safety profile of TV at alternative dosing regimens. The relationships between exposure and incidence, severity, and time course of grade ≥2 ocular AEs were characterized by a discrete-time Markov model (DTMM)[3]. The exposure impact of ADC and microtubule‐disrupting agent monomethyl auristatin E (MMAE), as well as other covariates (eg, co-administration of Pembro and baseline dry eye) on transition probabilities between grade ≤1 and ≥2 ocular AEs were evaluated. Various combinations of the exposure models (eg, direct and effect compartment [Ce]), transformations (eg, linear and hyperbolic), and functional inclusions of interindividual variability (IIV) (i.e., on transition probabilities, EC50, and the rate constant for the effect compartment (ke)) were also assessed during model development. The concentration-time profile of TV was predicted using individual pharmacokinetic (PK) parameters obtained from a population PK analysis [2]. Model-predicted and observed time profiles of ocular AEs and transition probabilities were compared to evaluate the DTMM model. The final DTMM model was used to simulate ocular AEs with an alternate TV dosing regimen that is currently under clinical evaluation.
Results: The exposure-ocular model was a longitudinal logistic regression model with a first-order Markov element and interindividual variability on baseline probabilities. The model (Model 1) with ADC in a hyperbolic function of Ce provided the best description of data (AIC: 5650). The effect of MMAE on risk of grade ≥2 ocular AEs was not significant, and its inclusion (as a direct, linear effect) led to marginal improvement in the fit (AIC: 5646, Model 2). Covariates were added to Model 1 as proportional factors on the effect of Ce (Model 3). The risk of grade ≥2 ocular AEs increased with ADC Ce and was associated with higher risk in those with baseline dry eye conditions relative to those without dry eye conditions, but lower in patients receiving prophylactic treatment. Moreover, the 90% CI of the associated effect comparing TV monotherapy (2.0 mg/kg Q3W) and in combination with Pembro included 1 (an effect of effect of 1.08 (90% CI: 0.931, 1.23)), indicating similar risk of ocular AEs between the two treatments.
Conclusions: This Markov model offers an alternative to existing exposure-response approaches to explore the occurrence, severity, and duration of ocular AEs. Model evaluation results showed that the observed time course of ocular AEs as well as transition probabilities were reasonably described by the DTMM. Based on currently available data, the DTMM suggested that there is an association between ADC exposure and risk of grade ≥2 ocular AEs. More frequent dosing may still provide a more favorable benefit-risk profile. The 1.7 mg/kg Q2W dose is being evaluated clinically in the ongoing innovaTV 207 clinical trial.
References:
[1] Voellinger J, et al. Tisotumab vedotin dose schedule optimization in non-cervical populations. ESMO Targeted Anticancer Therapies, Paris, Mar 6-8, 2023.
[2] Gibiansky L, et.al. Population pharmacokinetic analysis for tisotumab vedotin in patients with locally advanced and/or metastatic solid tumors. CPT Pharmacometrics Syst Pharmacol. 2022;11(10):1358-1370.
[3] Karlsson, M. O., R. C. Schoemaker, B. Kemp, A. F. Cohen, J. M. van Gerven, B. Tuk, C. C. Peck, and M. Danhof. 2000. “A Pharmacodynamic Markov Mixed-Effects Model for the Effect of Temazepam on Sleep.” Clinical Pharmacology and Therapeutics 68 (2): 175–88.