Tong Lu (1), Russ Wada (2), Jeannie Hou (1), Mukta Tripathi (1), Ivor Caro (1), Mark Dresser (1), Richard A. Graham (1), Jin Y. Jin (1)
(1) Genentech, Inc.; (2) Quantitative Solutions, Inc.
Objectives: To determine an optimal duration of treatment interruption of vismodegib (Vismo) to minimize adverse events (AE) in patients (pts) with operable basal cell carcinoma (oBCC) using longitudinal ordered categorical models of dysgeusia/ageusia and muscle spasms, which are common AEs leading to treatment interruption/discontinuation (Tx I/D).
Methods: AE data were obtained from a phase 2 trial designed to evaluate the efficacy and safety of short-term Vismo treatment in pts with nodular oBCC. In total, 74 pts were randomized to one of three cohorts receiving 150mg QD Vismo: 12 weeks (wk) dosing with 30 days (d) follow up, 12wk dosing with 24wk observation and 30d follow up, or 8wk dosing with 4wk observation followed by 8wk dosing and 30d follow up. Individual PK profiles for unbound Vismo were predicted based on a previous population PK model incorporating individual covariates and used as driving force in the following PK/AE analyses in NONMEM (Laplacian method). AE data used in the model included severity assessment by CTCAE version 4.0 of dysgeusia, ageusia and muscle spasm. Preferred terms of dysgeusia and ageusia were grouped as they represent the medical concept of taste disturbance.
Results: AE grades in pts experiencing at least one AE were best described by a mixed-effects longitudinal logistic regression model [1] with a linear AE turn-over rate (ke0) and a linear drug effect. Saturable drug effect was also tested with no significant improvement on model fitting. The half-life of the AE turn-over rate was around 20 days for muscle spasm, and 35 days for dysgeusia/ ageusia. Vismo showed higher potency for dysgeusia/ageusia, with a potency ratio (the ratio of linear drug effect “slope”) of 1.57 relative to muscle spasm. Simulations showed that after 12wk of 150mg QD treatment, a Tx I/D of 6wk could lead to a complete resolution of muscle spasms in 80% of pts, and a Tx I/D of 4wk could lead to a significant improvement of muscle spasms without ≥Grade 2 in 95% of pts. For dysgeusia/ageusia, Tx I/D of 12wk could lead to a complete resolution of AE in 80% of pts, and Tx I/D of 6wk could lead to a significant improvement without ≥Grade 2 in 95% of pts.
Conclusions: Longitudinal safety modeling and simulation were conducted to quantitatively understand AE time profile and determine the effect of treatment interruption of Vismo in oBCC pts. Our work exemplified how longitudinal PK/PD modeling can help elucidate the onset and offset of AE events and the effect of Tx I/D, thereby enabling optimization of regimen duration.
References:
[1] Kowalski KG, McFadyen L, Hutmacher MM, Frame B, Miller R. A Two-Part Mixture Model for Longitudinal Adverse Event Severity Data. J Pharmacokinet Pharmacodyn. 2003;30(5):315-36
Reference: PAGE 23 () Abstr 3254 [www.page-meeting.org/?abstract=3254]
Poster: Drug/Disease modeling - Oncology