(T-071) Population Pharmacokinetic/Pharmacodynamic Modeling of Tumor Size Dynamics in Lorlatinib-treated Patients with Non-small Cell Lung Cancer

Nathan Braniff – Pfizer; Alan Liu – Pfizer; Avanka Gunatilaka – ARC Center for Personalized Therapeutics Technologies; Joseph Chen – Genentech; Jerry Li – Pfizer; Yazdi Pithavala – Pfizer; Jennifer Hibma – Pfizer

Objectives: Lorlatinib is an approved small molecule inhibitor of Anaplastic Lymphoma Kinase (ALK) for the treatment of patients with ALK-positive non-small cell lung cancer (NSCLC). The present analysis sought to develop a population pharmacokinetic/pharmacodynamic model to predict tumor size dynamics over time in NSCLC patients treated with lorlatinib.

Methods: Tumor size data, in the form of sum of lesion diameter timeseries measurements, were used to develop a population tumor growth inhibition model. The model was developed using data from two studies: 1) a phase 1/2 study of lorlatinib (NCT01970865) in NSCLC patients (N=294) with ALK or ROS1 mutations receiving doses ranging from 10mg QD through 200mg QD, and 2) a phase 3 study of lorlatinib (NCT03052608; CROWN study) in previously untreated patients (N=142) with ALK-positive NSCLC receiving a 100mg QD dose of lorlatinib. Tumor size dynamics over the course of the follow-up period were modeled as a function of the net effect of cell killing and residual growth using nonlinear mixed effects modeling. NONMEM (v7.50) and R (v4.3.1) were used for model development. Several assumptions and model structures were tested including alternative parameterizations and the inclusion of different forms of resistance to capture the observed tumor size dynamics [1].

Results: The selected final model was a variant of a previously reported biexponential model [2], which can specifically capture instances of extended tumor stasis. The model characterized the tumor killing effect for lorlatinib via an exponential decay term and included a second exponential term to capture the effects of both stasis and resistance. Model selection and adequacy were evaluated using visual diagnostics and measures of goodness-of-fit. The model yielded typical value estimates of 2.06E-2 1/day for the tumor shrinkage rate (Kd) and 8.71E-5 1/day for the rate of resistant growth (Kg), with an approximately equal estimated typical fraction of sensitive and resistant tumor mass.

Conclusions: The model developed for this analysis adequately described lorlatinib anti-tumor activity over the study. This model may be used to characterize the time dynamics of tumor response to lorlatinib and help identify patterns with higher probability of tumor shrinkage across patient subpopulations.

Citations: [1] Ribba, B., et al. "A review of mixed‐effects models of tumor growth and effects of anticancer drug treatment used in population analysis." CPT: pharmacometrics & systems pharmacology 3.5 (2014): 1-10.
[2] Chatterjee, M. S., et al. "Population pharmacokinetic/pharmacodynamic modeling of tumor size dynamics in pembrolizumab‐treated advanced melanoma." CPT: pharmacometrics & systems pharmacology 6.1 (2017): 29-39.

Keywords: Tumor Growth Inhibition Model