Abstract
Lorlatinib is highly effective in patients with ALK-positive NSCLC but frequently causes toxicity, which requires dose modifications or treatment cessation. To explore exposure-guided dosing, we assessed the relationship between lorlatinib plasma exposure and severe toxicities using data from two cohorts of patients: a discovery cohort and an external validation cohort (n = 46 and n = 52, respectively). Severe toxicity, defined as adverse events grade more than or equal to 3 or those leading to dose interruption, reduction, or cessation or patient hospitalization, was observed in 43% and 46% of patients, predominantly neurocognitive symptoms, metabolism disorders, and edema. In a competing risk analysis, higher lorlatinib exposure significantly increased the risk of toxicity (log-transformed Ctrough adjusted subdistribution hazard ratio: 3.45 [95% confidence interval: 1.54-7.72; p = 0.003]) in the combined cohorts (n = 98), corrected for age and performance status. An optimal threshold of 147 ng/mL was identified in the discovery cohort using receiver operating characteristic analysis (area under the curve: 0.81). Patients below this threshold had a lower risk of toxicity in both the discovery cohort and the validation cohort (subdistribution hazard ratio: 0.26 and 0.35; p = 0.003 and 0.012, respectively). On the basis of results from these two independent cohorts of patients, early dose reduction in patients with a lorlatinib Ctrough above the toxicity threshold, representing approximately one-third of patients, might prevent severe toxicity and improve treatment tolerability.