Symptomatic adverse events in lung cancer clinical trials included in US drug labels from 2015 to 2021 to inform patient-reported outcomes.

Abstract

e13642 Background: FDA recommends assessing expected symptomatic adverse events (symAEs) as a core patient-reported outcome (PRO) in cancer clinical trials (CTs). Several core PRO symptom lists exist that measure the same symptoms across trial contexts; however, many were developed in an era when cytotoxic chemotherapy was the main mechanism of action (MOA). To further research relevant PROs in modern cancer CTs, we identified common symAEs from recently approved lung cancer drugs across MOAs. Methods: We identified lung cancer drug approvals from 2015 through 2021 using an FDA database. The drug US Prescribing Information (PI) was used to extract indication, MOA, CT information, and all-grade symAEs reported in ≥20% of patients in the investigational arm. To improve symAE attribution to the drug, we limited analysis to CTs with a monotherapy investigational arm. We determined the most common symAEs based on the number of CTs reporting a specific symAE in ≥20% of patients. Results: Of 54 approvals identified, 31 lung cancer CTs (30 for NSCLC, 1 for SCLC) with monotherapy investigational arms were included, with AE data from 9,829 patients. Drug MOAs were classified broadly as growth pathway inhibitors (GPI) for 21 CTs (68%), immunotherapy (IO) for 9 CTs (29%), and cytotoxic chemotherapy for 1 CT (3%). Most common symAEs are presented in Table 1 (truncated to symAEs in ≥9 CTs, based on All CTs). In comparing MOAs, symAEs reported at ≥20% incidence in ≥5 GPI CTs but in no IO CTs included: diarrhea (76% of GPI CTs), edema (38%), nail toxicity/paronychia (29%), vomiting (29%), stomatitis (29%), and dry skin (24%). Fifteen other symAEs were reported at ≥20% incidence in ≥3 GPI CTs, but in no IO CTs [data not shown]. Conclusions: While there were several common symAEs across lung cancer CTs, our analysis using AE data in PIs of recently approved drugs showed important variability between CTs and broad GPI and IO groups, demonstrating that consideration of MOA and expected toxicity are important for ensuring relevant high frequency symAEs are included as PROs. Final selection of expected symAEs from a PRO item library is context dependent, requiring consideration of all available clinical data and patient input.[Table: see text]