Oncologists’ preferences for first-line (1L) treatment of ALK-positive metastatic NSCLC in the US: A discrete choice experiment (DCE).

Abstract

e20631 Background: Trade-offs between efficacy, safety, and administration of 1L treatments for ALK+ mNSCLC are evolving and may influence shared decision making. This study assessed heterogeneity in provider preferences for attributes of 1L ALK+ mNSCLC treatment. Methods: An online survey of US oncologists who have treated at least 1 ALK+ mNSCLC patient in the past year was performed June-August 2024. Participants answered 12 questions comparing two 1L profiles with varying levels of 8 treatment attributes (Table) for a patient with no brain metastases. Latent class analysis (LCA) explored preference heterogeneity and a Wilcoxon test examined differences in respondents’ characteristics across latent groups. Aggregated relative attribute importance (RAI) (0-100%) was reported. Results: 201 oncologists (65% community) completed the DCE. LCA identified 2 groups. Group 1 (61%) treated fewer total cancer patients per month than Group 2 (39%) [Median=200 vs 300; p<0.001] and fewer ALK+ mNSCLC patients per year [Median=7 vs 10; p=0.016]. Groups 1 and 2 both balanced efficacy (RAI=44.5% vs 48.2%) and safety (RAI=39% vs 41.2%), but preferred different levels in 5 of the 8 attributes. For 3-year PFS, Group 1 preferred 64% over 46% or 43%, while Group 2 did not distinguish between different levels. For iDOR, Group 1 preferred >36 or 28 months over 17 months, while Group 2 preferred >36 months over 28 months, which was preferred over 17 months. For edema, Group 1 preferred 12% over 20%, which was preferred over 54%, while Group 2 preferred 12% or 20% over 54%. The table displays the statistically preferred levels for each group. Conclusions: We found two groups of oncologists, varied by patient volume, with different preferences for efficacy (3-year PFS, in particular), safety, and other attribute levels. With various 1L ALK+ mNSCLC treatments available, oncologists may balance multiple treatment objectives when making shared treatment decisions with patients. 1L treatment attributes, levels, and statistically preferred levels by latent group. Attributes Levels Group 1 (N=122) † Group 2 (N=79) † 3-year PFS, % (95% confidence interval) 43 (34, 51)46 (38, 55)64 (55, 71) 64 vs43 or 46 No differences* Median intracranial duration of response (iDOR), months 1728>36 >36 or 28 vs17 >36 vs28 vs17 16-month risk of adverse event, % Any edema 122054 12 vs20 vs54 12 or 20 vs54 Any cognitive effects <1020 <10 <10 Grade ≥3 interstitial lung disease /pneumonitis 0.20.61.9 0.2 or 0.6 vs1.9 0.2 vs0.6 or 1.9 Any myalgia 141821 No differences* No differences* 2L+ treatment ChemotherapyApproved 2L+ ALK TKIs Approved 2L+ ALK TKIs Approved 2L+ ALK TKIs Administration 1 pill a day (with or without food)4 pills twice a day with food 1 pill a day No differences* † Only statistically preferred levels are listed; vs = 95% CIs for the attribute levels did not overlap. *No statistically significant differences between any levels.