2171 - Characterization of Decision Points, Metastatic Trajectories and Outcomes in Advanced NSCLC

R. A. Weber¹, M. Rao¹, D. R. Gomez², and P. Iyengar²; ¹Memorial Sloan Kettering Cancer Center, New York, NY, ²Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY

Purpose/Objective(s): Multiple prior studies have suggested a promising role for RT in oligometastatic non-small cell lung cancer (NSCLC). Conversely, a randomized phase II/III trial (NRG-LU002), examined adding RT to maintenance systemic therapy in this context but recently closed for failure to meet its pre-specified progression free survival (PFS) endpoint. Optimal patient selection for local therapy may, in fact, underlie these conflicting data. Understanding and predicting metastatic trajectories may identify ideal local therapy candidates. Here, we assessed disease behavior in advanced NSCLC patients as a function PFS1 and consequent effects on PFS2 and OS. Materials/

Methods: We compiled a single institution database of advanced NSCLC patients from 2014-2022 whose tumors contained mutations in EGFR (543), ALK (78), or neither (WT) (1122). Next, we defined PFS1 and PFS2 as any radiographic progression after first- or second-line therapies, respectively. We dichotomized patients into the following groups: a) progression-dependent systemic therapy switch (immediately after progression); and b) progression-independent systemic therapy switch (5-week delay to switch). We then associated both: 1) timing of systemic therapy switch at PFS1 with OS, and 2) PFS2 as a function of short (=4 months) or long (>12 months) PFS1 intervals in EGFR, ALK, or WT cohorts.
Results: Progression-independent switch in systemic therapy was utilized in 73%, 79%, and 73% of patients in EGFR, ALK, or WT cohorts, respectively. Progression-independent change of first-line systemic therapy at progression was associated with a ~40% improvement in median OS for all three cohorts. This improvement was statistically significant for some (EGFR: p<0.0001, WT: p<0.0001) but not all cohorts (ALK). Next, we considered how PFS1 interval related to PFS2 for both progression-dependent and progression-independent systemic therapy switch groups. For the former group, in the EGFR cohort, the median PFS2 for a short PFS1 was 1.414 whereas for a long PFS1 it was 5.671, p<0.001. We saw a similar, albeit non-statistically significant, trend for ALK and WT patients. For the progression-independent systemic therapy switch group, in the EGFR cohort, the median PFS2 for a short PFS1 was 2.137 and for a long PFS1 it was 6.049, p<0.001. For the WT cohort, the median PFS2 for a short PFS1 was 2.236 and for a long PFS1 it was 3.682, p<0.001. We observed a similar albeit non-statistically significant trend in this direction for the ALK cohort.
Conclusion: Practice patterns in advanced NSCLC patients increasingly involve delaying change of 1st line therapy beyond progression. The delay in switch of 1st line therapy may be justified by its improvement in OS over patients who are switched at 1st progression. Importantly, longer PFS1 intervals portend for longer PFS2 intervals. Taken together, these data yield initial insight into potential factors that predict metastatic trajectories and can guide future use of local therapy.