Exploring systemic treatment approaches for advanced pure large cell neuroendocrine carcinoma (LCNEC): A multicenter retrospective analysis.

8106 Background: Due to lack of prospective data, the optimal first line treatment approach for patients (pts) with pure LCNEC histology remains uncertain. Methods: Across 17 centers, we conducted a retrospective analysis of metastatic pure LCNEC (diagnosed at local institutions) who received 1 st -line systemic therapy between 2015-2023. Pts were either treated with chemotherapy (chemo), immunotherapy (IO), or a combination (chemoIO). Clinical outcomes were progression-free survival (PFS), overall survival (OS), objective response rates (ORR), and treatment-related adverse events (trAE) as defined using CTCAE 5.0. Survival analysis by genomic alteration status ( TP53, RB1, STK11, KEAP1, KRAS) and PD-L1 status were performed. Results: We identified 161 pts with median age of 67 years (IQR: 58-83) at 1st line systemic therapy; 54% (n = 87) were males. Median follow-up was 55 months (mo), 62 mo, and 29 mo for the chemo, IO, and chemoIO groups, respectively. 79% (n = 127) were former or current smokers. 1st-line treatments were chemo (n = 94), IO (n = 11), or chemoIO (n = 56). Of 85 pts with PD-L1 status, 58 (68%) were 0% and 27 (32%) were ≥1%. The most common chemo regimen was platinum-etoposide (n = 72, 78%). ChemoIO regimens included carboplatin/etoposide/atezolizumab (n = 23, 43%) and carboplatin/pemetrexed/pembrolizumab (n = 17, 31%). In the IO group, 1 patient received dual IO. Lung (n = 74, 46%) and liver (n = 71, 41%) were the most common sites of metastasisThere was no significant difference in PFS across the groups (median PFS mPFS chemoIO: 5.7 mo, 95% CI: 5.0-6.3, chemo: 5.1 mo, 95% CI: 3.1-5.9; IO: 3.6 mo, 95% CI: 1.7-6.5) on multivariable analysis adjusting for ECOG and “M” stage (p = 0.2 and 0.24 for chemoIO-chemo and chemoIO-IO comparisons, respectively). There was no difference in OS (mOS chemo: 11 mo, 95% CI: 7.6-17.4; IO: 13.6 mo, 95% CI: 5.9-not reached; chemoIO: 12.2 mo, 95% CI: 7.4-20.6, p = 0.5 and 0.8 for chemoIO-chemo and chemoIO-IO comparisons, respectively). ORR was 35.2% (31/88) in chemo, 25% (2/8) in IO, and 35.7% (20/56) in chemoIO ( p= 0.46). There were no significant differences in PFS and OS outcomes by treatment group when divided by genetic profile (n = 79, methods), or by PD-L1 status (0% vs ≥1%). Any grade trAE occurred in 48 (51%), 5 (45%), and 28 (50%) pts treated with chemo, IO, and chemoIO, respectively. Grade ≥3 toxicity profiles are shown (table). Conclusions: For 1st-line treatment of LCNEC, similar PFS, OS, and ORR were seen for chemo, IO, and chemoIO. This questions the added benefit of chemoIO compared to chemo alone, and warrants future clinical trials to discern the optimal 1 st line treatment. Table: see text