AB0019 AUTOANTIBODY BIOMARKERS OF IMMUNE-RELATED ADVERSE EVENTS AND SURVIVAL IN MELANOMA PATIENTS RECEIVING IMMUNE CHECKPOINT INHIBITOR THERAPY

Background Immune checkpoint inhibitor (ICI) therapy has greatly improved the prognosis of advanced cancers but is often complicated by immune-related adverse events (irAEs). Biomarkers ICI safety and efficacy are needed. Objectives To identify autoantibodies (AAbs) associated with irAEs cancer outcomes in ICI-treated patients. Methods This analysis used clinical data biospecimens from a prospective trial 60 patients melanoma who received 2-4 doses combination (anti-PD1 + anti-CTLA4) followed anti-PD1 monotherapy. [1] Patients were for three years irAE development, progression, death. Plasma was collected at baseline six weeks after initiation assayed AAbs using PhIP-Seq human peptide library spanning entire proteome. Baseline levels ≥2 fold-change ≥10 follow-up (n = 833 AAbs) concurrently included bootstrapped LASSO Cox proportional hazards model (1000 replications; α 0.5; λ chosen cross-validation each replication) that separately run outcome: ≥1 severe (CTCAE grade ≥ 3), progression-free survival (PFS), overall (OS). The proportion times AAb replicated reported as its “validation frequency.” [2] Hazard ratio confidence intervals individually calculating bias-corrected accelerated (BCa) bootstrap replications) AAb. highest validation frequencies Partial Least Squares Discriminant Analysis (PLS-DA) clustering. Results Among patients, 47% developed irAE, 25% progressed, 23% died. Table 1 shows PFS, OS. PLS-DA able to cluster development only 10-fold cross-validated AUROC 0.86 ± 0.01 (p <.001) ( Figure ). performed similarly well PFS reflects higher identified worse OS (HR > 1); this pattern remained true across 25 (21/25, 84%) (23/25, 92%). Conversely, amongst top linked irAEs, lower majority (15/25, 60%) outcome < 1). Most unique outcome, there some overlap: shared seven (28%), two (8%), one (4%). No overlapped between more than outcomes. Conclusion These results suggest specific may be useful biomarkers predicting receiving therapy. However, these require validation. References [1]Postow MA, et al. Journal Clinical Oncology 2022;40:1059-1067 [2]Yang JJ, Blood 2012;120:4197-4204 1. ratios 95% Severe 60, 28) Progression-Free Survival (PFS) 15) Overall (OS) 14) Ratio (95% CI) VF* sVF** VF sVF AKAP9 1.25 (1.08-1.42) 76.4% 100 ACAN 1.72 (1.21-2.35) 66.3% KHSRP 1.34 (1.13-1.65) 52.8% KTN1 1.26 (1.02-1.50) 60.0% 79 POLR2A 1.29 (1.04-1.57) 62.7% 95 EN2 1.31 (1.05-1.61) 51.8% 98 GCN1 0.66 (0.45-0.95) 54.4% 78 ZGRF1 1.47 (1.03-1.81) 50.9% 77 SBNO2 1.33 (1.04-1.65) 50.0% FCHO2 1.21 (0.98-1.48) 49.5% 71 DYNC2H1 1.39 (0.97-1.85) 45.2% 68 PAK4 1.46 (1.16-1.87) 43.5% 82 NRXN2 0.72 (0.44-0.95) 44.2% 58 TRIOBP (1.05-1.43) 41.3% 62 1.35 (0.95-1.78) 39.0% 74 EPC1 1.28 (1.09-1.53) 43.2% 57 SDCCAG8 1.40 (1.08-1.74) 39.4% 59 1.43 (0.71-2.11) 34.0% 64 *VF Validation Frequency (proportion model) **sVF Scaled (validation frequency scaled value outcome) Acknowledgements: NIL. Disclosure Interests Carlos Aude: None declared, Nilasha Ghosh Grant/research support from: Hospital Special Surgery, Deanna Jannat-Khah Shareholder of: Walgreens Boots Alliance, AstraZeneca, Cytodyn, Karmela Kim Chan: Michael Postow Consultant BMS, Merck, Novartis, Eisai, Pfizer, Chugai, RGenix, Infinity, H Larman Scientific Advisory Board member TScan Therapeutics, Employee Founder ImmuneID, Portal Bioscience, Alchemab, Anne Bass Memorial Sloan Kettering Cancer Center, Rheumatology Research Foundation.