Abstract 130P
Background
The majority of patients treated with ICIs develop irAEs. It is currently not possible to predict the development of irAEs using biomarkers. Here we evaluated the IgG autoantibodies (AAbs) profiles in pre-treatment sera of metastatic melanoma patients treated with ICIs to identify AAbs that are associated with high-grade irAEs.
Methods
Clinical data of patients treated with pembrolizumab or nivolumab monotherapy (n=48) or combination ipilimumab and nivolumab (n=37) was retrospectively evaluated. irAEs were graded using CTCAE v5.0. Baseline sera were assessed for IgG AAbs using the HuProt™ microarray v4.0 covering 23,059 proteins. AAb profiles of patients with grade 3-4 irAEs were compared to those of patients with no irAEs (control group). Serum scores and a volcano plot (p < 0.05 and log2FC > 0.6) were used to identify differentially expressed AAbs. Principal component analysis and feature selection methods using the Boruta Package and Random Forrest feature importance were carried out in R. AAbs identified by two or more methods were selected and their diagnostic accuracy was further evaluated through ROC curve analysis. iPathwayGuide software (Advaita) was used to further explore their functional characteristics.
Results
Out of 85 patients, 40 experienced grade 1-2 irAEs, 20 grade 3-4 irAEs and 25 no irAEs. We found 473 differentially expressed AAbs in patients experiencing grade 3-4 irAEs relative to controls. Of these, a combination of 8 AAbs was able to correctly classify patients experiencing grade 3-4 irAEs with a ROC analysis showing an AUC =1.0. The proteins recognised by these 8 AAbs are expressed in tissues that are commonly affected by irAEs. Pathway analysis revealed enrichment of AAbs in pathways associated with immunity and autoimmunity.
Conclusions
We identified an 8 autoantibody signature with high accuracy that discriminates between patients experiencing grade 3-4 irAEs compared to those without any irAEs. Further analyses are being conducted to identify the correlation of irAE type and severity to specific autoantibodies. Prospective studies are required for validation of these AAbs specificities.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Centre for Precision Health, Edith Cowan University.
Funding
Edith Cowan University.
Disclosure
T. Meniawy: Non-Financial Interests, Personal, Advisory Board: GSK, AstraZeneca Australia, Takeda Pharmaceutical Australia Pty Ltd., Novartis Pharmaceutical Australia Pty Ltd., Bristol Myers Squibb; Non-Financial Interests, Institutional, Other: Eisai. M. Millward: Non-Financial Interests, Personal, Advisory Board: TOGA; Non-Financial Interests, Personal, Member of the Board of Directors: MASC Trials; Non-Financial Interests, Personal, Member: MOGA, WA Melanoma Health Study Management Committee, Cancer Trials Australia Phase I Group; Non-Financial Interests, Personal, Project Lead: Australian Genomic Cancer Medicine Program. All other authors have declared no conflicts of interest.