Robert Seitz, Head of Immune Oncology at Oncocyte speaks about AACR 2021 Abstract: 23 – Validation of a 27-gene immuno-oncology algorithm in metastatic urothelial carcinoma treated with an immune checkpoint inhibitor.
Link to Abstract:
https://www.abstractsonline.com/pp8/#!/9325/presentation/1088
Synopsis:
Foreground
Patients of metastatic urothelial carcinoma (mUC) have a poor prognosis, with just around 10% of patients remaining two years after diagnosis. The IMvigor210 study included patients with mUC who were treated with the immune checkpoint inhibitor (ICI) atezolizumab in a multicenter, single-arm, phase 2 trial. This trial’s RNA expression, total survival, tumor mutation burden (TMB), and PD-L1 IHC results are all open to the public. Using three related ICI molecules, the 27-gene immuno-oncology (IO) algorithm previously demonstrated an important correlation with ICI response in NSCLC and TNBC. Prior to testing in ICI-treated cohorts, the algorithm and positivity threshold were defined.
Methodologies
Using accessible RNA expression data from 348 patients in the IMvigor210 study, the prospectively specified 27-gene IO algorithm and threshold (IO Score) were used to identify patients as possible responders or non-responders. The coefficient for the IO Score was calculated in both a univariate and multivariate equation with PD-L1 IHC and TMB using Cox proportional hazard models for overall survival.
Conclusions
The IO Score was positive in 142 patients in all (41 percent ). In multivariate analyses of PD-L1 and TMB (HR=0.64; 95 percent CI = 0.45 to 0.92; p 0.02), positivity was substantially correlated with response (HR=0.61; 95 percent CI = 0.46 to 0.80; p 0.001). In univariate analysis, both PD-L1 IHC staining and TMB were important, but only TMB remained significant in the multivariate equation.
Final Thoughts
These findings show a connection between the 27-gene IO algorithm and the binary IO Score and ICI treatment response in a third tissue indication (TNBC, NSCLC, and mUC) using four separate ICI agents (two anti-PD-L1 and two anti-PD1). In all three trials, the algorithm and classification threshold were defined prior to research and did not vary between tissue indications. Furthermore, only the IO Score held relevance in multivariate regression with PD-L1 IHC and TMB in all three signs. Taken together, these findings point to the 27-gene IO signature as a pan-cancer indicator of ICI therapy response.