A prospective cohort study of HLA-B44 supertype as predictor of response to novel immune checkpoint combinations.

Abstract

e15071 Background: To date there are no clear predictors of response to novel immune checkpoint inhibitor (ICI) combinations. Recent reports suggest that HLA-I (Human Leukocyte Antigen Class I) supertype B44 may predict survival to PD1(L)1 inhibition in patients with melanoma and non-small cell lung cancer (NSCLC) ( Chowell et al, Science 2018). We sought to study if this effect is exclusive for PD(L)1 inhibition alone and the impact of HLA-B44 on ICI combinations with new immunotherapeutic agents. Methods: Between 2017 and 2019, 77 patients treated across 11 different early phase clinical trials (NSCLC: 64: (83,1%); melanoma: 4: (5,2%), bladder 8 (10,4%); and one nasopharynx carcinoma), ICI-naive were prospectively evaluated for HLA-I subclass. All patients were ECOG 0 or 1 (50 and 27 respectively). Most patients had ≤ 2 metastatic sites (55: 71,4%); 24 (31,2%) patients were treated with PD(L)1 inhibitors as monotherapy, and 53 (68.8%) with PD(L)1-containing combinations including ICI and agonists such as ICOS, OX40 or GITR monoclonal antibodies. HLA-I was classified for supertypes and the variables of progression-free survival (PFS) and overall survival (OS) were compared for patients with HLA-B44+ and B44-, in both groups (monotherapy and combinations). Results: 44 patients were HLA-B44+ and 33 were negative. For patients treated with anti-PD(L)1 agents as monotherapy, those that were HLA-B44+ had a longer PFS: [medianPFS: 22,38 months (m) (95% CI: 13,39 – 31,37); vs 3,8m (95% IC: 2,05 – 5,65); p: 0,002] and OS: [median: 33,2 m (95% CI: 23,7 – 42,6) vs 14,46 m (95% IC: 7,86 – 21,1); p: 0,13]. In contrast, these marked differences were not evident in patients receiving PD1 combination therapies [HLA-B44+ mPFS: 7,17m (95% CI: 4,41 - 9,93) vs 9,46 m (IC95% 6,8 m – 12,07); p:0,290], and OS: 11,53m (95% IC: 8,33 – 14,7) vs 18,9 m (IC 95% 13,3 – 24,5) p: 0,24. Lastly, response rate was also better for patients with HLA-B44 positive treated with PD1 monotherapy (66,7% vs 16,7% in HLA-B44 negative patients) p: 0,048, whilst when administering ICI combinations, HLA-B44 negative patients seemed to have a better response rate (44% vs 27% in HLA-B44 positive patients; p = 0,148). Conclusions: These results confirm in a prospective cohort the predictive impact of HLA-B44 supertype for patients treated with anti-PD(L)1 agent and suggest a potential benefit of combining checkpoint inhibitors in patients with other HLA supertypes (non-HLA-B44+).