Tumour mutational burden predicts survival in patients with recurrent/metastatic head and neck carcinoma treated with durvalumab with or without tremelimumab

Up till now, the predictive value of tumour mutational burden (TMB) in patients with recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC) remained unclear. The EAGLE study is the first exploratory analysis of a phase III trial to demonstrate that blood TMB is predictive of overall (OS) and progression-free survival (PFS) in patients with R/M HNSCC treated with durvalumab with or without tremelimumab. The benefit in OS and PFS seen with durvalumab-based therapy over chemotherapy crossed the boundary of significance above a bTMB threshold of 16 mut/Mb. Interestingly, this benefit further increased with a higher level of bTMB.

Introduction

Durvalumab is an immune checkpoint inhibitor (ICI) that blocks the interaction between PD-L1 and its receptors while tremelimumab is an anti-CTLA-4 monoclonal antibody. Immune checkpoint inhibitors have a demonstrated clinical efficacy in R/M HNSCC although clinical response was observed in only a minority of patients. Tumour mutational burden (TMB) is associated with clinical benefits with ICIs in R/M HNSCC, however, its predictive value is unclear. As blood TMB (bTMB), assessed from circulating tumour DNA, showed promise as a predictive survival biomarker for immunotherapy in patients with NSCLC, the phase III EAGLE study evaluates bTMB as a predictor of survival in R/M HNSCC.

The phase III EAGLE study enrolled a total of 736 patients with squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx or larynx. All patients had disease progression after platinum-containing regimen for R/M HNSCC or disease recurrence within six months of platinum-based multimodal therapy with a curative intent. In order to be eligible for the study, patients had to have an ECOG performance status of 0 or 1. Patients were randomly allocated (1:1:1) to durvalumab monotherapy (10 mg/kg every 2 weeks [Q2W]), durvalumab plus tremelimumab (durvalumab 20 mg/kg every 4 weeks [Q4W] plus tremelimumab 1 mg/kg Q4W up to four doses, followed by durvalumab 10 mg/kg Q2W), or investigator's choice of a single-agent standard-of-care therapy with cetuximab, docetaxel, paclitaxel, methotrexate, 5-fluorouracil, TS-1, or capecitabine. The primary endpoint of the trial was OS.

Results

As presented earlier, the primary endpoint of a longer OS was not met in the intention to treat population (HR [95%CI]: 0.88 [0.72-1.08]; p=0.20 for durvalumab versus chemotherapy and HR [95%BI]: 1.04 [0.85-1.26]; p=0.76 for durvalumab plus tremelimumab versus chemotherapy). Of the 736 patients in the intention to treat population, 247 patients (34%) were evaluable for bTMB. In total, 74 patients (27 in the durvalumab arm, 20 in the durvalumab plus tremelimumab arm and 27 in the chemotherapy arm) were bTMB high.

Interestingly, in patients with a high bTMB (≥16 mut/Mb), durvalumab was shown to be associated with a significantly longer OS compared to chemotherapy (HR: 0.385). For patients with a lower bTMB no significant OS difference was observed (HR: 0.912). In patients with a high bTMB, the investigators also observed an incremental improvement in OS with an increasing bTMB (e.g. HR for OS of 0.385 for ≥16 mut/Mb patients dropping to 0.257 in patients with a bTMB ≥24 mut/Mb). A similar observation was made for the comparison of durvalumab + tremelimumab versus chemotherapy. Also in this setting a significant OS benefit was seen in patients with a TMB ≥16 mut/Mb (HR: 0.381). In patients with a ≥24 mut/Mb, the HR for OS of durvalumab + tremelimumab versus chemotherapy dropped to 0.292.

PFS was a key secondary objective of the study and also for this endpoint bTMB proved to be predictive. In fact, in patients with a bTMB ≥16 mut/Mb durvalumab with or without tremelimumab was found to be associated with a significantly longer PFS compared to chemotherapy (HR: 0.482 and 0.447, respectively). Also for this endpoint the clinical benefit of the durvalumab-based treatment increased with an increasing bTMB above the 16 mut/Mb threshold.

In patients with bTMB ≥16 mut/Mb, the median OS was reported at 8.1 and 7.6 months for durvalumab with or without tremelimumab, respectively, as compared to only 4.0 months with chemotherapy (HR [95%CI]: 0.38 [0.19-0.78], p=0.0061 and HR [95%CI]: 0.39 [0.2-0.76], p=0.0044, respectively). At 18-months, this translates into a 22% OS rate for durvalumab and tremelimumab, of 33% with durvalumab alone and of 0% with chemotherapy in patients with high bTMB. In contrast, in patients with a low bTMB, the median OS with durvalumab with or without tremelimumab was 6.8 months and 6.2 months, respectively, which was shorter than the median OS of 8.6 obtained with chemotherapy in this cohort. With respect to PFS, bTMB high patients had a 55% and 52% lower risk of disease progression or death when they were treated with durvalumab +/- tremelimumab compared to chemotherapy. The 12-months PFS rates were 17%, 28% and 0% for patients in the durvalumab plus tremelimumab, durvalumab and chemotherapy arm, respectively. In patients with a bTMB <16 mut/Mb the median PFS was 2.0 months for the two durvalumab-containing treatments as compared to 3.7 months with chemotherapy.

Finally, patients with mutations in KMT2D, a known tumour suppressor gene in HNSCC, demonstrated an improved OS with durvalumab plus tremelimumab compared to chemotherapy (HR [95%CI]: 0.39 [0.17–0.85], p=0.014). A trend of improved OS for durvalumab plus tremelimumab versus chemotherapy (HR [95%CI] = 0.19 [0.03–1.03], p= 0.035) was also seen in patients with ATM mutations.

Conclusions

This is the first exploratory analysis of a phase III trial showing that bTMB is predictive of OS and PFS for checkpoint inhibitors in R/M HNSCC. Hazard ratios for OS and PFS improved with increasing bTMB levels for durvalumab versus chemotherapy and durvalumab plus tremelimumab versus chemotherapy. Other mutations may also be predictive of survival for immunotherapy and warrant further investigation. Based on this exploratory analysis, further investigation and prospective validation of bTMB as a predictive biomarker for immunotherapy in R/M HNSCC is warranted.

Reference

Li W, Wildsmith S, Ye J, et al. Plasma-based tumor mutational burden (bTMB) as predictor for survival in phase III EAGLE study: Durvalumab (D) ± tremelimumab (T) versus chemotherapy (CT) in recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC) after platinum failure. Presented at ASCO 2020; Abstract 6511.