HRAS Mutations Define a Distinct Subgroup in Head and Neck Squamous Cell Carcinoma

Abstract

Purpose: In head and neck squamous cell carcinoma (HNSCC), HRAS mutation is a new actionable oncogene driver. We aimed to evaluate HRAS mutational variants, comutation profile, and survival outcomes of this molecularly defined population.

Methods: We leveraged four deidentified patient data sets with HRAS-mutant HNSCC, MD Anderson Cancer Center, Kura Oncology, Inc trial, Foundation Medicine, and American Association for Cancer Research GENIE v.12. Patient demographic information and clinical courses were extracted, when available, in addition to HRAS mutation type and co-occurring mutations. Survival outcomes were analyzed (Kaplan-Meier method).

Results: Two hundred forty-nine patients with HRAS-mutant HNSCC were identified from the four data sets. Median age ranged from 55 to 65 years, with a higher frequency in male patients (64%); the majority of HRAS-mutant HNSCC occurred in human papillomavirus-negative HNSCC. HRAS mutation patterns were similar across data sets; G12S was the most common (29%). Treatment responses to tipifarnib were not codon-specific. Compared with wild-type, significantly co-occurring mutations with HRAS were Casp8 (Fisher's exact test, P < .00013), TERT (P < .0085), and NOTCH1 (P < .00013). Analysis of clinical courses from the MD Anderson Cancer Center and Kura Oncology, Inc data sets demonstrated poor clinical outcomes with a high rate of recurrence following primary definitive treatment (50%-67% relapse < 6 months) and short disease-free survival (4.0 months; 95% CI, 1.0 to 36.0) and overall survival (OS; 15.0 months; 95% CI, 6.0 to 52.0). Use of tipifarnib in this data set demonstrated improved OS (25.5 months; 95% CI, 18.0 to 48.0).

Conclusion: Oncogenic mutations in HRAS occur in 3%-4% of HNSCC, with G12S being the most frequent. Without targeted therapy, patients with HRAS-mutant HNSCC had poor clinic outcomes; observable trend toward improvement in OS has been noted in cohorts receiving treatments such as tipifarnib. The comutation pattern of HRAS-mutant in HNSCC is distinct, which may provide insight to future therapeutic combination strategies.

FIG 1.

Functional sites of HRAS mutations across four data sets: (A) MDACC cohort, (B) Kura cohort, (C) FMI cohort, and (D) AACR GENIE v.12 cohort. Lollipop plots indicate the location of HRAS mutations documented in all four cohorts. AACR, American Association for Cancer Research; FMI, Foundation Medicine Inc; Kura, Kura Oncology Inc; MDACC, MD Anderson Cancer Center.

FIG 2.

HRAS-mutant HNSCC versus HRAS wild-type populations in the Foundation Medicine Inc cohort. In this cohort, 155 unique HRAS-mutant HNSCC cases were identified from a total of 4,759 HNSCC sequenced. Patients with HRAS-mutant HNSCC are denoted in red; patients with HRAS wild-type are denoted by blue. Most patients were older than 59 years (n = 105 [67.7% in HRAS-mutant patients, n = 2,660], 58.3% in HRAS wild-type); female patients were in the minority in both groups (34.8% of HRAS-mutant patients were female [54/155]; 23.7% of HRAS wild-type patients were female [1,081/4,604]). Patients with both HRAS-mutant and HRAS wild-type HNSCC were more likely to be defined as HPV-negative disease (20.6% [32/155] of patients with HRAS-mutant HNSCC; 32.5% [1,485/4,604] of patients with HRAS wild-type, P < .0015). **P < .044; ***P < .0016. HNSCC, head and neck squamous cell carcinoma; HPV, human papillomavirus; MSI-H, microsatellite instability High; PD-L1, programmed death-ligand 1; TMB, tumor mutation burden.

FIG 3.

Comutational profile of HRAS-mutant versus WT populations in the Foundation Medicine Inc cohort. P values were calculated using a Fisher's exact test and were corrected for multiple testing. Data displayed using (A) bidirectional bar chart and (B) volcano plot. **P < .0085; ***P < .0003; ****P < .00013. Mutations in CASP8, NOTCH1, and TERT are identified as statistically significant comutations in the HRAS-mutant HNSCC population. Amplifications in FGF and CCND1 are identified as statistically significant and mutually exclusive in the HRAS-mutant HNSCC population. HNSCC, head and neck squamous cell carcinoma; WT, wild-type.

FIG 4.

Best overall response to tipifarnib by HRAS mutation in the Kura data set. Twenty patients were evaluable in the Kura data set. Tumor responses were demonstrated across codons 12, 13, 61, and 22. *Denotes the three patients with human papillomavirus–positive disease who achieved stable disease as best response. PR, partial response.

FIG 5.

Survival analyses performed in clinical data sets (MD Anderson Cancer Center and Kura): (A) Median DFS, (B) median OSmet, and (C) median OS in each respective cohort. DFS, disease-free survival; Kura, Kura Oncology, Inc; MDACC, OS, overall survival; OSmet, metastatic OS.