CYLD mutation characterizes a subset of HPV-positive head and neck squamous cell carcinomas with distinctive genomics and frequent cylindroma-like histologic features

Mutations in the tumor suppressor CYLD, known to be causative of cylindromas, were recently described in a subset of high-risk (hr) HPV-positive head and neck squamous cell carcinomas (HNSCC). Pathologic and genetic characterization of these CYLD-mutant carcinomas, however, remains limited. Here, we...

Full description

Saved in:
Bibliographic Details
Published inModern pathology Vol. 34; no. 2; pp. 358 - 370
Main Authors Williams, Erik A., Montesion, Meagan, Alexander, Brian M., Ramkissoon, Shakti H., Elvin, Julia A., Ross, Jeffrey S., Williams, Kevin Jon, Glomski, Krzysztof, Bledsoe, Jacob R., Tse, Julie Y., Mochel, Mark C.
Format Journal Article
LanguageEnglish
Published New York Elsevier Inc 01.02.2021
Nature Publishing Group US
Elsevier Limited
Subjects
Online AccessGet full text
ISSN0893-3952
1530-0285
1530-0285
DOI10.1038/s41379-020-00672-y

Cover

More Information
Summary:Mutations in the tumor suppressor CYLD, known to be causative of cylindromas, were recently described in a subset of high-risk (hr) HPV-positive head and neck squamous cell carcinomas (HNSCC). Pathologic and genetic characterization of these CYLD-mutant carcinomas, however, remains limited. Here, we investigated whether CYLD mutations characterize a histopathologically and genomically distinct subset of hrHPV-positive HNSCC. Comprehensive genomic profiling via hybrid capture-based DNA sequencing was performed on 703 consecutive head and neck carcinomas with hrHPV sequences, identifying 148 unique cases (21%) harboring CYLD mutations. Clinical data, pathology reports, and histopathology were reviewed. CYLD mutations included homozygous deletions (n = 61/148; 41%), truncations (n = 52; 35%), missense (n = 26; 18%) and splice-site (n = 9; 6%) mutations, and in-frame deletion (n = 1; 1%). Among hrHPV-positive HNSCC, the CYLD-mutant cohort showed substantially lower tumor mutational burden than CYLD-wildtype cases (n = 555) (median 2.6 vs. 4.4 mut/Mb, p < 0.00001) and less frequent alterations in PIK3CA (11% vs. 34%, p < 0.0001), KMT2D (1% vs. 16%, p < 0.0001), and FBXW7 (3% vs. 11%, p = 0.0018). Male predominance (94% vs. 87%), median age (58 vs. 60 years), and detection of HPV16 (95% vs. 89%) were similar. On available histopathology, 70% of CYLD-mutant HNSCC (98/141 cases) contained hyalinized material, consistent with basement membrane inclusions, within crowded aggregates of tumor cells. Only 7% of CYLD-wildtype cases demonstrated this distinctive pattern (p < 0.0001). Histopathologic patterns of CYLD-mutant HNSCC lacking basement membrane inclusions included nonkeratinizing (n = 22, 16%), predominantly nonkeratinizing (nonkeratinizing SCC with focal maturation; n = 10, 7%), and keratinizing (n = 11, 8%) patterns. The latter two groups showed significantly higher frequency of PTEN alterations compared with other CYLD-mutant cases (38% [8/21] vs. 7% [8/120], p = 0.0004). Within our cohort of hrHPV-positive HNSCCs, CYLD mutations were frequent (21%) and demonstrated distinctive clinical, histopathologic, and genomic features that may inform future study of prognosis and treatment.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:0893-3952
1530-0285
1530-0285
DOI:10.1038/s41379-020-00672-y