FGFR3::TACC3 fusions in head and neck carcinomas: a study of nine cases highlighting phenotypic heterogeneity, frequent HPV association, and a morphologically distinct subset in favor of a putative entity

The FGFR3::TACC3 fusion has been reported in subsets of diverse cancers including urothelial and squamous cell carcinomas (SCC). However, the morphology of FGFR3::TACC3-positive head and neck carcinomas has not been well studied and it is unclear if this fusion represents a random event, or if it might characterize a morphologically distinct tumor type. We describe nine FGFR3::TACC3 fusion-positive head and neck carcinomas affecting six males and three females aged 38 to 89 years (median, 59). The tumors originated in the sinonasal tract (n = 4), parotid gland (n = 2), and one case each in the oropharynx, submandibular gland, and larynx. At last follow-up (9-21 months; median, 11), four patients developed local recurrence and/or distant metastases, two died of disease at 11 and 12 months, one died of other cause, one was alive with disease, and two were disease-free. Three of six tumors harbored high risk oncogenic HPV infection (HPV33, HPV18, one unspecified). Histologically, three tumors revealed non-keratinizing transitional cell-like or non-descript morphology with variable mixed inflammatory infiltrate reminiscent of mucoepidermoid or DEK::AFF2 carcinoma (all were HPV-negative), and three were HPV-associated (all sinonasal) with multiphenotypic (1) and non-intestinal adenocarcinoma (2) pattern, respectively. One salivary gland tumor showed poorly cohesive large epithelioid cells with prominent background inflammation and expressed AR and GATA3, in line with a possible salivary duct carcinoma variant. Two tumors were conventional SCC. Targeted RNA sequencing revealed an in-frame FGFR3::TACC3 fusion in all cases. This series highlights heterogeneity of head and neck carcinomas harboring FGFR3::TACC3 fusions, which segregates into three categories: (1) unclassified HPV-negative category, morphologically distinct from SCC and other entities; (2) heterogeneous group of HPV-associated carcinomas; and (3) conventional SCC. A driver role of the FGFR3::TACC3 fusion in the first category (as a potential distinct entity) remains to be further studied. In the light of available FGFR-targeting therapies, delineation of these tumors and enhanced recognition is recommended.