The Impact of a Stepwise Approach to Primary Tumor Detection in Squamous Cell Carcinoma of the Neck With Unknown Primary.

Abstract

To examine the cumulative effect of diagnostic steps for primary tumor identification in patients with head and neck squamous cell carcinoma of unknown primary (HNSCCUP), including lingual tonsillectomy, and the impact of primary tumor identification on subsequent treatment. Retrospective analysis. We reviewed the records of 110 patients diagnosed with HNSCCUP between 2003 and 2015. Results of diagnostic imaging (fluorodeoxyglucose-positron emission tomography/computed tomography [FDG-PET/CT]), tumor detection with direct laryngoscopy with biopsies, palatine tonsillectomy, and transoral robotic surgery (TORS) lingual tonsillectomy were recorded. Associations between demographic and treatment variables with overall survival (OS) and progression-free survival (PFS) were modeled with Cox proportional hazards models. FDG-PET/CT was suspicious for a primary site in 23/77 (30%) patients. Direct laryngoscopy identified a primary tumor in 34/110 patients (31%). Forty-seven patients underwent palatine tonsillectomy, which identified 17 primaries (36%), yielding a cumulative primary tumor identification of 51/110 (46%). Fourteen patients underwent TORS lingual tonsillectomy, which identified eight primaries (57%), resulting in a cumulative identification of 59/110 (53%). The detection rate increased from 28/63 (44%) to 31/47 (66%) after the addition of TORS lingual tonsillectomy to our institutional approach. Detection rates varied by HPV status. Primary tumor identification altered subsequent radiation planning, as patients with an identified primary tumor received radiation to a smaller volume of tissue than did those without an identified primary tumor. However, there was no significant association between primary tumor identification and OS or PFS. A stepwise approach to primary tumor identification identifies a primary tumor in a majority of patients. 4 Laryngoscope, 129:1610-1616, 2019.