Narrow band imaging in head and neck unknown primary carcinoma: A systematic review and meta-analysis

Objective To estimate the diagnostic performance of Narrow Band Imaging (NBI) in patients with cervical metastasis from head and neck squamous cell carcinoma of unknown primary (SCCUP) origin. Methods PubMed, Embase, and Scopus databases were systematically scrutinized up to July 1, 2019, looking for studies that encompassed the NBI in the SCCUP diagnostic work up. The main inclusion criteria for eligible articles for the meta-analysis were non-evidence of primary tumor after physical examination and conventional cross-section imaging before NBI assessment and the availability of complete data on the diagnostic accuracy of NBI. A set of random-effects model meta-analyses was then performed following the Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines. Results Five studies, conducted between January 2003 and September 2016, comprising 169 patients imaged with NBI, were included in the meta-analysis. The pooled sensitivity and specificity of NBI in patients with head and neck SCCUP was 0.83 (99% CI, 0.54-0.95) and 0.88 (99% CI, 0.55-0.97), respectively. The positive and negative likelihood ratios were 6.38 (99% CI, 1.6-25.44) and 0.06 (99% CI, 0.005-0.86). The pooled diagnostic odds ratio (DOR) was 82.15 (99% CI, 7.06-955). The overall detection rate of NBI was 0.35 (99% CI, 0.18-0.53), which allowed localization the primary tumor in 61 out of 169 patients, otherwise not detected by the usual diagnostic work-up. Conclusions Current available evidence suggests that NBI has a considerable diagnostic accuracy in patients affected by head and neck SCCUP. Laryngoscope, 2019