Can videostroboscopy predict early glottic cancer? A systematic review and meta‐analysis. (3rd November 2015)

• Record Type: Journal Article
• Authors: Mehlum, Camilla S.; Rosenberg, Tine; Groentved, Aagot M.; Dyrvig, Anne‐Kirstine; Godballe, Christian
• Is Part Of: Laryngoscope. Volume 126:Number 9(2016:Sep.)
• Page Start: 2079
• Page End: 2084
• Publication Date: 2015-11-03
• Subjects: Videostroboscopy -- early glottic cancer -- vocal cords -- cancer -- premalignancy -- diagnostic accuracy -- biopsy -- meta‐analysis; Otolaryngology -- Periodicals; 617.51005
• DOI: 10.1002/lary.25745
• Languages: English
• ISSNs: 0023-852X

Abstract

Abstract : Objective: Correct assessment of patients with a glottic lesion is crucial for ensuring proper treatment in cases of cancer or premalignancy and for avoiding unnecessary surgery. For years, videostroboscopy (VS) has been the gold standard for assessing such lesions, but diagnostic difficulties have been described. We aim to estimate the diagnostic accuracy of VS in differentiating early glottic cancer from noninvasive lesions by conducting a systematic review and meta‐analysis of published studies. Data Sources: PubMed and Embase databases were searched without restrictions on publication date. Review Methods: A systematic review and subsequent meta‐analysis were performed according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis statement. We systematically searched the literature for publications on stroboscopic diagnosis of cancer or premalignant lesions on the vocal cords. All retrieved studies were reviewed and qualitatively assessed. The pooled sensitivity and specificity of VS were calculated, and bubble and summary receiver operating characteristics plots were created. Results: A meta‐analysis was conducted on five studies with a total of 307 patients. The sensitivities of VS within the single studies ranged from 86% to 100% and specificities ranged from 7% to 93%. The meta‐analysis showed that the sensitivity of the combined results was 0.96 (95% confidence interval [CI]: 0.89‐0.98), and the specificity was 0.65 (95% CI: 0.21‐0.93). Conclusion: VS is able to identify almost all patients with cancer, but only approximately two‐thirds of patients with noninvasive lesions are correctly identified as not having cancer. Further research concerning assessment of patients with vocal cord lesions is needed. Level of Evidence: NA Laryngoscope, 126:2079–2084, 2016