Elective neck dissection for salvage total laryngectomy: A systematic review, meta-analysis and "decision-to-treat" approach.

Abstract

The authors provide an updated, systematic and comprehensive summary of the literature concerning management of the N0 neck in patients for whom primary irradiation for squamous cell carcinoma of the larynx has been unsuccessful and salvage surgery in the form of total laryngectomy (TL) advocated. Bibliographic databases MEDLINE, Cochrane, PubMed and Embase were searched from inception to April 2019, with no language restrictions. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. Risk of bias was defined using the Joanna Briggs Institute guidelines. Outcome measures were defined as the impact of END on locoregional control, complication rate, disease-specific and overall survival (DSS and OS). The primary search identified 19 eligible articles, comprising 1353 patients, (1552 ENDs). The overall risk of occult metastases was 14% (9% of ENDs). The relative risk (RR) of developing complications was 1.29 when END was performed, compared to observation of the neck (CI 0.86-1.92). Contrariwise, patients in whom the neck was managed with neck dissection had a decreased risk of developing regional recurrence (RR 0.62, CI 0.35-1.08). There was no statistically significant variation between DSS and OS between END and neck observation groups, respectively. END during salvage TL may reduce the rate of regional recurrence, but not at the expense of improving DSS or OS. Rates of occult metastases, regional recurrence and "cure" through salvage neck dissection are not equivalent. Significant bias in all collated manuscripts should encourage the reader to interpret conclusions with caution. Patients should be fully involved in the decision-making process and their performance status and co-morbidities taken carefully into account when deciding to increase the extent of surgery, which we believe should remain limited to TL in the majority of cases.