Abstract
IMPORTANCE: Frozen section (FS) analysis of sentinel nodes offers potential on-table diagnosis and treatment for occult metastasis in oral squamous cell cancer. Systematic analysis of FS during sentinel node biopsy has not been illuminated in the literature. OBJECTIVE: To systematically review pooled data from studies using FS analysis in evaluating sentinel nodes in patients with cT1-T2 N0 oral squamous cell cancer. DATA SOURCES: An academic librarian led the search of CENTRAL, CINAHL, Cochrane Database of Systematic Reviews, Embase, and MEDLINE for studies published in English between January 2000 and January 2023. STUDY SELECTION: Two authors independently screened cohort studies, case series, and randomized clinical trials, in which FS analysis was used to evaluate sentinel nodes in patients with cT1-T2 N0 oral squamous cell cancer. DATA EXTRACTION AND SYNTHESIS: Data were extracted by 2 reviewers. Reporting quality was estimated using the Diagnostic Precision Study Quality Assessment Tool. Data analysis was performed between April and July 2023, and the meta-analysis was completed using the bivariate random-effects model. MAIN OUTCOMES AND MEASURES: The primary outcome was the pooled sensitivity of FS sentinel node analysis. Secondary outcomes included evaluation of the FS technique, rate of occult metastasis, false-negative rate, and survival. RESULTS: Seventeen articles with 878 patients met the eligibility criteria. Although protocols varied, confirmatory serial step sectioning was performed in all studies. Occult metastasis was found in 263 of 878 patients (30%), and FS analysis identified 173 cases (65.8%). Following serial sectioning, an additional 90 positive results were identified, leading to 47 patients undergoing staged completion neck dissection. The pooled sensitivity of FS was 0.71 (95% CI, 0.60-0.80), the diagnostic odds ratio was 110, and the false-negative rate was 34.2%. The Cochrane Q value was 15.62 (df = 16; P = .48) and τ2 = 0.36. CONCLUSION AND RELEVANCE: In this systematic review and meta-analysis, evaluated studies showed various techniques, in which pooled sensitivity reached 0.71, providing a benchmark for comparison to other 1-stop approaches. Due to the high false-negative rate of approximately one-third of patients, intraoperative FS must always be supplemented by serial sectioning. On-table diagnosis remains a key objective for sentinel node biopsy, and FS detection may be improved by standardizing protocols.