Does Close Margin Alone Warrant Postoperative Adjuvant Therapy in Oral Squamous Cell Carcinoma?
Sophie Corbett-Burns, Sydney Ch'ng, Norm Stanton, Kan Gao, Jonathan R Clark, Ruta Gupta. Royal Prince Alfred Hospital, Sydney, Australia; Sydney Head & Neck Cancer Institute, Sydney, NSW, Australia
Background: There are major variations between centers regarding postoperative adjuvant therapy for adverse factors in oral squamous cell carcinoma (SCC) with general consensus that close margins alone warrant adjuvant therapy. Our practice has not been to recommend adjuvant therapy for close but uninvolved margins alone. This study examines the implications of close margins in oral SCC using evidence based methods and assesses whether our local failure rate in this cohort was acceptable.
Design: The English literature (1976-2012) was systematically reviewed for best evidence of prognostic implications of close margins in oral SCC for meta-analysis (CMA, Biostat, NJ, USA). A retrospective, single arm, non-inferiority study with local failure as the primary end point and maximum 'acceptable' pre-specified local failure rate of 15% at 5 years (y) as not requiring adjuvant therapy was designed to evaluate clinical feasibility. Margin proximity, tumor differentiation, thickness, patterns of invasion (POI), perineural and lymphovascular invasion (PNI, LVI) were reviewed. Statistical analyses included power analysis, sample size estimation, Log-rank test and Kaplan-Meier survival curves (Stata 11, TX, USA).
Results: Twenty five retrospective (level III) studies reported close but uninvolved margins in 4833 oral SCC patients. Meta-analysis showed higher risk of adverse events in patients with close margins (OR:1.803, 95%CI:1.4-2.3, p 0.00). Funnel plot showed moderate data heterogeneity and asymmetry(I2:61). Our cohort included 144 patients (M:F=79:65, median age 64y, mean follow up 3.3y) with surgery alone, with histological close (<5mm) margins. The sample size needed to detect 10% detriment in local control (LC) from 100% with 80% power (p=0.05) was 80 patients with 5y follow up and no censoring. LC in all patients with surgery alone was 91% (95%CI 82%–95%) at 5y. Worse LC was not seen with ordered stratification of close margin. The 5y LC rates for having 0, 1, 2, 3 and 4 additional adverse features of tumor thickness >4mm, infiltrative POI, PNI and buccal mucosa sub-site were 100%, 96%, 83%, and 71%, respectively (p=0.004).
Conclusions: Available best evidence suggests that close margins carry higher risk of adverse events with significant data heterogeneity, chiefly due to variable definitions of close margins, follow up and end points. Our findings indicate that surgery alone without adjuvant therapy offers acceptable LC if the only adverse feature is close margin.
Category: Head & Neck
Monday, March 4, 2013 9:30 AM
Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 189, Monday Morning