Objectives: The concept of adequate surgical margins remains controversial in oral squamous cell carcinoma (OSCC) surgery. This study aimed to identify surgical margin-related indicators that might impact recurrence and survival of OSCC patients. Materials and Methods: Histopathological examination was performed using hematoxylin-eosin-stained surgical margin tissue sections in 235 OSCC patients. Axin2 and Snail expression at the surgical margin was detected by immunohistochemistry. The impact of the Axin2-Snail cascade on tumorigenesis of the immortalized human oral keratinocyte (IHOK) line was investigated in vivo. Results: The width and dysplasia of surgical margins were not significantly associated with the outcome of OSCC patients. In a multivariate analysis using variable clinicopathologic factors and with Axin2 and Snail expression as cofactors, higher age (hazard ratio [HR]:1.050; P=0.047), Axin2 (HR:6.883; P=0.014), and Snail abundance (HR:5.663; P=0.009) had independent impacts on worsened overall survival. Similarly, lesion site in retromolar trigone (HR:4.077; P=0.010), upper (HR:4.332; P=0.005) and lower gingiva (HR:3.545; P=0.012), presence of extranodal extension (HR:9.967; P<0.001), perineural invasion (HR:3.627; P=0.024), and Snail abundance (HR:3.587; P<0.001) had independent impacts on worsened recurrence-free survival. Furthermore, Axin2 knockdown induced decreased Snail expression and attenuated tumorigenesis in the IHOK line. Conclusion: Histopathological examination of surgical margins may not be reliable to predict OSCC patient outcome. Molecular analysis may provide a more accurate risk assessment of surgical margins in OSCC. In particular, Axin2 and Snail are potential predictive biomarkers for the risk assessment of surgical margins in OSCC.
Bibliographical noteFunding Information:
We thank Ana Luisa Carvalho for datasets 3 and 4, Débora Serrenho for testing the software and Asim Iqbal for reading the manuscript and providing feedback. A.Ö.A. was partially supported by Fundação para a Ciência e a Technologia (FCT) grant SFRH/BD/51264/2010. E.E. was supported by Personalized Health and Related Technologies (PHRT), project number 222, ETH domain. A.F.H. was supported by FCT grant SFRH/BD/51265/2010. The authors would also like to acknowledge the financial support of the European Research Council (ERC) via a Starting Grant to T.K (No. 679175), Fundação Bial (161/10-2010) and (PTDC/SAU-NMC/122035/2010) to I.I., Consejo Nacional de Ciencia y Tecnología (254878) and Programa de Apoyo a Proyectos de Invertigación e Inovación Tecnológica/UNAM (IN207420) to Y.R.C., and the Scientific and Technological Research Council of Turkey (113E603).
Copyright © 2022 Pei, Han, Kim, Kim, Nam, Kim, Cho, Kim, Cha and Zhang.
All Science Journal Classification (ASJC) codes
- Cancer Research