The importance of log odds of positive lymph nodes for locoregional recurrence in oral squamous cell carcinoma

Objectives: Log odds of positive lymph nodes (LODDS) has been demonstrated as a very promising staging model for multiple cancer sites, as it avoids singularity and predicts prognosis significantly better than conventional nodal staging and lymph node ratio. However, published studies on the influence of LODDS for patients with OSCC are very seldom. Material and methods: Retrospective chart review of 499 patients with treatment-naive oral squamous cell carcinoma. Exclusion criteria were neoadjuvant chemoradiotherapy, T4b classification, perioperative death, unresectable disease, synchronous malignancy, follow-up < 3 months and inadequate information to correctly determine nodal yield. Statistical analysis was performed using univariate and multivariate analysis. Results: A significant correlation was found between locoregional recurrence and pathologic T classification (p=0.030), pathologic N classification p=0.013), extracapsular spread (p=0.034), grading (p=0.021), number of positive lymph nodes (p=0.042), lymph node ratio (p=0.009), LODDS (p=0.007) and treatment strategy (p=0.039). Multivariate analysis indicated lymph node ratio (p=0.029) and LODDS (p=0.015) as independent indicators for locoregional recurrence. Within the analyzed models, Nagelkerke R2 index and Somers D showed the strongest discrimination ability for LODDS. Conclusion: For patients with oral squamous cell carcinoma, log odds of positive lymph nodes and lymph node ratio are independent indicators for locoregional recurrence. LODDS predicts locoregional recurrence better than conventional nodal staging system, lymph node ratio and the number of positive lymph nodes. (C) 2017 Elsevier Ltd. All rights reserved.