RESUMEN
OBJECTIVE: To generate a nomogram for predicting the risk of neck node metastasis in pathologically node-negative patients using a combination of variables comprising of protein expression, ultrastructural alterations and clinicopathological parameters. MATERIALS AND METHODS: Surgically removed oral tumours (n = 103) were analysed for the expression of desmosomal and hemidesmosomal assembly proteins by immunohistochemistry and ultrastructural alterations by transmission electron microscopy (TEM). Protein expression, ultrastructural alterations and clinicopathological variables were used to construct nomogram from the training set in 75 patients. Clinical utility of the nomogram was validated in a discrete set of 28 patients. RESULTS: Univariate and multivariate analyses were performed on the training set, and obtained significant variables comprising of integrin ß4 expression (p = .027), number of hemidesmosomes (p = .027)/desmosomes (p = .046), tumour differentiation grade (p = .033) and tumour thickness (p = .024) were used for construction of the nomogram. The area under the curve was calculated for both training 0.821 (95% CI 0.725-0.918) and validation sets 0.880 (95% CI 0.743-1.000). The nomogram demonstrated a predictive accuracy of 73.3% and 78.6% with the sensitivity of 81.4% and 83.3% in the training and validation sets, respectively. CONCLUSIONS: The nomogram constructed on postsurgical tumour samples will be a value addition to histopathology for the detection of neck node metastasis in pathologically node-negative patients.