Development and validation of a nomogram prediction model for malignant transformation of oral potentially malignant disorders

Abstract

Objective: Oral potentially malignant disorders have increased the risk of oral squamous cell carcinoma. This study developed a nomogram model to assess the risks of malignant transformation of oral potentially malignant disorders.

Materials and methods: A retrospective analysis of patients diagnosed with oral potentially malignant disorders confirmed by pre-treatment biopsy was performed between 2010 and 2017 at the Peking University Hospital of Stomatology. The candidate risk factors for malignant transformation were screened from clinicopathological variables using Cox and stepwise regression analyses. The nomogram model was constructed based on the regression results and was validated through receiver operating characteristic curves and calibration curves. Decision curve analysis was used to estimate clinical usefulness.

Results: A total of 6964 cases of oral potentially malignant disorders were assessed. The malignant transformation rate of oral potentially malignant disorders was 2.00%. Risk factors (age, site, kind of oral potentially malignant disorder, existence of dysplasia and its grade, and other cancers) derived from the regression analyses were entered into the nomogram model. Time-dependent receiver operating characteristic curve, calibration curve, and decision curve analyses showed high levels of predictive value and clinical relevance, although not for all oral potentially malignant disorders.

Conclusion: A specific dynamic nomogram could be adopted to predict the malignant transformation of oral potentially malignant disorders and implement interventions.

Keywords: Dysplasia; Nomogram prediction model; Oral leukoplakia; Oral lichen planus; Oral potentially malignant disorders; Oral squamous cell carcinoma.