Development of a Clinical Prediction Model to Predict Malignant Transformation of Sinonasal Inverted Papilloma Based on Hematological Indices and Clinical Features

Abstract

Objective

Sinonasal inverted papilloma (SNIP) has a risk of malignant transformation into SNIP with squamous cell carcinoma (SNIP-SCC). Early detection of SNIP-SCC is crucial for patient outcomes. This study aimed to evaluate the utility of tumor-related hematological indices in the early diagnosis of SNIP-SCC and to develop a nomogram incorporating imaging and clinical features.

Methods

This study included 159 patients with SNIP with (n = 34) or without (n = 125) SCC. Univariate and multivariate logistic regression analyses were employed to identify independent risk factors and develop nomogram. The diagnostic model was evaluated using receiver operating characteristics (ROC), standard, clinical decision, and clinical impact curves.

Results

Elevated serum cytokeratin fragment antigen 21-1 (CYFRA 21-1), the loss of convoluted cerebriform pattern, bone destruction, headache/facial pain, and epistaxis/blood-tinged mucus were identified as independent risk factors for SNIP-SCC. The optimal cut-off value for serum CYFRA 21-1 was 3.51 ng/mL, with the nomogram area under the ROC curve of 0.966 based on the above indicators.

Conclusion

The findings suggest that CYFRA 21-1 is a promising diagnostic marker for SNIP-SCC. The nomogram incorporating serum CYFRA 21-1, imaging features, and clinical factors demonstrates strong clinical utility and can be a reference tool for clinical decision-making.