Abstract:

Objective To investigate the enhanced CT and MRI imaging features of nasal sinus squamous cell carcinoma （SCC） and lymphoma （NHL）， and to analyze the efficacy of different imaging features in differentiating nasal sinus SCC from NHL. Methods The imaging， clinical and pathological data of 67 patients with sinus SCC and NHL who underwent sinus CT and MRI with contrast CT and MRI in our hospital and confirmed by surgical pathology were retrospectively analyzed， and the tumor origin， maximum diameter， CT density， MRI signal intensity， enhancement degree， tumor internal necrosis， adjacent bone destruction， invasion of surrounding tissues， and The imaging features such as cervical lymph node metastasis within the scanning range were analyzed， and the receiver operating characteristic （ROC） curve and area under the curve （AUC） were used to analyze the efficacy of different imaging features to distinguish nasal sinus SCC and NHL. Results There were statistically significant differences between the five imaging features of nasal sinus SCC and NHL， including tumor origin， maximum diameter， internal tumor necrosis， surrounding bone destruction and peripheral tissue invasion （P < 0.05）， and the AUC of differentiating SCC and NHL were 0.708， 0.694， 0.785， 0.850 and 0.629， respectively. The AUC of SCC and NHL was 0.969， and the sensitivity and specificity were 83.9% and 97.2%， respectively. Conclusion On contrast?enhanced CT and MRI， the imaging signs of tumor origin， maximum diameter， tumor internal necrosis， bone destruction and surrounding tissue invasion are helpful to distinguish nasal sinus SCC from NHL， especially if the tumor originates in the nasal cavity， necrosis is rare， bone destruction is mild， and the possibility of nasal sinus NHL should be given priority. Contrast?enhanced CT and MRI can help differentiate nasal sinus SCC from NHL， and the combination can help improve differential diagnostic performance.

Key words: nasal sinuses, lymphoma, squamous cell carcinoma, image features, differential diagnosis