Magnetic resonance imaging findings of inverted papilloma: Differential diagnosis with malignant sinonasal tumors

Background: Computed tomography and magnetic resonance imaging (MRI) are the techniques of choice for pretreatment staging in neoplasms of the sinonasal tract. In inverted papilloma (IP), the information provided by computed tomography regarding characterization of the lesion is rather nonspecific. The present retrospective study was performed in the attempt to identify distinguishing features of IP on MRI. Methods: MRI examinations of 23 patients affected by IP (16 primary and 7 recurrent) and 23 patients affected by malignant tumors (MT; 12 adenocarcinomas, 9 squamous cell carcinomas, and 2 neuroendocrine carcinomas) of the sinonasal tract were evaluated. IP arose from the lateral nasal wall in 17 cases, the maxillary sinus in 5 cases, and the nasal septum in 1 case. The signal intensity of IP and MT was compared with muscles on spin-echo (SE) T2 and SE T1 images; contrast enhancement was compared with nasal septum mucosa. Possible specific MRI patterns in the two groups of patients were investigated. Bone involvement was graduated as remodeling or erosion (focal, less than or equal to15 mm; intermediate, >15 mm and less than or equal to30 mm; extended, >30 mm). The size of the lesions was assessed by measuring the greatest diameter on MRI Parametric statistics in the form of Student's t-test or chi-squared test was used for data comparison. Results: IP showed a columnar pattern in all 23 cases by enhanced SE T1 images and in 16 of 23 lesions (>20 mm in diameter) by SE T2. This pattern was observed in only 1 of the 23 MTs; pathological examination of that specimen showed multiple foci of IP associated with squamous cell carcinoma. Bone remodeling was observed in 19 of 23 IPs, which in five patients was associated with focal (two cases) or intermediate (three cases) erosion. In MT, remodeling was present, which was always combined with focal (2 cases) or extended (21 cases) erosion. A strong correlation was found between the pattern of bone changes and histology (p = 0.00001). Bone alterations did not correlate with the size of the IP. The mean size of the IPs was significantly less than that of MT (33.9 +/- 15.7 mm versus 59 +/- 16 mm; p = 0.0003). Conclusion: A columnar pattern is a reliable MRI indicator of IP and reflects its histological architecture (positive predictive value of 95.8%). The combination of this finding with the absence of extended bone erosion allows for the confident discrimination of IPs from MTs.