Diffusion restriction - low ADC - has been observed in Warthin tumors (94,7%) and malignant tumors (94,7%).
96,5% of pleomorphic adenomas presented with high ADC values.
ROC curve analysis allowed us to establish an ADC value of 1,3 x 10-3 mm2/s to be cut-off value in differentiation between pleomorphic adenomas and other lesions (with the area under the curve of 0,918).
Solid tumors and tumors with solid and cystic components with ADC values larger than 1,3 x 10-3 mm2/s are suggestive of benign lesions,
with exclusion of Warthin tumor,
and most frequently correspond with pleomorphic adenomas and other types of adenomas.
ADC values lower than 1,3 x 10-3 mm2/s were shown by both Warthin tumors and malignancies (e.g.
carcinomas,
lymphomas),
therefore those lesions may require cytology examination to exclude malignancy. In our material,
we have not found a statistically significant difference in ADC values between Warthin tumors (benign lesions) and malignant lesions.
Type A enhancement curve has been observed in 93% of pleomorphic adenomas (Fig.
2) and 71% and 71,4% of other adenomas.
36,8% of malignant lesions presented type A enhancement curve.
Type B TIC has been reported in 55% of Warthin tumors (Fig.
3) and in one case of malignant tumor.
42% of Warthin tumors and 57,8% of malignant lesions presented with type C enhancement curve (Fig.
4).
Type D TIC was observed in case of 2 cystic lesions.
Tables 3 and 4 present signs of diffusion restriction in reference to histopathological diagnosis and time-signal intensity curves of evaluated parotid tumors based on observations of the firest observer; note the excellent intrerobserver agreement (Table 5).
Lymphadenopathy was present in 41 cases - accompanied 24 benign lesions and 17 out of 19 malignant masses.
Signs of perineural spread were present in 13 cases,
all corresponding to malignancy.
With the use of the Cohen statistics we found substantial agreement between radiological and histopathological diagnosis - kappa 0,76,
while agreement in accuracy of diagnosis between FNAC and histopathologic examination was moderate - kappa 0,63.
Interobserver agreement in tumor characterization was excellent with kappa value of 0,94.
No significant difference was found between the efficacy of radiologic evaluation and histopathological result.
Results of statistical analysis are displayed in table 5.
The highest agreement on diagnosis for all three methods (radiological assessment,
FNAC and histopathology) was reported in malignancies,
pleomorphic adenomas and Warthin tumors.