Histogram Analysis Method On Multimodality MRI For Differential Diagnosis Of Pituitary Adenomas

Part oneHistogram Analysis method for differentiation of Cystic Pituitary Adenomas and Rathke Cleft CystsPurpose:Similar radiological features on MRI imaging between cystic pituitary adenomas and Rathke cleft cysts often make difficulty of differential diagnosis,and it is critical to distinguish them preoperatively.The aim of this study was to estimate the potential value of histogram analysis on T1-,T2-weighted and post-contrast T1-weighted images for differentiating diagnosis between these two diseases.Methods:The retrospective study involved 85 patients with pathological confirmed cystic pituitary adenomas(n=47)and Rathke cleft cysts(n=38).T1-,T2-weighted,and post-contrast T1-weighted images of lesions were used for histogram analysis.Group comparisons of parameters from histogram analysis were performed using independent sample t-test.The area under the curve(AUC)with receiver operating characteristic curve(ROC)analyses was conducted between the two patient groups.Furthermore,a support vector machines(SVD)classification analysis was used,aiming to detect the best combining strategy of histogram parameters to improve the differential diagnosis capability,and the outcomes were compared with the results of conventional radiological reading.Results:Group comparisons showed that there were significant differences(P<0.05)of the histogram parameters between these two lesions,including variance and kurtosis derived from T2-weighted images,standard deviation,variance,energy,entropy and range derived from post-contrast T1-weighted images.The optimal multi-parameter model for differential diagnosis calculated in the SVM classification consisted of energy,entropy,range and mean value derived from post-contrast T1-weighted images,of which the diagnostic accuracy rate was high up to 89.4%.In contrast,the diagnostic accuracy rate of conventional radiological reading was 69.4%..Conclusion:Histogram analysis of post-contrast T1-weighted Images can improve the accuracy rate for differential diagnosis between cystic pituitary adenoma and Rathke cleft cyst in contrast to conventional radiological reading.Part TwoDCE-MRI quantitative analysis and MRI histogram analysis for differentiation between prolactinoma and nonfunctional pituitary adenomaPurpose:Identification of nonfunctional pituitary adenoma and prolactinoma is of great importance for the choice of treatment.The aim of this study was to investigate the diagnostic value of dynamic contrast-enhanced-magnetic resonance imaging(DCE-MRI)quantitative parameters and histogram parameters for differentiating nonfunctioning pituitary adenoma and prolactinoma.Methods:25 patients with nonfunctioning pituitary adenoma confirmed clinically and 25 patients with prolactinoma confirmed clinically were performed routine MRI plain and enhanced scan,and DCE-MRI.Histogram parameters on the T1WI,T2WI,T1WI enhanced images,DCE-MRI perfusion quantitative parameters were calculated by software,and histogram parameters of DCE-MRI corresponding perfusion quantitative parameters were calculated respectively.The independent sample t test was used to compare the differences between two groups,The area under the curve(AUC)with receiver operating characteristic curve(ROC)analyses was conducted.And a support vector machines(SVM)classification analysis was used,aiming to detect the best combining strategy of histogram parameters.Results:Group comparisons caculated by independent sample t test showed that there were no significant differencesof the histogram parameters between nonfunctioning pituitary adenoma and prolactinoma on T1WI,T2WI and post-contrast T1WI(P>0.05).DCE-MRI perfusion quantitative parameters including Ktrans and Kep were statistically significant(P<0.05).The DCE-MRI histogram parameters were statistically significant on the mean value,mean deviation,skewness,quantile25,quantile 50,quantile 75 from Ktrans map and energy and entropy fromVe map(P<0.05).The optimal multi-parameter model for differential diagnosis among prolactinoma and nonfunctioning pituitary adenoma calculated in the SVM classification consisted of Kep,mean value derived from Ktrans map,entropy and energy derived from Ve map,of which the diagnostic accuracy rate was high up to 82%.Conclusion:DCE-MRI quantitative analysis and DCE-MRI histogram analysis are useful for differential diagnosis between nonfunctioning pituitary adenoma and prolactinoma.