Quantitive Assessment Of Pituitary Microadenoma With Dynamic Mr Imaging

Objective: To quantitively analyse the dynamic MR parameters of normal adenohypophysis and microadenomas and explor the optimal diagnosistic threshold of these parameters for pituitary microadenomas.Method: Dynamic MR enhancement Imaging was undertaken in 30 normal femal adenohypophysis and 30 patients with pituitary microadenoma and SI-T curves were divided into 3 types, Shape of SI-T curves, peak time, slop and enhance rate were statistically compared between normal adenohypophysis and microadenoma, the diagnostic threshold of pituitary microadeoma was determined with ROC, the sensitivity, specificity were calculated as well. The regression equation and coefficient of the variance were acquired by logisitic regression analysis.Result:1. Curves of type I was mainly seen in nomal adenohypophysis, and curves of typ III SI-T curves was only seen in microadenoma, and curves of typ II was seen in both normal and abnormal tissue.2. There was no statistical difference between respective position of normal adenohypophysis.3. Peak time, enhance rate, slope of SI-T curves were statistically different between normal tissue and microadenoma ,the average value of peak time, enhance rate, slope in normal tissue were 48.53s, 158.97% and 3.53, and microadenoma were 86.4s, 139.35% and 1.82.4. ROC analysis showed that the area under ROC curve of peak time, curve of enhance rate and curve of enhance slope was 0.896, 0.643 and 0.868 and the area under conbined ROC curve of 3 variances 0.907.5. The regression equation of the variance were get by logisitic regression analysis: P=1/[1+e-(3.699-0.081X1+0.012X2+0.345X3)]. Conclusion:1. Peak time, slop of dynamic enhanced MR is valuable in diagnosis of pituitary microadenoma. With 66 seconds peak time as the threshold of pituitary microadenoma, the diagnostic sensitivity, specificity were 76.7%, 84.4%, respectively. With slop of SI-T curves 2.015 as threshold of microadenoma, the sensitivity, specificity were 88.5%, 67%, respectively. With all of the three variance 0.5663 as threshold, the sensitivity, specificity were 92%, 87%, respectively.2. The three types of SI-T curves can discriminate the normal and abnormal pituitary tissue in a certain degree.