Value Of MRI Features And Texture Analysis To Evaluate The Texture Of Pituitary Adenomas

Objective: To explore the differences in texture of pituitary adenomas with age,gender,and secretory function;to compare the differences in surgical methods and resection levels of different texture adenomas;to explore the MRI features and texture analysis to evaluate the value of pituitary adenoma textures.Methods: The clinical and imaging data of 103 patients with pituitary gland adenoma confirmed by surgery and pathology were analyzed retrospectively at Yijishan Hospital of Wannan Medical College.According to the intraoperative findings,it is divided into the soft(n= 59)and the hard group(n = 44);according to the secretory function of the tumor,it is divided into functional groups(including the secretion of one or more hormones)(n = 79)And non-functional(n = 24).All patients underwent conventional MRI plain scan and enhanced examination before operation,and the pathological diagnosis was clear.Analyze the tumor MRI features;manually draw the ROI layer by layer on the MRI image,and use GE's AK software to extract the texture features.Categorical variables were compared using chi-square or Fisher's exact test,continuous variables were compared using two independent sample t-tests or Mann-Whitney U test;feature selection was performed using LASSO regression analysis,and MRI features and texture parameters were analyzed using single-factor and multi-factor logistic regression.Modeling,drawing ROC to evaluate MRI features,texture parameters,and diagnostic effectiveness of the model.Results: There were no statistically significant differences in pituitary adenomas between the soft and hard groups with respect to age,sex,and hormone secretion function,and the differences between the surgical methods and the degree of resection were statistically significant(p <0.05).The MRI features of the two texture pituitary adenomas were statistically different between the largest tumor diameters and T2 WI signals.The AUC values of the largest tumor diameters and T2 WI discrimination between the two tumors were 0.774 and0.688,respectively.AK software extracted a total of 3132 texture parameters from images ofdifferent sequences of MRI(including T1 WI,T2WI,and T1 enhanced sequences),and finally screened a total of 9 texture parameters with higher value.The texture parameters identified the two tumors with AUC values of 0.659 to 0.847.Among T2 WI weighted sequences,Compactness2 has the highest efficacy in identifying two tumors(AUC value 0.847),and its sensitivity and specificity for differential diagnosis are 0.864 and 0.763,respectively.In multivariate logistic regression analysis,the combined model of MRI features and texture parameters has the highest efficiency(AUC value of 0.995)in identifying the two tumors,which is superior to that of MRI features and texture parameter models(AUC values of 0.774 and 0.990 respectively).Conclusions:1.The texture of pituitary adenomas has nothing to do with sex,age,and secretory function of the tumor,and is related to the maximum diameter of the tumor.When the maximum diameter is greater than 2.35 cm,the texture of the tumor is significantly harder.2.The texture of pituitary adenomas has nothing to do with MRI T1 WI signals and T1 WI enhanced signals;it is related to T2 WI signals,and tumors with low T2 WI signals have a harder texture.3.The diagnostic performance of MRI texture parameter model,MRI features combined with texture parameter model is significantly better than that of MRI features alone.It has high diagnostic value for judging the texture of pituitary adenomas,and has good sensitivity and specificity.4.Massive soft tumors prefer transsphenoidal resection,most of which are total or subtotal.Hard tumors tend to be craniotomy,and most of them are partial resection.