The Role Of Texture Analysis Of Diffusion Weighted Images In Differential Diagnosis Of Intracranial Tumors Based On Gray Level Co-occurrence

PartⅠ:Value of Apparent Diffusion Coefficient in detection primary tumors in patients with brain metastasesObjective:To investigate the value of diffusion weighted imaging（DWI）and apparent diffusion coefficient（ADC）in detection primary tumors in patients with brain metastases.Materials and methods:105 patients with brain metastases confirmed by clinical and imaging follow-up or histopathology were enrolled,including 53 cases from lung cancer,23 cases from breast cancer and 29 cases from gastrointestinal tumor.The ADC of solid portion of tumors and peri-tumor edema were measured,the clinical ability of ADC value was evaluated by analyzing the data of different originate brain metastases in statistically.Results:The ADC values of solid portion in brain metastases of lung cancer,breast cancer,gastrointestinal tumors were 0.693±0.120×10^-3 mm²/s,0.857±0.088×10^-3mm²/s,0.782±0.150×10^-33 mm²/s,with statistically difference between the three groups（F=15.321,P<0.001）;the ADC value of peri-tumor edema were 1.630±0.158×10^-3 mm²/s,1.590±0.238×10^-3 mm²/s,1.601±0.181×10^-3 mm²/s,with no statistically difference（F=0.462,P=0.631）.Conclusion:The ADC values of the common brain metastases is related to the type of primary tumor.Therefore,ADC values may be used to detect the origin tumor for those patients who could not identify primary tumors,and provide help for clinical treatment.Part Ⅱ: Texture analysis of diffusion weighted images of glioblastoma and solitary brain metastases based on gray level co-occurrence matrixObjective: To investigate the value of diffusion weighted imaging texture analysis of glioblastoma and solitary brain metastases based on gray level co-occurrence matrix.Materials and methods: 51 patients with glioblastoma and 48 patients with solitary brain metastases confirmed by histopathology were enrolled in our study.Five commonly used textural features of Energy,Entropy,Inertia,Correlation and Inverse Difference Moment（IDM）of tumor area and peritumoral edema area in two groups of images were extracted through gray level co-occurrence matrix.The independent sample t test was used to compare the differences of the texture parameters between the two groups.Receiver operating characteristic（ROC）curve was constructed to analyze the diagnostic efficacy of every textural feature.Results: Four textural features of Energy,Inertia,Correlation and IDM of tumor area and peritumoral edema area were statistically significant in the differential diagnosis of glioblastoma and solitary brain metastases（P<0.05）.With ROC analysis,all the parameters have diagnostic efficiency in the differentiation between glioblastoma and solitary brain metastases,of which the Energy and Inertia value of tumor area had higher accuracy.Conclusion: Diffusion weighted image texture analysis based on gray co-occurrence matrix is helpful to identify glioblastoma and solitary brain metastases.