基于常规T2WI灰度共生矩阵纹理参数鉴别诊断卵巢实性肿瘤良恶性

Gray level co-occurrence matrix based on T2WI in differential diagnosis of benign and malignant ovarian solid tumors

目的探讨基于常规T2WI的灰度共生矩阵(GLCM)纹理参数鉴别诊断良恶性卵巢实性肿瘤的价值。方法回顾性分析经病理证实的19例卵巢原发实性良性肿瘤和27例恶性肿瘤患者的术前MRI和临床资料。采用Image J软件于轴位T2WI提取病灶的GLCM纹理参数,包括能量、对比、相关、逆差距和熵,比较卵巢实性良恶性肿瘤间各参数的差异;绘制ROC曲线,评价其鉴别诊断卵巢实性良恶性肿瘤的效能。结果卵巢实性恶性肿瘤的熵和对比高于良性肿瘤,逆差距和能量低于良性肿瘤(P均<0.05);而相关差异无统计学意义(P=0.074)。ROC曲线结果显示,能量、对比、逆差距和熵诊断卵巢实性良恶性肿瘤的AUC值分别为0.70、0.68、0.74和0.81(P均<0.05),敏感度分别为68.4%、74.1%、73.7%和77.8%,特异度分别为77.8%、68.4%、77.8%和73.7%。结论基于T2WI的GLCM纹理参数鉴别诊断良恶性卵巢实性肿瘤具有一定价值,其中熵值的诊断效能较高。

Objective To explore the diagnostic value of gray level co-occurrence matrix(GLCM)based on conventional T2WI in differential diagnosis of benign and malignant ovarian solid tumors.Methods Preoperative MRI and clinical data of patients with pathologically confirmed primary ovarian solid tumor(19 benign and 27 malignant ones)were retrospectively analyzed.On axial T2WI,the texture parameters of GLCM(energy,contrast,correlation,inverse difference,entropy)of lesions were extracted using Image J software.The differences of texture parameters between benign and malignant solid tumors were compared.ROC curve was drawn to evaluate the efficacy of texture parameters in diagnosis of benign and malignant ovarian solid tumors.Results The values of inverse difference and energy in malignant ovarian solid tumors were lower than those in benign ovarian solid tumors,while the values of entropy and contrast were higher than those in benign ovarian solid tumors(all P<0.05).In addition,no significant difference was observed in the correlation between benign and malignant ovarian solid tumors(P=0.074).In differentiation between benign and malignant ovarian solid tumors,the AUC of energy,contrast,inverse difference and entropy was 0.70,0.68,0.74 and 0.81,respectively,with the sensitivity of 68.4%,74.1%,73.7%,77.8%,the specificity of 77.8%,68.4%,77.8%and 73.7%,respectively.Conclusion The texture parameters of GLCM based on T2WI have some value in differential diagnosis of benign and malignant solid ovarian tumors,among which,the diagnostic efficacy of entropy value is the highest.