利用零差K分布模型参数识别乳腺肿瘤

Use of homodyned K-distribution parameter for characterizing breast neoplasms

目的探讨利用零差K分布模型进行超声组织定征的临床价值。方法利用零差K分布模型参数对恶性、良性乳腺肿瘤进行分析。首先,从43例乳腺肿瘤(包含乳腺纤维瘤以及乳腺浸润性导管癌)病例图像中选取出感兴趣区域进行分析。然后计算良性、恶性乳腺肿瘤区域的零差K分布标准化参数值βnorm,比较良性、恶性乳腺肿瘤所提取的βnorm值的差异。结果良恶性肿瘤的βnorm值存在显著差异。结论利用零差K分布模型的参数能有效地区分良性、恶性乳腺肿瘤,具有临床应用价值。

Objective To investigate the efficacy of the parameter in characterizing breast lesions as benign or malignant. Methods The parameter in homodyned K-distribution was used in this paper. First, regions of interest were chosen from 43 patients with breast masses （including breast adenofibroma and breast cancer）. Second, the normalized values of the parameterβnorm were calculated and compared. Results The parameterβnorm gained both from benign and malignant breast masses showed an obvious discrepancy. ConcIusion The parameter in homodyned K-distribution is efficient in classifying the breast lesions as benign and malignant.