基于灰度特征值的混杂性肺原位腺癌与微浸润性腺癌的鉴别

Differentiation of AIS and MIA Based on Gray Feature Yalue

为了探讨灰度特征值对表现为混杂的肺原位腺癌(AIS)和微浸润性腺癌(MIA)的鉴别价值,文章首先利用大津法与掩模法等方法将肺实质提取出来,随后利用计算机分别计算了0°、45°、90°、135°方向上的灰度特征值(能量、熵、逆差矩和相关性)、均值与统计学P值。发现灰度共生矩阵中病灶的惯性矩能够作为两者的判别标准:AIS与MIA的惯性矩均值分别为13.8513±0.0001、13.8562±0.0021。得出结论:灰度特征值对于AIS与MIA的诊断具有一定的指导意义,当惯性矩值≤13.8513时,病灶为AIS的可能性很大,当惯性矩值≥13.8513时,病灶为MIA的可能性较大,计算机在鉴别肺原位腺癌与微浸润腺癌中具有重要的作用。

In order to explore the value of gray-scale feature values in the identification of mixed lung adenocarcinoma in situ(AIS)and micro-invasive adenocarcinoma(MIA),in this article,the lung parenchyma is extracted first by using the Otsu method and the mask method.Then,the computer calculated the gray feature values(energy,entropy,inverse moment and correlation),mean value and statistical P value in the directions of 0°,45°,90°,and 135°.It is found that the moment of inertia of the lesion in the gray-level co-occurrence matrix can be used as the criterion of the two:the mean values of the moments of inertia of AIS and MIA are 13.8513±0.0001,13.8562±0.0021,respectively.It is concluded that the gray feature value has certain guiding significance for the diagnosis of AIS and MIA.When the moment of inertia value is less than or equal to 13.8513,the lesion is likely to be AIS,and when the moment of inertia value is greater than or equal to 13.8513,the lesion is likely to be MIA.It is more sexual,and the computer plays an important role in distinguishing lung adenocarcinoma in situ from microinvasive adenocarcinoma.