Carvalho Alysson R S, Guimarães Alan, Basilio Rodrigo, Conrado da Silva Marco A, Colli Sandro, Galhós de Aguiar Carolina, Pereira Rafael C, Lisboa Liseane G, Hochhegger Bruno, Rodrigues Rosana S
Department of Radiology and Imaging Diagnosis, Hospital Universitário Polydoro Ernani de São Thiago, Universidade Federal de Santa Catarina, Florianópolis.
D'Or Institute for Research and Education.
J Thorac Imaging. 2025 Mar 1;40(2):e0804. doi: 10.1097/RTI.0000000000000804.
To compare texture-based analysis using convolutional neural networks (CNNs) against lung densitometry in detecting chest computed tomography (CT) image abnormalities.
A U-NET was used for lung segmentation, and an ensemble of 7 CNN architectures was trained for the classification of low-attenuation areas (LAAs; emphysema, cysts), normal-attenuation areas (NAAs; normal parenchyma), and high-attenuation areas (HAAs; ground-glass opacities, crazy paving/linear opacity, consolidation). Lung densitometry also computes (LAAs, ≤-950 HU), NAAs (-949 to -700 HU), and HAAs (-699 to -250 HU). CNN-based and densitometry-based severity indices (CNN and Dens, respectively) were calculated as (LAA+HAA)/(LAA+NAA+HAA) in 812 CT scans from 176 normal subjects, 343 patients with emphysema, and 293 patients with interstitial lung disease (ILD). The correlation between CNN-derived and densitometry-derived indices was analyzed, alongside a comparison of severity indices among patient subgroups with emphysema and ILD, using the Spearman correlation and ANOVA with Bonferroni correction.
CNN-derived and densitometry-derived severity indices (SIs) showed a strong correlation (ρ=0.90) and increased with disease severity. CNN-SIs differed from densitometry SIs, being lower for emphysema and higher for moderate to severe ILD cases. CNN estimations for normal attenuation areas were higher than those from densitometry across all groups, indicating a potential for more accurate characterization of lung abnormalities.
CNN outputs align closely with densitometry in assessing lung abnormalities on CT scans, offering improved estimates of normal areas and better distinguishing similar abnormalities. However, this requires higher computing power.
比较使用卷积神经网络（CNN）基于纹理的分析与肺密度测定法在检测胸部计算机断层扫描（CT）图像异常中的效果。
使用U-Net进行肺部分割，并训练了7种CNN架构的集成模型，用于对低衰减区域（LAA；肺气肿、囊肿）、正常衰减区域（NAA；正常实质）和高衰减区域（HAA；磨玻璃影、铺路石样/线状影、实变）进行分类。肺密度测定法还计算（LAA，≤-950 HU）、NAA（-949至-700 HU）和HAA（-699至-250 HU）。在来自176名正常受试者、343名肺气肿患者和293名间质性肺疾病（ILD）患者的812例CT扫描中，基于CNN和基于密度测定法的严重程度指数（分别为CNN和Dens）计算为（LAA+HAA）/（LAA+NAA+HAA）。使用Spearman相关性分析和带有Bonferroni校正的方差分析，分析了基于CNN和基于密度测定法得出的指数之间的相关性，同时比较了肺气肿和ILD患者亚组之间的严重程度指数。
基于CNN和基于密度测定法得出的严重程度指数（SI）显示出很强的相关性（ρ=0.90），并且随着疾病严重程度的增加而增加。CNN-SI与密度测定法SI不同，肺气肿患者的该指数较低，中度至重度ILD病例的该指数较高。在所有组中，CNN对正常衰减区域的估计高于密度测定法的估计，表明在更准确地描述肺部异常方面具有潜力。
CNN输出在评估CT扫描上的肺部异常时与密度测定法密切相关，能提供对正常区域的改进估计，并能更好地区分相似的异常。然而，这需要更高的计算能力。
