Lukhumaidze Leila, Hogg James C, Bourbeau Jean, Tan Wan C, Kirby Miranda
Toronto Metropolitan University, Toronto, ON, Canada (L.L., M.K.).
Center for Heart, Lung Innovation, University of British Columbia, Vancouver, BC, Canada (J.C.H., W.C.T.).
Acad Radiol. 2025 Jan;32(1):543-555. doi: 10.1016/j.acra.2024.08.030. Epub 2024 Aug 26.
The structural lung features that characterize individuals with preserved ratio impaired spirometry (PRISm) that remain stable overtime are unknown. The objective of this study was to use machine learning models with computed tomography (CT) imaging to classify stable PRISm from stable controls and stable COPD and identify discriminative features.
A total of 596 participants that did not transition between control, PRISm and COPD groups at baseline and 3-year follow-up were evaluated: n = 274 with normal lung function (stable control), n = 22 stable PRISm, and n = 300 stable COPD. Investigated features included: quantitative CT (QCT) features (n = 34), such as total lung volume (%TLC) and percentage of ground glass and reticulation (%GG+Reticulation), as well as Radiomic (n = 102) features, including varied intensity zone distribution grainy texture (GLDZM). Logistic regression machine learning models were trained using various feature combinations (Base, Base+QCT, Base+Radiomic, Base+QCT+Radiomic). Model performances were evaluated using area under receiver operator curve (AUC) and comparisons between models were made using DeLong test; feature importance was ranked using Shapley Additive Explanations values.
Machine learning models for all feature combinations achieved AUCs between 0.63-0.84 for stable PRISm vs. stable control, and 0.65-0.92 for stable PRISm vs. stable COPD classification. Models incorporating imaging features outperformed those trained solely on base features (p < 0.05). Compared to stable control and COPD, those with stable PRISm exhibited decreased %TLC and increased %GG+Reticulation and GLDZM.
These findings suggest that reduced lung volumes, and elevated high-density and ground glass/reticulation patterns on CT imaging are associated with stable PRISm.
肺活量比值降低但肺功能正常（PRISm）且长期保持稳定的个体所具有的肺部结构特征尚不清楚。本研究的目的是使用机器学习模型和计算机断层扫描（CT）成像，将稳定的PRISm与稳定的对照者和稳定的慢性阻塞性肺疾病（COPD）患者进行分类，并识别具有鉴别性的特征。
对596名在基线和3年随访期间未在对照组、PRISm组和COPD组之间转换的参与者进行了评估：n = 274名肺功能正常者（稳定对照组），n = 22名稳定的PRISm患者，以及n = 300名稳定的COPD患者。研究的特征包括：定量CT（QCT）特征（n = 34），如肺总量（%TLC）以及磨玻璃影和网状影百分比（%GG + 网状影），以及影像组学特征（n = 102），包括不同强度区域分布颗粒纹理（GLDZM）。使用各种特征组合（基础特征、基础特征 + QCT、基础特征 + 影像组学、基础特征 + QCT + 影像组学）训练逻辑回归机器学习模型。使用受试者工作特征曲线下面积（AUC）评估模型性能，并使用德龙检验对模型进行比较；使用夏普利值对特征重要性进行排名。
对于稳定的PRISm与稳定对照组的分类，所有特征组合的机器学习模型的AUC在0.63 - 0.84之间，对于稳定的PRISm与稳定的COPD患者的分类，AUC在0.65 - 0.92之间。纳入影像特征的模型优于仅基于基础特征训练的模型（p < 0.05）。与稳定对照组和COPD患者相比，稳定的PRISm患者表现出%TLC降低，%GG + 网状影和GLDZM增加。
这些发现表明，肺容积减小以及CT成像上高密度和磨玻璃影/网状影模式升高与稳定的PRISm相关。
