Applications of artificial intelligence in computed tomography imaging for phenotyping pulmonary hypertension

Abstract

Purpose of review: Pulmonary hypertension is a heterogeneous condition with significant morbidity and mortality. Computer tomography (CT) plays a central role in determining the phenotype of pulmonary hypertension, informing treatment strategies. Many artificial intelligence tools have been developed in this modality for the assessment of pulmonary hypertension. This article reviews the latest CT artificial intelligence applications in pulmonary hypertension and related diseases.

Recent findings: Multistructure segmentation tools have been developed in both pulmonary hypertension and nonpulmonary hypertension cohorts using state-of-the-art UNet architecture. These segmentations correspond well with those of trained radiologists, giving clinically valuable metrics in significantly less time. Artificial intelligence lung parenchymal assessment accurately identifies and quantifies lung disease patterns by integrating multiple radiomic techniques such as texture analysis and classification. This gives valuable information on disease burden and prognosis. There are many accurate artificial intelligence tools to detect acute pulmonary embolism. Detection of chronic pulmonary embolism proves more challenging with further research required.

Summary: There are numerous artificial intelligence tools being developed to identify and quantify many clinically relevant parameters in both pulmonary hypertension and related disease cohorts. These potentially provide accurate and efficient clinical information, impacting clinical decision-making.

Box 1

no caption available

FIGURE 1

Computer tomography findings in pulmonary hypertension. (a) Dilatation of the main pulmonary artery. Comparison to the aorta is used to normalize for body size. (b) The pulmonary arteries are assessed for thromboembolic disease and pulmonary obstruction. These provide information on AV malformations, aneurysms, and large vessel vasculitis. (c) Mediastinal structures including the dilated bronchial arteries, dilated oesophagus, lymphadenopathy and pericardial effusions provide evidence of pulmonary hypertension and potential cause. (d) Size and shape of cardiac chambers and myocardial hypertrophy provide evidence of left-sided and right-sided heart failure. Assessment of congenital heart disease and anomalous arterial venous drainage can also be made. (e) Thickening of the right ventricle (RV) outflow tract is suggestive of RV hypertrophy. (f) Other thoracic evidence such as pleural effusion, pericardial effusion and ascites and features of left and right heart failure should be considered. (g) Lung parenchymal assessment provides information as to the presence and severity of lung disease as important factors for differentiation of pulmonary hypertension phenotypes.

FIGURE 2

Cardiac and great vessel segmentation. Segmentation of the cardiac structures and great vessels including ascending and descending aorta (yellow and pink, respectively), pulmonary artery (light blue), right atrium (red), right ventricular chamber (orange) and myocardium (light green), left atrium (dark blue), left ventricular chamber (dark green) and myocardium (purple). For a colour version of this figure, see the online version of this article.

FIGURE 3

Lung disease texture quantification and localization. Patient with interstitial lung disease overlaid with artificial intelligence lung disease texture classifications; normal (dark blue), pure ground-glass (red), ground-glass with reticulation (yellow), honeycombing (light blue) and low attenuation (pink). For a colour version of this figure, see the online version of this article.