Clinical Report: Radiomic Analysis of Pulmonary Artery Thrombus Predicts Mortality in PE
Overview
Radiomic features extracted from manually segmented pulmonary artery thrombus on CT pulmonary angiograms correlate with 30-day mortality and troponin levels in acute pulmonary embolism (APE) patients. This study demonstrates the potential of radiomics to enhance prognostic assessment beyond conventional imaging markers.
Background
Acute pulmonary embolism is a common cardiovascular emergency with high mortality, necessitating rapid and accurate diagnosis. CT pulmonary angiogram (CTPA) is the standard imaging modality used to detect emboli and assess parameters such as right ventricular enlargement and inferior vena cava reflux, which correlate with clinical outcomes. Radiomics, a quantitative image feature extraction technique, has shown promise in oncology and is emerging as a tool to improve diagnosis and prognosis in other diseases, including APE. Prior studies have linked texture features of emboli on CTPA with patient outcomes, but direct evaluation of radiomic features of the thrombus in relation to mortality and troponin levels has not been extensively explored.
Data Highlights
Parameter
Value
Number of patients
86
Male/Female
58/28
Mean age (years)
64.7 ± 14.8
CT scanners used
Siemens Somatom Definition AS+, Canon Aquilion Prime
Contrast agent volume
60–150 ml
Slice thickness and interval
1 mm
Follow-up duration
30 days
Key Findings
Radiomic features were successfully extracted from 2D and 3D manually segmented thrombus masks on CTPA images.
Significant correlations were found between radiomic features of the thrombus and 30-day mortality in APE patients.
Radiomic features also correlated with troponin levels, a biomarker linked to APE severity and mortality risk.
Binary classification models based on radiomic features could predict mortality and elevated troponin with balanced datasets after over- and undersampling.
Manual segmentation was validated by radiologists, ensuring accuracy of thrombus delineation for radiomic analysis.
Clinical Implications
Radiomic analysis of pulmonary artery thrombus on routine CTPA can provide additional prognostic information beyond traditional imaging markers and clinical scores. Incorporating radiomics into clinical workflows may improve risk stratification and guide management decisions in patients with acute pulmonary embolism. Further validation could enable personalized treatment planning based on quantitative thrombus characteristics.
Conclusion
Radiomic features derived from pulmonary artery thrombus on CTPA are promising biomarkers for predicting short-term mortality and troponin elevation in acute pulmonary embolism. This approach may enhance prognostic accuracy and support personalized patient care.
References
Jakob Leonhardi et al. 2021 -- Texture analysis of pulmonary embolism correlates with blood markers and outcomes
Jennifer Gotta et al. 2020 -- Radiomics for classification and survival prediction in pulmonary embolism
Dawei Wang et al. 2019 -- Prognostic value of texture features in pulmonary embolism
Otto-von-Guericke University Ethics Committee 2021 -- Institutional review board approval Nr. 145/21
