Deep Learning-Based Algorithm for Automatic Detection of Pulmonary Embolism in Chest CT Angiograms

Abstract

Purpose: Since the prompt recognition of acute pulmonary embolism (PE) and the immediate initiation of treatment can significantly reduce the risk of death, we developed a deep learning (DL)-based application aimed to automatically detect PEs on chest computed tomography angiograms (CTAs) and alert radiologists for an urgent interpretation. Convolutional neural networks (CNNs) were used to design the application. The associated algorithm used a hybrid 3D/2D UNet topology. The training phase was performed on datasets adequately distributed in terms of vendors, patient age, slice thickness, and kVp. The objective of this study was to validate the performance of the algorithm in detecting suspected PEs on CTAs.

Methods: The validation dataset included 387 anonymized real-world chest CTAs from multiple clinical sites (228 U.S. cities). The data were acquired on 41 different scanner models from five different scanner makers. The ground truth (presence or absence of PE on CTA images) was established by three independent U.S. board-certified radiologists.

Results: The algorithm correctly identified 170 of 186 exams positive for PE (sensitivity 91.4% [95% CI: 86.4-95.0%]) and 184 of 201 exams negative for PE (specificity 91.5% [95% CI: 86.8-95.0%]), leading to an accuracy of 91.5%. False negative cases were either chronic PEs or PEs at the limit of subsegmental arteries and close to partial volume effect artifacts. Most of the false positive findings were due to contrast agent-related fluid artifacts, pulmonary veins, and lymph nodes.

Conclusions: The DL-based algorithm has a high degree of diagnostic accuracy with balanced sensitivity and specificity for the detection of PE on CTAs.

Keywords: artificial intelligence; chest CT; computed tomography angiography; deep learning tool; pulmonary embolism.

Figure 1

Receiver-operating characteristic (ROC) curve for CINA-PE.

Figure 2

Axial CTA image. Example of segmental pulmonary embolus (PE) in the right lower lobe (filling defect within the arterial lumen), correctly detected by the CINA-PE software considered as a true positive case. Top (a): The red arrow shows the clot within the arterial lumen. Bottom (b): Axial image targeted on the right lower lobe.

Figure 3

Case of the false negative response of the algorithm in the left upper lobe. Top (a): Axial image showing excellent opacification of pulmonary arteries. The red arrow shows the small filling defect within the termination of the segmental artery just close to the origin of subsegmental arteries. Note the presence of bilateral pleural effusion. Bottom (b): Axial image targeted on the left lung illustrating the presence of PE.

Figure 4

Case of the false positive response of the algorithm in the right lower lobe. Top (a): Axial image at the level of lung bases. The red arrow shows the pseudo-lesion. Bottom (b): Subsequent axial image showing motion artifacts blurring the pseudo-lesion.