Cross-Phase Mutual Learning Framework for Pulmonary Embolism Identification on Non-Contrast CT Scans

TL;DR

Pulmonary embolism identification on non-contrast CT is challenging due to low contrast and limited clinical radiology guidance. The authors introduce CPMN, a Cross-Phase Mutual Learning framework that transfers knowledge from contrast-enhanced CTPA to non-contrast CT to jointly perform PE classification and embolism segmentation, using Inter-Feature Alignment and Intra-Feature Discrepancy to align and sharpen representations. On a large dual-phase ALD-PE dataset and a public benchmark, CPMN achieves 95.4% patient-level sensitivity and 99.6% specificity on NCT, with a segmentation Dice of 78.5%, outperforming radiologists and state-of-the-art baselines and providing CAM visualizations and predicted masks for interpretability. The approach offers a cost-effective, interpretable tool suitable for rapid PE triage when contrast imaging is unavailable, with strong potential for real-world clinical deployment.

Abstract

Pulmonary embolism (PE) is a life-threatening condition where rapid and accurate diagnosis is imperative yet difficult due to predominantly atypical symptomatology. Computed tomography pulmonary angiography (CTPA) is acknowledged as the gold standard imaging tool in clinics, yet it can be contraindicated for emergency department (ED) patients and represents an onerous procedure, thus necessitating PE identification through non-contrast CT (NCT) scans. In this work, we explore the feasibility of applying a deep-learning approach to NCT scans for PE identification. We propose a novel Cross-Phase Mutual learNing framework (CPMN) that fosters knowledge transfer from CTPA to NCT, while concurrently conducting embolism segmentation and abnormality classification in a multi-task manner. The proposed CPMN leverages the Inter-Feature Alignment (IFA) strategy that enhances spatial contiguity and mutual learning between the dual-pathway network, while the Intra-Feature Discrepancy (IFD) strategy can facilitate precise segmentation of PE against complex backgrounds for single-pathway networks. For a comprehensive assessment of the proposed approach, a large-scale dual-phase dataset containing 334 PE patients and 1,105 normal subjects has been established. Experimental results demonstrate that CPMN achieves the leading identification performance, which is 95.4\% and 99.6\% in patient-level sensitivity and specificity on NCT scans, indicating the potential of our approach as an economical, accessible, and precise tool for PE identification in clinical practice.

Figures (2)

• Figure 1: Overview of our proposed Cross-Phase Mutual learNing framework (CPMN) that contains the CTPA-pathway network ($\mathbf{\Omega_{\text{1}}}$) and the NCT-pathway network ($\mathbf{\Omega_{\text{2}}}$). Each pathway network comprises an encoder-decoder pair ($\Phi_1/\Psi_1$, $\Phi_2/\Psi_2$) that extracts features from the corresponding volume. The presented Inter-Feature Alignment (IFA) strategy through an affinity graph captures pair-wise spatial feature similarities in the encoder. The predicted PE probabilities ($p_1$, $p_2$) are harmonized using KL divergence to align feature distributions without altering the CTPA-pathway network. The dense center loss is designed to refine the segmentation feature space ($\Sigma_1, \Sigma_2$).
• Figure 2: (a) ROC curve for our model versus three radiologists on the hold-out test set (n = 289) for binary classification. (b) Visualization example in the test set. This PE case is miss-detected by three radiologists but our model succeeds in locating the embolism by CAM zhou2016learning and predicted mask. Green contours represent the regions of embolism (best viewed in color).