Back to all papers

Multimodal Contrast-Free Pulmonary Perfusion Imaging by Integrating CT and MRI for Enhanced Lung Function Assessment.

July 30, 2026pubmed logopapers

Authors

Hu D,Li B,Li H,Ge H,Xiong T,Chen Z,Huang Y,Li T,Ren G,Cai J

Affiliations (5)

  • Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong SAR, China. Electronic address: [email protected].
  • Department of Radiation Oncology, The Affiliated Cancer Hospital of Zhengzhou University and Henan Cancer Hospital, Zhengzhou, China.
  • Department of Radiation Oncology, The Affiliated Cancer Hospital of Zhengzhou University and Henan Cancer Hospital, Zhengzhou, China. Electronic address: [email protected].
  • Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong SAR, China.
  • Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hong Kong SAR, China. Electronic address: [email protected].

Abstract

Anatomy image-driven lung function imaging methods have been explored for thoracic radiotherapy, but most contrast-free approaches rely on unimodal surrogates. This study aimed to develop a multimodal contrast-free pulmonary perfusion reconstruction framework (MCF-Q) that integrates computed tomography (CT) and magnetic resonance imaging (MRI) to leverage complementary anatomical and functional information from routinely acquired CT and non-contrast MRI, improve agreement with single-photon emission computed tomography perfusion (SPECT-Q), and explore its potential to support functional lung avoidance radiotherapy (FLART). This prospective analysis included 21 patients with lung cancer who underwent pulmonary SPECT-Q, CT, and <sup>1</sup>H MRI. MCF-Q adopted a dual-branch deep learning architecture to extract complementary features from CT and MRI and fuse them into pulmonary perfusion maps. Seven-fold cross-validation was performed to evaluate voxel-wise and function-wise agreement between MCF-Q and SPECT-Q, including Spearman's correlation coefficient (R) and the Dice similarity coefficient (DSC). The dosimetric analysis was also conducted by comparing a conventional radiotherapy (ConvRT) plan with FLART plans guided by different perfusion maps. For voxel-wise assessment, the MCF-Q achieved an R value of 0.7831 ± 0.0821. For function-wise similarity, the MCF-Q gained the DSC value of 0.8396 ± 0.0379 in high-functional regions, and 0.7680 ± 0.0555 in low-functional regions. All metrics calculated from MCF-Q showed significant improvement over single-modality-based lung function imaging methods. In dosimetric performance, the MCF-Q-guided FLART achieved better dose sparing in high-functional regions, while maintaining comparable whole-lung and organ-at-risk dose metrics. In this study, the proposed MCF-Q demonstrated the feasibility of multimodal perfusion reconstruction from CT and MRI, with improved agreement with SPECT-Q, and provided radiotherapy-planning-relevant functional information that may facilitate functional lung avoidance strategies. These findings support the value of integrating routinely acquired CT and MRI for contrast-free, planning-relevant perfusion estimation, warranting validation in larger cohorts.

Topics

Journal Article

Ready to Sharpen Your Edge?

Subscribe to join 11k+ peers who rely on RadAI Slice. Get the essential weekly briefing that empowers you to navigate the future of radiology.

We respect your privacy. Unsubscribe at any time.