Lung-DDPM+: Efficient thoracic CT image synthesis using diffusion probabilistic model.
Authors
Affiliations (3)
Affiliations (3)
- Centre de recherche du CHU de Québec-Université Laval, 2260 boul. Henri-Bourassa, Québec, G1J 0J9, QC, Canada; Department of Molecular Biology, Medical Biochemistry and Pathology, Université Laval, Ferdinand Vandry Pavillon, 1050 Rue de la Médecine, Québec, G1V 0A6, QC, Canada; Cancer Research Center, Université Laval, 9 Rue McMahon, Québec, G1R 3S3, QC, Canada; Big Data Research Center, Université Laval, Adrien Pouliot Pavilion, 1065 Av. de la Médecine, Québec, G1V 0A6, QC, Canada.
- Department of Radiology, Carver College of Medicine - The University of Iowa, 200 Hawkins Drive, Iowa City, 52242, IA, United States.
- Centre de recherche du CHU de Québec-Université Laval, 2260 boul. Henri-Bourassa, Québec, G1J 0J9, QC, Canada; Department of Molecular Biology, Medical Biochemistry and Pathology, Université Laval, Ferdinand Vandry Pavillon, 1050 Rue de la Médecine, Québec, G1V 0A6, QC, Canada; Cancer Research Center, Université Laval, 9 Rue McMahon, Québec, G1R 3S3, QC, Canada; Big Data Research Center, Université Laval, Adrien Pouliot Pavilion, 1065 Av. de la Médecine, Québec, G1V 0A6, QC, Canada. Electronic address: [email protected].
Abstract
Generative artificial intelligence (AI) has been playing an important role in various domains. Leveraging its high capability to generate high-fidelity and diverse synthetic data, generative AI is widely applied in diagnostic tasks, such as lung cancer diagnosis using computed tomography (CT). However, existing generative models for lung cancer diagnosis suffer from low efficiency and anatomical imprecision, which limit their clinical applicability. To address these drawbacks, we propose Lung-DDPM+, an improved version of our previous model, Lung-DDPM. This novel approach is a denoising diffusion probabilistic model (DDPM) guided by nodule semantic layouts and accelerated by a pulmonary DPM-solver, enabling the method to focus on lesion areas while achieving a better trade-off between sampling efficiency and quality. Evaluation results on the public LIDC-IDRI dataset suggest that the proposed method achieves 8× fewer FLOPs (floating point operations per second), 6.8× lower GPU memory consumption, and 14× faster sampling compared to Lung-DDPM. Moreover, it maintains comparable sample quality to both Lung-DDPM and other state-of-the-art (SOTA) generative models in two downstream segmentation tasks. We also conducted a Visual Turing Test by an experienced radiologist, showing the advanced quality and fidelity of synthetic samples generated by the proposed method. These experimental results demonstrate that Lung-DDPM+ can effectively generate high-quality thoracic CT images with lung nodules, highlighting its potential for broader applications, such as general tumor synthesis and lesion generation in medical imaging. The code and pretrained models are available at https://github.com/Manem-Lab/Lung-DDPM-PLUS.