Diffuse Until You Fake It—Synthesizing High-Fidelity Chest CT Volumes from the LIDC-IDRI Dataset Using Diffusion Models

• Roney Nogueira de Sousa IFCE
• Saulo Anderson Freitas Oliveira IFCE

Resumo

We propose a diffusion-based generative framework for synthesizing realistic thoracic CT volumes from the LIDC–IDRI dataset, aiming to address data scarcity, privacy concerns, and augmentation needs in medical imaging. Our approach leverages a 3D U-Net architecture with residual connections as the denoising function within a Denoising Diffusion Probabilistic Model, enabling volumetric reconstruction from isotropic Gaussian noise across 1,000 reverse diffusion steps. The model is trained end-to-end on standardized CT scans resampled to 1283 voxels, with intensities normalized in the Hounsfield scale. Due to the high computational demands of volumetric diffusion, the training was distributed across two consumer-grade GPUs over 58 days, incorporating memory-efficient strategies such as gradient checkpointing and small batch sizes. Evaluation is conducted using Fréchet Inception Distance (FID) and Multi-Scale Structural Similarity Index (MS-SSIM), with results computed against a held-out test set. Qualitative inspection and quantitative metrics jointly demonstrate that the generated samples exhibit high anatomical plausibility, cross-slice coherence, and distributional alignment with real CT scans. These findings highlight the potential of diffusion models to surpass GAN-based alternatives in generating clinically meaningful synthetic 3D medical images.

Referências

H.-C. Shin, N. A. Tenenholtz, J. K. Rogers, T. Schwarz, M. L. Senjem, J. L. Gunter, K. P. Andriole, and M. Michalski, “Medical image synthesis for data augmentation and anonymization using generative adversarial networks,” in Simulation and Synthesis in Medical Imaging. Springer International Publishing, 2018, pp. 1–11.

V. Sandfort, K. Yan, P. J. Pickhardt, and R. M. Summers, “Data augmentation using generative adversarial networks (cyclegan) to improve generalizability in ct segmentation tasks,” Scientific Reports, vol. 9, no. 1, Nov. 2019. [Online]. DOI: 10.1038/s41598-019-52737-x

D. Nie, R. Trullo, J. Lian, C. Petitjean, S. Ruan, Q. Wang, and D. Shen, Medical Image Synthesis with Context-Aware Generative Adversarial Networks. Springer International Publishing, 2017, p. 417–425. [Online]. DOI: 10.1007/978-3-319-66179-7_48

J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 33, 2020, pp. 6840–6851.

A. Kazerouni, E. K. Aghdam, M. Heidari, R. Azad, M. Fayyaz, I. Hacihaliloglu, and D. Merhof, “Diffusion models in medical imaging: A comprehensive survey,” Medical Image Analysis, vol. 88, p. 102846, Aug. 2023. [Online]. DOI: 10.1016/j.media.2023.102846

Z. Yang, Z. Chen, Y. Sun, A. Strittmatter, A. Raj, A. Allababidi, J. S. Rink, and F. G. Zöllner, “seg2med: a bridge from artificial anatomy to multimodal medical images,” 2025. [Online]. Available: [link]

S. G. I. Armato, G. McLennan, L. Bidaut, M. F. McNitt-Gray, C. R. Meyer, A. P. Reeves, B. Zhao, D. R. Aberle, C. I. Henschke, E. A. Hoffman, E. A. Kazerooni, H. MacMahon, E. J. Van Beek, D. Yankelevitz, A. M. Biancardi, P. H. Bland, M. S. Brown, R. M. Engelmann, G. E. Laderach, D. Max, R. C. Pais, D. P.-Y. Qing, R. Y. Roberts, A. R. Smith, A. Starkey, P. Batra, P. Caligiuri, A. Farooqi, G. W. Gladish, C. M. Jude, R. F. Munden, I. Petkovska, L. E. Quint, L. H. Schwartz, B. Sundaram, L. E. Dodd, C. Fenimore, D. Gur, N. Petrick, J. Freymann, J. Kirby, B. Hughes, A. Vande Casteele, S. Gupte, M. Sallam, M. D. Heath, M. H. Kuhn, E. Dharaiya, R. Burns, D. S. Fryd, M. Salganicoff, V. Anand, U. Shreter, S. Vastagh, B. Y. Croft, and L. P. Clarke, “The lung image database consortium (lidc) and image database resource initiative (idri): A completed reference database of lung nodules on ct scans,” Medical Physics, vol. 38, no. 2, pp. 915–931, 2011.

A. M. P. Ferreira, “3d lung computed tomography synthesis using generative adversarial networks,” Master’s thesis, Faculdade de Ciências da Universidade do Porto, Porto, Portugal, 2021.

S. Hong, R. Marinescu, A. V. Dalca, A. K. Bonkhoff, M. Bretzner, N. S. Rost, and P. Golland, 3D-StyleGAN: A Style-Based Generative Adversarial Network for Generative Modeling of Three-Dimensional Medical Images. Springer International Publishing, 2021, p. 24–34. [Online]. DOI: 10.1007/978-3-030-88210-5_3

F. Khader, G. Müller-Franzes, S. Tayebi Arasteh, T. Han, C. Haarburger, M. Schulze-Hagen, P. Schad, S. Engelhardt, B. Baeßler, S. Foersch, J. Stegmaier, C. Kuhl, S. Nebelung, J. N. Kather, and D. Truhn, “Denoising diffusion probabilistic models for 3d medical image generation,” Scientific Reports, vol. 13, no. 1, May 2023. [Online]. DOI: 10.1038/s41598-023-34341-2

R. MaríMolas, P. Subías-Beltrán, C. Pitarch Abaigar, M. Galofré Cardo, and R. Redondo Tejedor, Characterization of Synthetic Lung Nodules in Conditional Latent Diffusion of Chest CT Scans. IOS Press, Sep. 2024. [Online]. DOI: 10.3233/FAIA240408

A. Brys, “dicom2nifti: Python library for converting dicom files to nifti,” [link], 2018.

A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan et al., “Pytorch: An imperative style, high-performance deep learning library,” in Advances in Neural Information Processing Systems, vol. 32, 2019, pp. 8026–8037.

F. Bieder, J. Wolleb, A. Durrer, R. Sandkuehler, and P. C. Cattin, “Memory-efficient 3d denoising diffusion models for medical image processing,” in Medical Imaging with Deep Learning. PMLR, 2024, pp. 552–567.

K. D. B. J. Adam et al., “A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, vol. 1412, no. 6, 2014.

M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler, and S. Hochreiter, “Gans trained by a two time-scale update rule converge to a local nash equilibrium,” 2018. [Online]. Available: [link]

Z. Wang, A. Bovik, H. Sheikh, and E. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE Transactions on Image Processing, vol. 13, no. 4, p. 600–612, Apr. 2004. [Online]. DOI: 10.1109/TIP.2003.819861

S. Chen, K. Ma, and Y. Zheng, “Med3d: Transfer learning for 3d medical image analysis,” CoRR, vol. abs/1904.00625, 2019. [Online]. Available: [link]

S. Dayarathna, K. T. Islam, S. Uribe, G. Yang, M. Hayat, and Z. Chen, “Deep learning based synthesis of mri, ct and pet: Review and analysis,” Medical Image Analysis, vol. 92, p. 103046, Feb. 2024. [Online]. DOI: 10.1016/j.media.2023.103046

S. U. H. Dar, A. Ghanaat, J. Kahmann, I. Ayx, T. Papavassiliu, S. O. Schoenberg, and S. Engelhardt, “Investigating data memorization in 3d latent diffusion models for medical image synthesis,” 2023. [Online]. Available: [link]

V. Mudeng, M. Kim, and S.-w. Choe, “Prospects of structural similarity index for medical image analysis,” Applied Sciences, vol. 12, no. 8, p. 3754, Apr. 2022. [Online]. DOI: 10.3390/app12083754

H. Ali, S. Murad, and Z. Shah, Spot the Fake Lungs: Generating Synthetic Medical Images Using Neural Diffusion Models. Springer Nature Switzerland, 2023, p. 32–39. [Online]. DOI: 10.1007/978-3-031-26438-2_3