DiffSegLung: Diffusion Radiomic Distillation for Unsupervised Lung Pathology Segmentation
For medical imaging researchers, this work addresses the challenge of unsupervised multi-pathology segmentation in CT by leveraging physics-grounded radiomic features without annotations.
DiffSegLung introduces a diffusion radiomic distillation framework that uses handcrafted radiomic descriptors as a teacher to guide unsupervised segmentation of lung pathologies in CT, achieving improved segmentation across four pathology classes over unsupervised baselines and better generation fidelity than prior CT diffusion models.
Unsupervised segmentation of pulmonary pathologies in CT remains an open challenge due to the absence of annotated multi pathology cohorts and the failure of existing diffusion-based methods to exploit the quantitative Hounsfield Unit (HU) signal that physically distinguishes tissue classes. To address this, we propose DiffSegLung,a framework that introduces Diffusion Radiomic Distillation, in which handcrafted radiomic descriptors serve as a physics grounded teacher to shape the bottleneck of a 3D diffusion U-Net via a contrastive objective, transferring pathology discriminative structure into the learned representation without any annotations. At inference, the teacher is discarded and multitimestep bottleneck features are clustered by a Gaussian Mixture Model with HU-guided label assignment, followed by Sobel Diffusion Fusion for boundary refinement. Evaluated on 190 expert annotated axial slices drawn from four heterogeneous CT cohorts, Diff-SegLung improves segmentation across all four pathology classes over unsupervised baselines and improves generation fidelity over prior CT diffusion models.