Radiomic features derived from medical images and segmentation masks are used to support decision making in clinical imaging pipelines. In practice, these features are often computed from predicted masks, but segmentation models can be overconfident or poorly calibrated, making derived measurements appear more reliable than they are. Conformal prediction (CP) provides distribution-free prediction intervals with finite-sample marginal coverage guarantees, but black-box intervals for segmentation-derived radiomics can be inefficient because they ignore test-time information about image appearance, mask geometry, and segmentation uncertainty. We propose ConRad, a conformal framework for scalar radiomic targets that uses covariates derived from the predicted mask, input image, predicted radiomics, and boundary uncertainty to construct adaptive intervals while maintaining coverage. Across five 2D medical imaging datasets and 171 retained radiomic targets, we show that ConRad improves feature-level efficiency compared to baselines while maintaining near-nominal empirical coverage. Ablation results further indicate that segmentation boundary uncertainty features are the largest contributors to interval efficiency.
This paper provides a rigorous, self-contained investigation of consensus segmentation. We derive the mathematical foundations from first principles -- the generative model, EM algorithm, Van Leemput's marginalization analysis, identifiability conditions, Spatial STAPLE, and deep…
Magnetic particle imaging (MPI) enables real-time, radiation-free tracking of magnetic nanoparticle-coated instruments, making it highly suitable for interventional procedures. This study proposes a harmonic-aware transformer framework that directly predicts catheter tip position…
Despite exposing rich intermediate representations, Vision Transformers (ViTs) are almost exclusively utilized as black-box feature extractors, where only the last layer is considered for downstream tasks. We challenge this convention by introducing the notion of recoverability:…
Arrhythmogenic left ventricular cardiomyopathy is a genetic myocardial disease difficult to diagnose due to the lack of gold standard criteria. Simultaneous PET/MR imaging, combined with multiparametric quantitative analysis, could facilitate the identification of different profi…
A popular route to interpretable zero-shot classification asks a large language model (LLM) to describe each class name and prompts CLIP with the resulting descriptors. We show that these descriptors carry little visual evidence of their own: removing the class name from the prom…
Where should the question go in a vision-language model (VLM) prompt: before the image or after it? Intuition says before: knowing what is asked should tell the model where to look. Yet across visual question answering benchmarks, question-first prompting consistently underperfor…