Searcharxiv⌕ Search

arXiv · 2610.03733

Rare Disease Classification in Video Capsule Endoscopy

Abstract

Video capsule endoscopy offers a non-invasive way to examine the small bowel, however, the rare but clinically important findings that appear only in less than 1\% of labelled frames, provide very less signal for detection. As a result, while models can achieve high overall performance, they almost completely miss these rare findings. We condition each frame's decision on evidence from the rest of that patient's video, not on the frame's own posterior alone, supplying information a single frozen frame can never contain. Building upon Expected Gain Arbitration, a per-frame, macro-F1-optimal correction over the confusion matrix, we further compose it with video-disease-presence conditional resulting in our decoding method Hierarchical Conditional-Expected-gain (HCE). We evaluate HCE on two independent VCE datasets, Kvasir-Capsule and GALAR, against a static temporal-smoothing baseline and an optimal-transport (OT) frame-level correction. HCE significantly improves rare-class recall over a temporal-smoothing baseline on GALAR (2.7--3.1x; 95\% CI excludes zero in both scopes) and shows a similar, though not statistically confirmed, gain on Kvasir-Capsule (up to 3.4x). Against OT, HCE still improves rare-class recall in all four evaluation settings (+0.016 to +0.029), and additionally improves macro-F1 and MCC.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Amit Kumar Patel, Debesh Jha. 2026-09-14. Rare Disease Classification in Video Capsule Endoscopy. https://searcharxiv.com/abs/2610.03733

Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.

KEEP EXPLORING

Related papers

An Octahedral Fibrous Constitutive Model for Heart Valve Mechanics and Function

Fibrous soft tissues derive their nonlinear mechanical response from networks of extracellular matrix fibers. The organization of fibers and their load-bearing characteristics give rise to strain stiffening, the reverse Poynting effect, and anisotropic mechanical behavior. Motivated by these coupled features, we develop an anisotropic hyperelastic model for fibrous biological tissues that accounts for the contribution of the fiber network under both tensile and compressive deformation in an asymmetric fashion. The model uses an octahedral arrangement of fibers embedded in a ground matrix, allowing anisotropy, fiber volume fraction, and pre-stretch to influence the material response. We calibrate the model to experimental data for mitral valve leaflets using an inverse finite element approach that is coupled to automatic differentiation to facilitate efficient parameter calibration. The calibrated model reproduces responses across biaxial, pure-shear, and constrained-uniaxial tests, while separate numerical tests demonstrate its ability to represent substantial volume reduction and the reverse Poynting effect. Using the calibrated model, we investigate how anisotropy and fiber reorientation affect valve deformation under physiological loading. In the idealized valve, the reference fiber orientation, which provided greater radial compliance, yielded complete leaflet coaptation. Prescribed local fiber reorientation redistributed leaflet stress but did not compromise coaptation. Swapping the preferred material directions throughout the leaflets, by contrast, prevented closure in the simulated valve. Finally, we examine how fiber pre-stretch and collagen volume fraction affect leaflet strain and contact, finding a stronger response to changes in pre-stretch over the parameter ranges studied.

q-bio.TO↗

Arginine metabolism drives tumor immune escape by shaping an immunosuppressive microenvironment

The tumor microenvironment evolves through the ``elimination-equilibrium-escape'' paradigm, yet the quantitative mechanisms by which metabolic reprogramming contributes to immune evasion remain unclear. We develop a mathematical model of tumor-immune interactions to investigate how arginine metabolism shapes an immunosuppressive microenvironment. We establish the positivity and boundedness of solutions and analyze the existence and stability of equilibrium states. The model is calibrated using tumor growth data from three mouse cohorts and reproduces tumor growth patterns under distinct immune conditions. The calibrated model predicts faster tumor growth in immunocompetent mice than in immunodeficient controls, with macrophage dynamics accounting for this difference. It further predicts that arginine administration promotes tumor progression by enhancing macrophage polarization and weakening T cell-mediated tumor control. To account for interindividual heterogeneity, we develop the TVInfABC algorithm to generate a virtual mouse cohort. Moreover, our analysis indicates that tumor-associated macrophages can promote transitions from immunologically ``hot'' to ``cold'' states. Systematic pharmacodynamic simulations further uncover a synergy between arginine depletion (anti-ARG1) and immune checkpoint blockade (anti-PD-1/PD-L1). These results provide a mathematical basis for studying metabolic-immune interactions and for evaluating mechanism-based combination therapies.

q-bio.TO↗

Clinical predictions using artificial intelligence in neurosurgery and its subspecialties: a systematic review and meta-analysis

Artificial intelligence models for clinical prediction are proliferating across neurosurgery, but their aggregate discriminative performance, distribution along the care pathway and methodological quality are unclear. We systematically reviewed and meta-analysed these studies in neurosurgery and its subspecialities along with their risk of bias appraisal. We performed a PRISMA search for peer-reviewed original studies using AI to predict a clinical outcome in neurosurgical populations. Metrics were pooled using random effects modelling. Each study was classified as diagnostic, prognostic or forecasting and was assigned an evidence level. Clinical prediction task sub-stratification analysis was performed. Risk-of-bias was assessed using PROBAST. 1063 studies were included. Across 451 studies reporting AUC, pooled discrimination was 0.86, ranging from 0.84 (spine) to 0.87 (neuro-oncology), with extreme heterogeneity. Performance varied by clinical pathway stage, highest for diagnosis, grading and molecular tasks and lowest for post-operative tasks. Only 51 studies were using prospective data with external validation, whereas 619 rested on internal validation alone or no validation. 89 percent studies were at high risk-of-bias, driven by the analysis domain. AI models report high performance across all the neurosurgical subspecialities, concordant across AUC, accuracy, sensitivity and specificity and robust to risk-of-bias stratification. Performance is highest for imaging-anchored diagnostic tasks and lowest at decision-facing post-operative stage. However, the literature is dominated by optimism-prone, internally-validated and poorly calibrated models with extreme heterogeneity. External validation, calibration analysis and reporting guidelines adherence remain prerequisites for clinical adoption.

q-bio.TO↗