arXiv subjects

Ting Yin

Publications and source records attributed to Ting Yin.

4 recordsLinked to original sources

arXiv2026

CorePath: A Breast-Specialized Pathology Foundation Model for Core Needle Biopsy Diagnosis and Risk-Controlled Report Generation

Breast core needle biopsy (CNB) is central to breast cancer diagnosis yet remains challenging because limited tissue sampling, lesion heterogeneity, and subtle morphologic overlap can obscure subtype distinctions. We developed CorePath, a breast-specialized multimodal pathology foundation model fine-tuned from PRISM using 7901 paired CNB whole-slide images and diagnostic reports from two centers. Evaluated across six CNB cohorts and two public breast pathology benchmarks without task-specific retraining, CorePath consistently outperformed PRISM across cancer detection, invasion assessment, and histological subtyping. It achieved weighted area under the receiver operating characteristic curves (AUCs) of 0.9526-0.9735 for five-class CNB histological subtyping across private centers. On public benchmarks, CorePath outperformed leading pathology foundation models, achieving the highest weighted AUCs of 0.7780 for BCNB invasive carcinoma subtyping, 0.8178 for BRACS lesion stratification, and 0.8252 for BRACS fine-grained classification. In report generation, CorePath reduced the overall non-breast hallucinations from 30.1% to 2.8%, demonstrating improved domain fidelity after breast-specific adaptation. CorePath-CRG further combined conformal subtype-confidence gating with Learn-Then-Test risk control to enable selective report release, subtype-level fallback, and deferral. CorePath-CRG achieved zero non-breast hallucinations among released outputs and showed the strongest overall performance in pathologist-validated LLM-based Evaluation Scores and quantitative report-generation metrics across most centers. These results demonstrate that domain-specialized foundation models with statistical risk control offer a promising approach for accurate breast CNB diagnosis and reliable report generation.

cs.CV↗

arXiv2025

Denoising Mutual Knowledge Distillation in Bi-Directional Multiple Instance Learning

Multiple Instance Learning is the predominant method for Whole Slide Image classification in digital pathology, enabling the use of slide-level labels to supervise model training. Although MIL eliminates the tedious fine-grained annotation process for supervised learning, whether it can learn accurate bag- and instance-level classifiers remains a question. To address the issue, instance-level classifiers and instance masks were incorporated to ground the prediction on supporting patches. These methods, while practically improving the performance of MIL methods, may potentially introduce noisy labels. We propose to bridge the gap between commonly used MIL and fully supervised learning by augmenting both the bag- and instance-level learning processes with pseudo-label correction capabilities elicited from weak to strong generalization techniques. The proposed algorithm improves the performance of dual-level MIL algorithms on both bag- and instance-level predictions. Experiments on public pathology datasets showcase the advantage of the proposed methods.

cs.CV↗

arXiv2021

PET CMR$_{glc}$ mapping and $^{1}$H MRS show altered glucose uptake and neurometabolic profiles in BDL rats

Type C hepatic encephalopathy (HE) is a complex neuropsychiatric disorder occurring as a consequence of chronic liver disease. Alterations in energy metabolism have been suggested in type C HE, but $\textit{in vivo}$ studies on this matter remain sparse and have reported conflicting results. Here, we propose a novel preclinical $^{18}$F-FDG PET methodology to compute quantitative 3D maps of the regional cerebral metabolic rate of glucose (CMR$_{glc}$) from a labelling steady-state PET image of the brain and an image-derived input function. This quantitative approach shows its strength when comparing groups of animals with divergent physiology, such as HE animals. PET CMR$_{glc}$ maps were registered to an atlas and the mean CMR$_{glc}$ from the hippocampus and the cerebellum were associated to the corresponding localized $^{1}$H MR spectroscopy acquisitions. This study provides for the first time local and quantitative information on both brain glucose uptake and neurometabolic profile alterations in a rat model of type C HE. A 2-fold lower brain glucose uptake, concomitant with an increase in brain glutamine and a decrease in the main osmolytes was observed in the hippocampus and in the cerebellum. These novel findings are an important step towards new insights into energy metabolism in the pathophysiology of HE.

physics.med-ph↗

arXiv2020

An optimized pipeline for functional connectivity analysis in the rat brain

Resting state functional MRI (rs-fMRI) is a widespread and powerful tool for investigating functional connectivity and brain disorders. However, functional connectivity analysis can be seriously affected by random and structured noise from non-neural sources such as physiology. Thus, it is essential to first reduce thermal noise and then correctly identify and remove non-neural artefacts from rs-fMRI signals through optimized data processing methods. However, existing tools that correct for these effects have been developed for human brain and are not readily transposable to rat data. Therefore, the aim of the present study was to establish a data processing pipeline that can robustly remove random and structured noise from rat rs-fMRI data. It includes a novel denoising approach based on the Marchenko-Pastur Principle Component Analysis (MP-PCA) method, FMRIB's ICA-based Xnoiseifier (FIX) for automatic artefact classification and cleaning, and global signal regression. Our results show that: I) MP-PCA denoising substantially improves the temporal signal-to-noise ratio; II) the pre-trained FIX classifier achieves a high accuracy in artefact classification; III) both artefact cleaning and global signal regression are essential steps in minimizing the within-group variability in control animals and identifying functional connectivity changes in a rat model of sporadic Alzheimer's disease, as compared to controls.

q-bio.NC↗