arXiv subjects

Guanghua Xiao

Publications and source records attributed to Guanghua Xiao.

17 recordsLinked to original sources

arXiv2026

GlaKG: A Biomarker-Centric Fundus Knowledge Graph for Explainable Glaucoma Diagnosis and Risk Assessment

Glaucoma is a leading cause of irreversible blindness worldwide, yet most automated diagnosis systems rely on opaque deep-learning models that offer little clinical interpretability. We present GlaKG, a biomarker-centric fundus knowledge graph that integrates structural biomarkers, clinically grounded rules, and image features to produce traceable reasoning for glaucoma diagnosis and risk stratification. GlaKG encodes six entity types (Fundus Image, Optic Disc, Neural Rim, Pathology, Diagnosis, Risk Level), eight relation types, and 11 clinically validated rules into a unified graph, so that every prediction is accompanied by an explicit reasoning chain linking biomarker evidence to activated clinical rules. To keep knowledge-based reasoning strictly separate from label information, we adopt a post-processing fusion framework that combines ResNet50 image embeddings with a normalized KG reasoning-chain score via a tunable weight alpha, with all fitting confined to the training split. On a publicly available, AI-annotated fundus dataset, GlaKG reaches F1 = 0.9953 for binary glaucoma classification and 0.930 accuracy with 0.922 weighted F1 for four-class risk stratification; we report openly that the dataset's biomarker annotations are highly label-correlated, and therefore frame these figures as an upper bound attainable with clean structured biomarkers rather than as leakage-free image-only performance. Feature-importance analysis shows KG-derived and biomarker features contributing near-equally (51.1% vs. 48.9%), and the reasoning chain flags borderline cases by exposing low chain scores rather than failing silently. GlaKG's central contribution is therefore a clinically auditable reasoning framework that complements raw predictive performance by explicitly exposing the biomarker evidence and rule activations behind each decision.

cs.CV↗

arXiv2026

Deep reflective reasoning in interdependence constrained structured data extraction from clinical notes for digital health

Extracting structured information from clinical notes requires navigating a dense web of interdependent variables where the value of one attribute logically constrains others. Existing Large Language Model (LLM)-based extraction pipelines often struggle to capture these dependencies, leading to clinically inconsistent outputs. We propose deep reflective reasoning, a large language model agent framework that iteratively self-critiques and revises structured outputs by checking consistency among variables, the input text, and retrieved domain knowledge, stopping when outputs converge. We extensively evaluate the proposed method in three diverse oncology applications: (1) On colorectal cancer synoptic reporting from gross descriptions (n=217), reflective reasoning improved average F1 across eight categorical synoptic variables from 0.828 to 0.911 and increased mean correct rate across four numeric variables from 0.806 to 0.895; (2) On Ewing sarcoma CD99 immunostaining pattern identification (n=200), the accuracy improved from 0.870 to 0.927; (3) On lung cancer tumor staging (n=100), tumor stage accuracy improved from 0.680 to 0.833 (pT: 0.842 -> 0.884; pN: 0.885 -> 0.948). The results demonstrate that deep reflective reasoning can systematically improve the reliability of LLM-based structured data extraction under interdependence constraints, enabling more consistent machine-operable clinical datasets and facilitating knowledge discovery with machine learning and data science towards digital health.

cs.AI↗

arXiv2026

Spatial Analysis for AI-segmented Histopathology Images: Methods and Implementation

Quantitative characterization of cellular spatial organization is critical for understanding tumor progression and immune response. Recent advances in artificial intelligence (AI) enable large-scale segmentation and classification of nuclei from digitized histopathology slides, producing massive point pattern and marked point pattern data. However, accessible and standardized tools for downstream spatial statistical analysis remain limited. We present SASHIMI (Spatial Analysis for Segmented Histopathology Images using Machine Intelligence), a browser-based platform for real-time spatial analysis of AI-segmented histopathology images. Rather than proposing new spatial methods, SASHIMI systematically organizes and operationalizes 27 widely used spatial summary statistics, areal indices, and topological features within a unified computational framework. The platform computes mathematically grounded descriptors including K-, L-, G-, F-, and J-functions, pair correlation and mark connection functions, spatial autocorrelation measures, similarity indices, and persistent homology-based topological summaries. Outputs include both functional curves and scalar feature tables suitable for downstream statistical modeling. We illustrate the framework using two cancer cohorts: oral potentially malignant disorders and non-small-cell lung cancer. Across datasets, cross-type spatial interactions and topological descriptors show associations with patient survival, demonstrating that complementary spatial features capture distinct aspects of tumor microenvironment architecture. SASHIMI provides an accessible, reproducible platform for single-cell-level spatial profiling of tumor tissue, enabling interactive visualization and standardized feature extraction without requiring programming expertise.

stat.AP↗

arXiv2026

MEDVISTAGYM: A Scalable Training Environment for Thinking with Medical Images via Tool-Integrated Reinforcement Learning

Vision language models (VLMs) achieve strong performance on general image understanding but struggle to think with medical images, especially when performing multi-step reasoning through iterative visual interaction. Medical VLMs often rely on static visual embeddings and single-pass inference, preventing models from re-examining, verifying, or refining visual evidence during reasoning. While tool-integrated reasoning offers a promising path forward, open-source VLMs lack the training infrastructure to learn effective tool selection, invocation, and coordination in multi-modal medical reasoning. We introduce MedVistaGym, a scalable and interactive training environment that incentivizes tool-integrated visual reasoning for medical image analysis. MedVistaGym equips VLMs to determine when and which tools to invoke, localize task-relevant image regions, and integrate single or multiple sub-image evidence into interleaved multimodal reasoning within a unified, executable interface for agentic training. Using MedVistaGym, we train MedVistaGym-R1 to interleave tool use with agentic reasoning through trajectory sampling and end-to-end reinforcement learning. Across six medical VQA benchmarks, MedVistaGym-R1-8B exceeds comparably sized tool-augmented baselines by 19.10% to 24.21%, demonstrating that structured agentic training--not tool access alone--unlocks effective tool-integrated reasoning for medical image analysis.

cs.CV↗

arXiv2025

Scaling Agentic Reinforcement Learning for Tool-Integrated Reasoning in VLMs

While recent vision-language models (VLMs) demonstrate strong image understanding, their ability to "think with images", i.e., to reason through multi-step visual interactions, remains limited. We introduce VISTA-Gym, a scalable training environment for incentivizing tool-integrated visual reasoning capabilities in VLMs. VISTA-Gym unifies diverse real-world multimodal reasoning tasks (7 tasks from 13 datasets in total) with a standardized interface for visual tools (e.g., grounding, parsing), executable interaction loops, verifiable feedback signals, and efficient trajectory logging, enabling visual agentic reinforcement learning at scale. While recent VLMs exhibit strong text-only reasoning, both proprietary and open-source models still struggle with tool selection, invocation, and coordination. With VISTA-Gym, we train VISTA-R1 to interleave tool-use with agentic reasoning via multi-turn trajectory sampling and end-to-end reinforcement learning. Extensive experiments across 11 public reasoning-intensive VQA benchmarks show that VISTA-R1-8B outperforms state-of-the-art baselines with similar sizes by 9.51%-18.72%, demonstrating VISTA-Gym as an effective training ground to unlock the tool-integrated reasoning capabilities for VLMs.

cs.AI↗

arXiv2025

MedAgentGym: A Scalable Agentic Training Environment for Code-Centric Reasoning in Biomedical Data Science

We introduce MedAgentGym, a scalable and interactive training environment designed to enhance coding-based biomedical reasoning capabilities in large language model (LLM) agents. MedAgentGym comprises 72,413 task instances across 129 categories derived from 12 authentic real-world biomedical scenarios. Tasks are encapsulated within executable sandbox environments, each featuring detailed task specifications, interactive feedback mechanisms, verifiable ground truth annotations, and scalable training trajectory generation. Extensive benchmarking of 29 LLMs reveals substantial performance disparities in biomedical data science between commercial and open-source LLMs. Leveraging efficient multi-threaded and multi-turn trajectory sampling in MedAgentGym, Med-Copilot achieves performance gains of +43.02% and +45.28% from offline and online reinforcement learning, respectively, demonstrating MedAgentGym as an effective training ground while establishing itself as a cost-effective, privacy-preserving alternative competitive with proprietary LLMs (gpt-4o). By offering a unified execution environment with a comprehensive benchmark and accessible, extensible training resources, MedAgentGym delivers an integrated platform to develop LLM-based coding assistants for advanced biomedical data science.

cs.CL↗

arXiv2025

GAS-MIL: Group-Aggregative Selection Multi-Instance Learning for Ensemble of Foundation Models in Digital Pathology Image Analysis

Foundation models (FMs) have transformed computational pathology by providing powerful, general-purpose feature extractors. However, adapting and benchmarking individual FMs for specific diagnostic tasks is often time-consuming and resource-intensive, especially given their scale and diversity. To address this challenge, we introduce Group-Aggregative Selection Multi-Instance Learning (GAS-MIL), a flexible ensemble framework that seamlessly integrates features from multiple FMs, preserving their complementary strengths without requiring manual feature selection or extensive task-specific fine-tuning. Across classification tasks in three cancer datasets-prostate (PANDA), ovarian (UBC-OCEAN), and breast (TCGA-BrCa)-GAS-MIL consistently achieves superior or on-par performance relative to individual FMs and established MIL methods, demonstrating its robustness and generalizability. By enabling efficient integration of heterogeneous FMs, GAS-MIL streamlines model deployment for pathology and provides a scalable foundation for future multimodal and precision oncology applications.

cs.CV↗

arXiv2025

Large Language Models Powered Multiagent Ensemble for Mitigating Hallucination and Efficient Atrial Fibrillation Annotation of ECG Reports

This study introduces a LLMs powered multiagent ensemble method to address challenges in hallucination and data labeling, particularly in large-scale EHR datasets. Manual labeling of such datasets requires domain expertise and is labor-intensive, time-consuming, expensive, and error-prone. To overcome this bottleneck, we developed an ensemble LLMs method and demonstrated its effectiveness in two real-world tasks: (1) labeling a large-scale unlabeled ECG dataset in MIMIC-IV; (2) identifying social determinants of health (SDOH) from the clinical notes of EHR. Trading off benefits and cost, we selected a pool of diverse open source LLMs with satisfactory performance. We treat each LLM's prediction as a vote and apply a mechanism of majority voting with minimal winning threshold for ensemble. We implemented an ensemble LLMs application for EHR data labeling tasks. By using the ensemble LLMs and natural language processing, we labeled MIMIC-IV ECG dataset of 623,566 ECG reports with an estimated accuracy of 98.2%. We applied the ensemble LLMs method to identify SDOH from social history sections of 1,405 EHR clinical notes, also achieving competitive performance. Our experiments show that the ensemble LLMs can outperform individual LLM even the best commercial one, and the method reduces hallucination errors. From the research, we found that (1) the ensemble LLMs method significantly reduces the time and effort required for labeling large-scale EHR data, automating the process with high accuracy and quality; (2) the method generalizes well to other text data labeling tasks, as shown by its application to SDOH identification; (3) the ensemble of a group of diverse LLMs can outperform or match the performance of the best individual LLM; and (4) the ensemble method substantially reduces hallucination errors. This approach provides a scalable and efficient solution to data-labeling challenges.

cs.AI↗

arXiv2023

Using Persistent Homology Topological Features to Characterize Medical Images: Case Studies on Lung and Brain Cancers

Tumor shape is a key factor that affects tumor growth and metastasis. This paper proposes a topological feature computed by persistent homology to characterize tumor progression from digital pathology and radiology images and examines its effect on the time-to-event data. The proposed topological features are invariant to scale-preserving transformation and can summarize various tumor shape patterns. The topological features are represented in functional space and used as functional predictors in a functional Cox proportional hazards model. The proposed model enables interpretable inference about the association between topological shape features and survival risks. Two case studies are conducted using consecutive 133 lung cancer and 77 brain tumor patients. The results of both studies show that the topological features predict survival prognosis after adjusting clinical variables, and the predicted high-risk groups have worse survival outcomes than the low-risk groups. Also, the topological shape features found to be positively associated with survival hazards are irregular and heterogeneous shape patterns, which are known to be related to tumor progression.

cs.CV↗

arXiv2021

A Bayesian Modified Ising Model for Identifying Spatially Variable Genes from Spatial Transcriptomics Data

A recent technology breakthrough in spatial molecular profiling has enabled the comprehensive molecular characterizations of single cells while preserving spatial information. It provides new opportunities to delineate how cells from different origins form tissues with distinctive structures and functions. One immediate question in spatial molecular profiling data analysis is to identify genes whose expressions exhibit spatially correlated patterns, called spatially variable genes. Most current methods to identify spatially variable genes are built upon the geostatistical model with Gaussian process to capture the spatial patterns, which rely on ad hoc kernels that could limit the models' ability to identify complex spatial patterns. In order to overcome this challenge and capture more types of spatial patterns, we introduce a Bayesian approach to identify spatially variable genes via a modified Ising model. The key idea is to use the energy interaction parameter of the Ising model to characterize spatial expression patterns. We use auxiliary variable Markov chain Monte Carlo algorithms to sample from the posterior distribution with an intractable normalizing constant in the model. Simulation studies using both simulated and synthetic data showed that the energy-based modeling approach led to higher accuracy in detecting spatially variable genes than those kernel-based methods. When applied to two real spatial transcriptomics datasets, the proposed method discovered novel spatial patterns that shed light on the biological mechanisms. In summary, the proposed method presents a new perspective for analyzing spatial transcriptomics data.

stat.AP↗

arXiv2020

Discovering Clinically Meaningful Shape Features for the Analysis of Tumor Pathology Images

With the advanced imaging technology, digital pathology imaging of tumor tissue slides is becoming a routine clinical procedure for cancer diagnosis. This process produces massive imaging data that capture histological details in high resolution. Recent developments in deep-learning methods have enabled us to automatically detect and characterize the tumor regions in pathology images at large scale. From each identified tumor region, we extracted 30 well-defined descriptors that quantify its shape, geometry, and topology. We demonstrated how those descriptor features were associated with patient survival outcome in lung adenocarcinoma patients from the National Lung Screening Trial (n=143). Besides, a descriptor-based prognostic model was developed and validated in an independent patient cohort from The Cancer Genome Atlas Program program (n=318). This study proposes new insights into the relationship between tumor shape, geometrical, and topological features and patient prognosis. We provide software in the form of R code on GitHub: https://github.com/estfernandez/Slide_Image_Segmentation_and_Extraction.

stat.AP↗

arXiv2020

Bayesian Modeling of Spatial Molecular Profiling Data via Gaussian Process

The location, timing, and abundance of gene expression (both mRNA and proteins) within a tissue define the molecular mechanisms of cell functions. Recent technology breakthroughs in spatial molecular profiling, including imaging-based technologies and sequencing-based technologies, have enabled the comprehensive molecular characterization of single cells while preserving their spatial and morphological contexts. This new bioinformatics scenario calls for effective and robust computational methods to identify genes with spatial patterns. We represent a novel Bayesian hierarchical model to analyze spatial transcriptomics data, with several unique characteristics. It models the zero-inflated and over-dispersed counts by deploying a zero-inflated negative binomial model that greatly increases model stability and robustness. Besides, the Bayesian inference framework allows us to borrow strength in parameter estimation in a de novo fashion. As a result, the proposed model shows competitive performances in accuracy and robustness over existing methods in both simulation studies and two real data applications. The related R/C++ source code is available at https://github.com/Minzhe/BOOST-GP.

stat.AP↗

arXiv2020

Bayesian Landmark-based Shape Analysis of Tumor Pathology Images

Medical imaging is a form of technology that has revolutionized the medical field in the past century. In addition to radiology imaging of tumor tissues, digital pathology imaging, which captures histological details in high spatial resolution, is fast becoming a routine clinical procedure for cancer diagnosis support and treatment planning. Recent developments in deep-learning methods facilitate the segmentation of tumor regions at almost the cellular level from digital pathology images. The traditional shape features that were developed for characterizing tumor boundary roughness in radiology are not applicable. Reliable statistical approaches to modeling tumor shape in pathology images are in urgent need. In this paper, we consider the problem of modeling a tumor boundary with a closed polygonal chain. A Bayesian landmark-based shape analysis (BayesLASA) model is proposed to partition the polygonal chain into mutually exclusive segments to quantify the boundary roughness piecewise. Our fully Bayesian inference framework provides uncertainty estimates of both the number and locations of landmarks. The BayesLASA outperforms a recently developed landmark detection model for planar elastic curves in terms of accuracy and efficiency. We demonstrate how this model-based analysis can lead to sharper inferences than ordinary approaches through a case study on the 246 pathology images from 143 non-small cell lung cancer patients. The case study shows that the heterogeneity of tumor boundary roughness predicts patient prognosis (p-value < 0.001). This statistical methodology not only presents a new model for characterizing a digitized object's shape features by using its landmarks, but also provides a new perspective for understanding the role of tumor surface in cancer progression.

stat.AP↗

arXiv2019

A Bayesian Zero-Inflated Negative Binomial Regression Model for the Integrative Analysis of Microbiome Data

Microbiome `omics approaches can reveal intriguing relationships between the human microbiome and certain disease states. Along with the identification of specific bacteria taxa associated with diseases, recent scientific advancements provide mounting evidence that metabolism, genetics and environmental factors can all modulate these microbial effects. However, the current methods for integrating microbiome data and other covariates are severely lacking. Hence, we present an integrative Bayesian zero-inflated negative binomial regression model that can both distinguish differentially abundant taxa with distinct phenotypes and quantify covariate-taxa effects. Our model demonstrates good performance using simulated data. Furthermore, we successfully integrated microbiome taxonomies and metabolomics in two real microbiome datasets to provide biologically interpretable findings. In all, we proposed a novel integrative Bayesian regression model that features bacterial differential abundance analysis and microbiome-covariate effects quantifications, which makes it suitable for general microbiome studies.

stat.AP↗

arXiv2019

Bayesian Modeling of Microbiome Data for Differential Abundance Analysis

The advances of next-generation sequencing technology have accelerated study of the microbiome and stimulated the high throughput profiling of metagenomes. The large volume of sequenced data has encouraged the rise of various studies for detecting differentially abundant taxonomic features across healthy and diseased populations, with the ultimate goal of deciphering the relationship between the microbiome diversity and health conditions. As the microbiome data are high-dimensional, typically featuring by uneven sampling depth, overdispersion and a huge amount of zeros, these data characteristics often hamper the downstream analysis. Moreover, the taxonomic features are implicitly imposed by the phylogenetic tree structure and often ignored. To overcome these challenges, we propose a Bayesian hierarchical modeling framework for the analysis of microbiome count data for differential abundance analysis. Under this framework, we introduce a bi-level Bayesian hierarchical model that allows a flexible choice of the count generating process, and hyperpriors in the feature selection scheme. We particularly focus on employing a zero-inflated negative binomial model with a Bayesian nonparametric prior model on the bottom level, and applying Gaussian mixture models for differentially abundant taxa detection on the top level. Our method allows for the simultaneous modeling of sample heterogeneity and detecting differentially abundant taxa. We conducted comprehensive simulations and summarized the improved statistical performances of the proposed model. We applied the model in two real microbiome study datasets and successfully identified biologically validated differentially abundant taxa. We hope that the proposed framework and model can facilitate further microbiome studies and elucidate disease etiology.

stat.ME↗

arXiv2018

ConvPath: A Software Tool for Lung Adenocarcinoma Digital Pathological Image Analysis Aided by Convolutional Neural Network

The spatial distributions of different types of cells could reveal a cancer cell growth pattern, its relationships with the tumor microenvironment and the immune response of the body, all of which represent key hallmarks of cancer. However, manually recognizing and localizing all the cells in pathology slides are almost impossible. In this study, we developed an automated cell type classification pipeline, ConvPath, which includes nuclei segmentation, convolutional neural network-based tumor, stromal and lymphocytes classification, and extraction of tumor microenvironment related features for lung cancer pathology images. The overall classification accuracy is 92.9% and 90.1% in training and independent testing datasets, respectively. By identifying cells and classifying cell types, this pipeline can convert a pathology image into a spatial map of tumor, stromal and lymphocyte cells. From this spatial map, we can extracted features that characterize the tumor micro-environment. Based on these features, we developed an image feature-based prognostic model and validated the model in two independent cohorts. The predicted risk group serves as an independent prognostic factor, after adjusting for clinical variables that include age, gender, smoking status, and stage.

cs.CV↗

arXiv2018

A Bayesian Mark Interaction Model for Analysis of Tumor Pathology Images

With the advance of imaging technology, digital pathology imaging of tumor tissue slides is becoming a routine clinical procedure for cancer diagnosis. This process produces massive imaging data that capture histological details in high resolution. Recent developments in deep-learning methods have enabled us to identify and classify individual cells from digital pathology images at large scale. The randomly distributed cells can be considered from a marked point process, where each point is defined by its position and cell type. Reliable statistical approaches to model such marked spatial point patterns can provide new insight into tumor progression and shed light on the biological mechanisms of cancer. In this paper, we consider the problem of modeling spatial correlations among three commonly seen cells (i.e. lymphocyte, stromal, and tumor) observed in tumor pathology images. A novel marking model of marked point processes, with interpretable underlying parameters (some of which are clinically meaningful), is proposed in a Bayesian framework. We use Markov chain Monte Carlo (MCMC) sampling techniques, combined with the double Metropolis-Hastings (DMH) algorithm, to sample from the posterior distribution with an intractable normalizing constant. On the benchmark datasets, we demonstrate how this model-based analysis can lead to sharper inferences than ordinary exploratory analyses. Lastly, we conduct a case study on the pathology images of 188 lung cancer patients from the National Lung Screening Trial. The results show that the spatial correlation between tumor and stromal cells predicts patient prognosis. This statistical methodology not only presents a new model for characterizing spatial correlations in a multi-type spatial point pattern, but also provides a new perspective for understanding the role of cell-cell interactions in cancer progression.

stat.ME↗