Papers for

clinical decision support developers

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Combining models improves ICU vital sign forecasts for each variable

Which Model, For Which Variable? Per-Variable Forecast Combination in Sparse ICU Time Series

Abstract: Clinical forecasting papers typically report a single best architecture. We show that on multivariate ICU physiology this question conceals variable-specific differences: no architecture is best on every variable, and which model wins changes from variable to variable. Across five ICU databases (MIMIC-III, MIMIC-IV, eICU, HiRID, and UMM (University Medicine Mannheim), a German single-center ICU) we forecast 94 channels over a 24-hour horizon with a pool of 14 models, and combine them with weights fitted per variable. On test, the combination beats the best single model on all five databases by 6.1-11.0% mean squared error (mean 8.7%). The gain is concentrated where clinical data is thinnest: the sparsest third of channels improves by 11.9% and laboratory values by 7.5%, against 3.1% for the densest third. Rank-based harmonic weighting of the best five models per variable, with no fitted weights at all, recovers about four-fifths of the gain. Extending the pool afterwards with a per-channel gradient-boosted tree, ARIMA and ETS makes the tree the validation-selected best single model on all five databases; per-variable combination still improves on it by 2.9-5.9% (significant on four of five), while equal weighting of the same members does not. As on univariate benchmarks, learned combination beats simple averaging - but only once the weights are constrained. We also report which model to prefer when, and show that the ordering changes with database size.

Fri 13 NovMachine Learning
The gist
Forecasting medical variables in intensive care units is tricky because no single prediction model works best for all measurements. The authors show that taking the best model for each variable and combining predictions improves accuracy, especially for data that is sparse or less frequently recorded. They tested this approach on five different ICU datasets and found consistent improvement over any single model. Their method also adapts to different data sizes and types, helping to choose the right model for each variable.
Open → 2611.18149v1

Medication recommendation system improves safety with clinical evidence

Clinical Evidence-Guided Residual Learning for Medication Recommendation with Dynamic DDI Regularization

Abstract: Medication recommendation must combine patient-specific treatment evidence with the interaction structure of a prescription. The central challenge is to determine which evidence supports each medication and how strongly to regularize interactions while preserving predictive quality. We propose EviDrug, a clinical evidence-guided residual learning framework. Its clinical-memory backbone integrates a self-trained visit-aware encoder, training-case memory and one medication-set interaction block to establish a patient-specific prescription. A gated refinement head combines the backbone states with clinical associations and history-conditioned medication identities through bounded logit corrections. Drug-drug interaction (DDI) feedback adapts regularization strength using training-query predictions. Across MIMIC-IV, NWICU and SICdb, EviDrug leads twelve published-method baselines in AP, F1 and Jaccard, improving AP by 2.26, 2.06 and 1.71 percentage points over the strongest comparator. These gains come at lower parameter and runtime costs than MetaDrug. Matched controls demonstrate DDI reductions with stable predictive quality. Code is available at https://github.com/LIZESHENG13/EviDrug.

Sat 7 NovMachine LearningInformation Retrieval
The gist
Recommending the right combination of medicines for a patient is hard because some medicines can interact in harmful ways. The authors created EviDrug, a system that learns from clinical evidence and patient history to suggest safer prescriptions. It uses smart ways to understand each patient's needs while reducing risky drug interactions. Their tests showed EviDrug works better and faster than older methods on real hospital data.
Open → 2611.10613v1

Graph neural network improves multimodal retinal disease detection accuracy

Interpretable Context-Aware Multi-Label Retinal Classification with Dual-Hub Heterogeneous Graph Neural Network

Abstract: Multi-label retinal screening requires capturing co-occurring pathologies and clinical context, yet standard multimodal fusion remains opaque. We propose a dual-hub heterogeneous graph neural network (GNN) that couples spatial fundus features and structured clinical covariates through cross-modal message passing while providing inherent interpretability. Rather than relying on post-hoc saliency approximations, our architecture exposes multi-level decision pathways during forward aggregation: spatially biased self-attention captures lesion-level patch interactions, while modality-specific graph attention coefficients explicitly quantify the contribution of visual regions and clinical risk factors to each disease node. Evaluated on the BRSET dataset (16,266 photographs from 8,524 patients), the proposed model achieves 76.1% Macro-AUPRC across seven conditions, matching or exceeding image-only (75.9%) and ResNet-50+MLP fusion (75.9%) baselines, with largest contextual gains in macular edema (+3.5%) and diabetic retinopathy (+2.1%). Codebase is available at: https://github.com/proshanto-c/ca-mlc-refactor.

Fri 6 NovComputer Vision and Pattern RecognitionMachine Learning
The gist
Detecting multiple eye diseases from retinal images is tricky because different conditions can appear together and patient information matters. The authors developed a special kind of graph-based AI that looks at both the images and clinical data together, clearly showing how each part contributes to the decision. Their method matched or beat existing image-only techniques on a big eye health dataset, especially improving detection of macular edema and diabetic retinopathy. This approach also helps doctors understand why the AI made each diagnosis.
Open → 2611.09309v1

MedCaseAgent generates accurate rare disease reports from clinical data

MedCaseAgent: From Multimodal Clinical Evidence to Rare Disease Case Reports

Abstract: Although Large Language Models (LLMs) have demonstrated potential in clinical documentation, generating the rigorous, evidence-based reasoning required for unprecedented, atypical, and rare disease case reports remains a significant challenge. We present MedCaseAgent, an end-to-end agentic reasoning paradigm that bridges this gap by synthesizing publication-quality reports from unstructured clinical notes. MedCaseAgent leverages a hierarchical structure, integrating patient-level data mining with iterative, tool-integrated reasoning. By dynamically querying medical literature and knowledge bases, the agent maintains high clinical fidelity and assesses case novelty. The framework is inherently modular, supporting various backbone LLMs in a plug-and-play fashion. To facilitate this research, we release MedCase-150K, a large-scale multimodal clinical case-report corpus, and MedCase-Bench, a robust temporally held-out evaluation suite. We also introduce an open-source model post-trained from Qwen3.5-9B and optimized specifically for this task. Under both LLM-judged and rule-based evaluation, MedCaseAgent outperforms existing general-purpose and specialized medical LLMs and agents. Blinded expert reviews and ablations further support the system's clinical relevance and evidence grounding.

Fri 23 OctArtificial Intelligence
The gist
Writing detailed reports for rare diseases is very hard because it needs careful evidence and new insights. The authors created MedCaseAgent, a tool that reads patient notes and checks medical sources to write accurate, publication-ready reports. It can work with different language models and learns to keep reports trustworthy and novel. They also built a large dataset of medical cases and a testing system, and made a version of their model available to use. Experts reviewing the reports found them useful and well supported by evidence.
Open → 2610.32258v1

Clinical semantic guidance improves sepsis treatment with decision transformer

ClinSDT: LLM-Encoded Clinical Semantic Guidance for Sepsis Treatment via Decision Transformer

Abstract: Offline sepsis treatment aims to learn treatment policies from retrospective ICU trajectories. Decision Transformer-based methods provide a natural sequence modeling framework for this task, but they mainly rely on raw physiological states and historical clinician actions. Such raw trajectory tokens may not explicitly capture clinical semantics that clinicians use in practice, such as patient severity, recent treatment response, and current risk priorities. To address this limitation, we propose ClinSDT, an LLM-assisted Decision Transformer framework for offline sepsis treatment. ClinSDT uses a frozen LLM as a semantic bridge rather than a treatment decision maker. It constructs three clinical semantic views from ICU trajectories, namely Profile, Context, and Advisory, and inserts their encoded representations as independent tokens into a Decision Transformer policy. To support policy refinement and rollout, ClinSDT further trains a Treatment-Response Forecaster to estimate the next physiological state, SAPS-II change, and predictive uncertainty from state-action inputs. In the second training stage, the forecaster screens local counterfactual treatment candidates, and only low-uncertainty candidates with lower predicted SAPS-II increase are distilled into the policy after semantic regeneration. During inference, ClinSDT performs closed-loop treatment rollout by refreshing semantic inputs from the currently available patient state. Experiments on MIMIC-III and MIMIC-IV show that ClinSDT achieves consistent relative improvements over representative imitation learning, offline reinforcement learning, and Decision Transformer baselines under the same rollout protocol. Code and data are available at https://github.com/KAG778/ClinSDT.

Sun 18 OctMachine LearningArtificial Intelligence
The gist
Sepsis treatment in intensive care involves complex decisions based on patient data. The authors present ClinSDT, a method that helps computer models understand clinical meanings like patient condition and treatment effects by using a large language model to summarize data. This approach improves decision-making by combining these clinical summaries with a system that predicts patient responses to treatments. Their experiments show this method performs better than previous models when tested on real ICU patient records.
Open → 2610.25121v1

Medical concept IDs improve predictions by capturing clinical hierarchy

SMI: Ontology-Grounded Semantic IDs for Hierarchy-Aware Medical Concept Representation

Abstract: Biomedical language models (LMs) provide strong representations for clinical tasks, but they often fail to capture the hierarchical structure among medical concepts. This limitation is especially pronounced for rare and fine-grained clinical codes in long-tailed vocabularies, where hierarchy-agnostic LM representations become less informative as concept frequency decreases. We introduce Semantic Medical Identifier (SMI), a training-free framework for constructing ontology-grounded Semantic IDs and hierarchy-aware medical concept representations. For each diagnosis, procedure, or medication concept, SMI defines its Semantic ID as the coarse-to-fine path through an expert-defined medical ontology. SMI then aggregates frozen biomedical LM embeddings across the corresponding ontology levels to form hierarchy-aware medical concept representations that preserve coarse-to-fine clinical semantics. The base SMI uses uniform aggregation across ontology levels, while a level-weighted variant selects global level weights on validation data. Our theoretical analysis shows that SMI yields more informative medical concept representations than hierarchy-agnostic LM embeddings, particularly for rare and fine-grained clinical concepts. Experiments show that SMI captures meaningful hierarchical structure and improves clinical relation prediction. On MIMIC-IV, SMI improves mortality prediction, readmission prediction, drug recommendation, and rare drug prediction. On eICU, SMI yields more robust patient representations under cross-hospital distribution shift.

Tue 13 OctComputation and LanguageArtificial Intelligence
The gist
Medical computer programs often struggle to understand how different medical terms relate to each other, especially for rare conditions. The authors introduce Semantic Medical Identifier (SMI), which uses expert medical classifications to give each concept a meaningful ID showing its place in a hierarchy. This helps existing language models better understand and represent medical concepts, improving predictions for patient outcomes and drug recommendations. Testing on real hospital data showed SMI made predictions more accurate and robust across different settings.
Open → 2610.18672v1

Large language models struggle to use evidence in dynamic clinical diagnosis

From Hindsight to Foresight: Dissecting Clinical Evidence Acquisition in Large Language Models

Abstract: Large language models (LLMs) often diagnose full sight, static clinical cases well yet struggle in foresight, dynamic diagnosis. Prior work has improved this limitation by quantifying the information models acquire and by developing stronger follow-up strategies to improve diagnostic accuracy. In contrast, we introduce a distinctive hindsight to systematically dissect why LLMs fail to acquire diagnostically critical evidence, tracing the failure from evidence identification, utilization and downstream diagnostic updating through not only diagnostic accuracy, but also evidence coverage, trajectory dynamics, and elicited probability and information entropy. Starting experiments from the breakdown at the level of capabilities, to the diagnosis of the pre-event state before the evidence arrives, and to the post-event state transition after the evidence arrives, our research has revealed the profound limitations and internal causes existing in the dynamic diagnostic process. At the same time, we have also conducted sufficient verification and generalization experiments to prove that these limitations are widespread. Together, these findings shift the central challenge from whether LLMs possess relevant clinical knowledge to whether they can identify, acquire, and effectively use the right evidence under clinical diagnosis and conduct a more detailed mechanism analysis to identify the exact location of the capability gap.

Tue 13 OctComputation and Language
The gist
Large language models are good at diagnosing medical cases when all information is available at once but have trouble when diagnosis requires thinking ahead and updating with new evidence over time. The authors studied where exactly these models fail in identifying, using, and learning from important clinical evidence during dynamic diagnosis. They found that the problem is not lack of medical knowledge but the models’ difficulty in handling evolving information. Their experiments show this issue is widespread and pinpoint the capability gaps for future improvements.
Open → 2610.18605v1

MedEvo improves medical reasoning by repairing student errors

MedEvo: Medical Reasoning Evolution through Knowledge Navigation

Abstract: Medical multimodal reasoning models are typically trained on teacher-generated rationales, discarding the valid reasoning in students' own failed trajectories, while general-domain correction methods offer little guidance for errors in image reading, anatomical context, or differential diagnosis. We propose MedEvo, which uses medical knowledge, namely a tag taxonomy and error-type diagnosis, to navigate both trajectory repair and data selection. For each failed rollout, a gold-blind auditor locates the first consequential error and provides a minimal, answer-free hint; the same student retains the valid prefix and regenerates the erroneous suffix. After each round, MedEvo identifies capability gaps across medical tags on a held-out feedback set and prioritizes related questions for new rollouts and repairs. With supervised fine-tuning (SFT) alone, MedEvo raises Qwen3.5-9B mean accuracy across seven benchmarks from 65.57% to 69.12% (47.40% to 54.50% on the multimodal track of MedXpertQA (MedXpertQA-MM)), surpassing teacher-distillation SFT (66.93%). It also raises Qwen3.5-35B and Qwen3.5-397B mean accuracy from 67.91% to 73.04% and from 71.83% to 74.95%, respectively. Our analyses show that failed trajectories contain reasoning worth preserving, that repair success does not imply learning value, and that independent feedback is key to sustained improvement.

Tue 13 OctArtificial IntelligenceComputer Vision and Pattern RecognitionMachine Learning
The gist
Medical AI models often ignore mistakes made during reasoning, missing chances to learn from them. The authors present MedEvo, which identifies the exact mistakes in medical reasoning tests and gives hints without revealing answers so the AI can try better answers. This method focuses training on weak areas and improves performance on medical exams. Their results show that fixing these errors leads to better scores compared to traditional training.
Open → 2610.17690v1

Medication recommendations improved for better safety and clarity

RES-MR: Risk-Aware Reasoning for Explainable and Safe Medication Recommendation

Abstract: Medication recommendations seek to deliver personalized drug combinations based on patients' clinical information, where ensuring transparent decision-making and personalized safety are of paramount importance. However, existing approaches predominantly operate as ``black boxes'', offering limited interpretability and hindering clinician trust. Furthermore, they typically rely on fixed, patient-agnostic safety constraints, neglecting the heterogeneity of risk tolerance across individuals, which poses significant risks to vulnerable populations. To address these gaps, we propose \textbf{RES-MR}, a novel Risk-aware Reasoning framework for Explainable and Safe Medication Recommendation using LLMs. RES-MR follows a two-stage training paradigm: (1) an explainable clinical reasoning distillation stage to elicit diagnostic capabilities by fine-tuning on synthesized reasoning trajectories, grounded in patient-specific knowledge from graph retrieval; and (2) a risk-aware policy optimization stage to dynamically balance therapeutic efficacy with personalized safety. This stage employs disentangled drug factors to construct patient-specific reference prototypes that capture desired therapeutic drug features and risk features to avoid. These prototypes enable risk boundary calibration and guide the model toward safe recommendations via reward shaping and clipping. Extensive experiments on benchmark datasets demonstrate that RES-MR significantly outperforms state-of-the-art baselines in terms of accuracy, safety, and explainability. Our code and data are available at: https://github.com/wangcong2001/RES-MR.

Sat 10 OctMachine LearningArtificial IntelligenceInformation Retrieval
The gist
Medication recommendation systems help doctors choose medicines for patients based on their health data, but many current systems don’t explain their choices well and use the same safety rules for everyone. The authors introduce RES-MR, a method that teaches computers to explain their reasoning when suggesting drug combinations while adjusting safety rules to each patient's individual risk. This approach uses large language models and special training steps that focus on understanding patient data and balancing treatment benefits with personalized safety. Their tests show it performs better at giving accurate, safe, and explainable medication advice.
Open → 2610.14474v1

Accurate thyroid nodule segmentation with uncertainty maps in ultrasound images

Thyroid Nodule Segmentation in Ultrasound Images with Conformal Uncertainty Estimation

Abstract: Thyroid nodules affect a large proportion of the general population. Although most are benign, a non-negligible fraction is malignant. Ultrasound imaging is the primary diagnostic modality, and accurate nodule segmentation is a prerequisite for CAD systems to classify malignancy and assess metastatic risk.\\ In this work, we propose a thyroid nodule segmentation framework that integrates a DeepLabV3+, channel-attention modules and post-hoc conformal prediction, providing both accurate mask and pixel-level uncertainty estimates. We address thyroid ultrasound nodule segmentation as the first stage of a pipeline that also includes malignancy classification and metastatic risk prediction. We adopt DeepLabV3+, whose Atrous Spatial Pyramid Pooling (ASPP) captures multi-scale context, with an EfficientNet-B5 backbone for its accuracy-efficiency trade-off. Additionally, we test three channel-attention modules: Squeeze-and-Excitation, CBAM and Grouped Channel Attention (GCA), paired with two loss formulations (Dice-BCE and Boundary).\\ Since binary masks do not indicate prediction confidence, we wrap the trained model with a split conformal prediction layer that produces pixel-level uncertainty maps with statistical coverage guarantees.\\ We evaluate all configurations on the TN3K public benchmark using a fixed split and directly compare them with state-of-the-art methods. The plain backbone already improves upon previously reported results, with CBAM + Dice-BCE emerging as the best overall configuration.

Fri 9 OctComputer Vision and Pattern Recognition
The gist
Thyroid nodules are common, and while most are harmless, some can be cancerous. Doctors use ultrasound images to check nodules, but they need precise outlines of these nodules for better diagnosis. The authors created a computer method that not only finds these nodules accurately in the images but also shows how confident it is about each pixel it labels. This helps improve the first step in diagnosing whether a nodule might be dangerous. Their method improves on previous techniques and provides a way to understand uncertainty in the segmentation.
Open → 2610.13604v1

SepsisLens predicts early sepsis alerts while showing key warning signals

SepsisLens: Structure-Preserving Sequence Modelling for Decomposable Early Sepsis Warning

Abstract: Early sepsis warning from ICU records can be cast as a structure-preserving prediction problem. A model needs to detect deterioration from irregular measurements while keeping each alert connected to the physiological signals that support it. Many temporal models fuse clinical variables into a patient-level representation, supporting scalar risk prediction but weakening the structure needed for clinical decomposition. We present SepsisLens, which preserves variable-indexed temporal states until risk composition. Observation-aware representations encode each variable's dynamics and measurement history, while a shared temporal encoder models each trajectory without collapsing the variable axis. The StructuredRiskHead composes multi-horizon risk from explicit variable-level and organ-level components. We evaluate SepsisLens on three public ICU cohorts and one private-hospital cohort under a common pre-onset protocol. SepsisLens achieves strong discrimination on all four cohorts and lower alert burden at matched event recall on MIMIC-IV. Structural ablations support the design, while input-side masking shows that the ranked components reflect variables with greater influence on prediction.

Tue 6 OctMachine Learning
The gist
SepsisLens is a computer model designed to warn doctors early about sepsis, a dangerous infection that often occurs in intensive care units. Unlike other approaches that mix many patient measurements into a single score, SepsisLens keeps track of each vital sign separately and explains which ones drove the warning. This helps doctors understand why an alert happens and which parts of the body might be deteriorating. The authors tested SepsisLens on four different hospital data collections, where it made accurate predictions while reducing unnecessary alerts.
Open → 2610.08046v1

Budget-aware language model improves medical risk predictions with verifiable knowledge

Cite What You Explore: Budget-Aware LLM Reasoning over Medical KGs with Verifiable Evidence

Abstract: Post-discharge risk prediction from electronic health records (EHRs) is difficult because many dependencies that link discharge-time observations to downstream complications, such as comorbidity cascades and drug-disease interactions, are absent from the record. External medical knowledge graphs (KGs) can supply these missing dependencies, but tracing them demands three properties: KG exploration must remain cost-bounded, retrieved evidence must be differentiated by source quality, and the resulting rationale must be citable for retrospective review. Large language models (LLMs) can plan and verify over structured evidence, making them natural candidates for KG reasoning, but existing LLM-based methods do not satisfy these three properties jointly. In this paper, we propose BAR, a Budget-Aware LLM Reasoning framework over medical KGs with three contributions. First, BAR refines the raw KG into disease-specific evidence graphs whose edges carry support scores and provenance records, turning the KG into a quality-annotated reasoning space rather than a static feature source. Second, an LLM then reasons over this graph through a plan-navigate-verify loop that decomposes the question into steps, retrieves evidence under a patient-specific budget, and revises when verification fails. Third, a reasoning policy is trained with a reward that compares predictions with and without acquired evidence, combined with acquisition cost and citation-integrity terms. Across 8 diseases and 3 prediction horizons on MIMIC-III and MIMIC-IV, BAR improves AUPRC by 3.4 points over the strongest baseline, raises citation precision from 59.8% to 77.9%, and consumes only 62-65% of the budget cap.

Tue 6 OctMachine Learning
The gist
Predicting health risks after patients leave the hospital is hard because medical records don’t show all important links between diseases and treatments. The authors created a system called BAR that uses large language models to explore medical knowledge graphs carefully and within a budget. BAR retrieves good-quality evidence to support its predictions and clearly shows where its information comes from, making it easier for doctors to check the reasoning. This approach improves accuracy and trustworthiness while keeping the cost of retrieving information low.
Open → 2610.07739v1

Patient-centered treatment planning improves health outcomes in chronic multimorbidity

Patient-Centered Treatment Planning for Chronic Multimorbidity: A Hierarchical Reinforcement Learning Framework for Preference Modeling

Abstract: Patient preference, defined as a patient's demonstrated willingness and capacity to adhere to clinical recommendations, is a primary determinant of therapeutic effect yet remains structurally absent from existing computational treatment planning models. We address this gap by presenting patient-centered factored-action hierarchical option-critic (FAHOC), a hierarchical reinforcement learning (HRL) framework that jointly learns high-level options corresponding to therapeutic strategies and factored intra-option policies that decompose the joint action space into disease- and intervention-specific subcomponents, while imposing a cooperation-aware action masking mechanism. This enables structured exploration, improved credit assignment across hierarchy levels, and more interpretable decision pathways, while enforcing patients' preferences. Formal guarantees establish that cooperative patients achieve higher optimal expected health outcomes than non-cooperative patients, and that the factored Q-function approximation error is provably bounded. The framework is evaluated using longitudinal data collected from approximately 50,000 comorbid hypertension and type 2 diabetes mellitus patients from five hospitals in the Southeast U.S. FAHOC achieves a quality-adjusted life year expectancy equivalent improvement of 0.669 (vs -0.133 observed clinician practice), correctly identifies cooperative patients in 95.9% of cases and never violates a patient's preference in held-out test, demonstrating that HRL with explicit preference constraints can support preference-consistent, clinically safe decision-making in multimorbidity management.

Wed 30 SeptMachine Learning
The gist
People with multiple chronic illnesses like hypertension and diabetes often have complex treatment plans. This paper presents a new computer method that learns treatment strategies while respecting patients' preferences on what treatments they are willing and able to follow. The method uses a hierarchical learning approach to break down decisions by disease and therapy, making it easier to understand and safer for patients. The authors show with real patient data that following this method leads to better health outcomes and respects patient choices.
Open → 2609.39911v1

Physician context improves heart risk detection from chest x rays

From Image Interpretation to Clinical Reasoning: Upstream Physician-Context-Aware Multimodal Learning with Causal Reinforcement Learning

Abstract: Major adverse cardiovascular events (MACE) remain the leading cause of mortality worldwide. Opportunistic screening using routinely acquired clinical data offers a scalable approach for identifying high-risk individuals before acute events occur. Although chest X-rays (CXRs) capture latent cardiovascular biomarkers and clinical histories provide complementary patient context, existing medical vision-language models are primarily optimized for radiology interpretation rather than prognostic reasoning. We propose a causal reinforcement learning framework for multimodal clinical reasoning that integrates CXRs and physician-authored clinical histories for opportunistic MACE prediction. The framework introduces (1) a role-decoupled dual-LLM architecture that separates reasoning from risk prediction, (2) a dual-action causal reinforcement learning policy for evidence selection and reasoning optimization, and (3) causal token pruning to learn compact multimodal representations. Evaluated on an internal cohort, an emergency department cohort, and the external MIMIC dataset, the proposed framework consistently outperformed unimodal baselines and state-of-the-art medical vision-language models, achieving AUROCs of 0.720, 0.760, and 0.845, respectively. It also substantially improved reasoning quality, achieving higher GREEN scores and higher expert preference while maintaining robust predictive performance across diverse patient populations.

Wed 30 SeptComputer Vision and Pattern Recognition
The gist
Heart attacks and other major heart problems cause many deaths worldwide. Doctors often take chest x rays to check patients, which hold clues about heart health, but current AI systems mainly focus on reading images instead of predicting future risks. The authors created a new AI method that combines chest x rays with doctors’ notes to better predict which patients might have heart problems later. Their system learns to pick out important information and think through the risks more clearly, leading to better predictions across multiple patient groups.
Open → 2609.38924v1

Llm accuracy drops as answer choices grow in number

Overwhelmed by Choice: Studying LLM Decision Making at Scale

Abstract: Multiple-choice and candidate-selection evaluations are widely used to assess LLM reasoning and decision-making, yet most benchmarks contain relatively small candidate sets. It remains unclear whether conclusions drawn from these settings remain valid as the candidate space scales. We systematically evaluate LLMs as the number of competing candidates increases and find substantial accuracy degradation across tasks, prompting strategies, and model scales. Controlled analyses show that standard long-context retrieval explanations cannot fully account for this degradation. Instead, we identify two systematic failure patterns. First, gold-margin collapse: the score gap between the correct answer and the strongest distractor progressively shrinks, driven primarily by weakening confidence in the correct answer. Second, earlier candidate preferences become increasingly difficult to overturn, with later candidates exerting progressively weaker influence on the final prediction. Motivated by these findings, we evaluate hierarchical partitioning and permutation-based inference, which improve accuracy by roughly 20 percentage points at $N=160$ on both HotpotQA and MIMIC. Overall, our results identify candidate-set scale as an important evaluation-protocol variable and show that strong small-option performance does not necessarily imply robust large-scale candidate comparison.

Sat 26 SeptArtificial IntelligenceComputation and Language
The gist
Choosing the right answer from many options is harder for large language models than previously thought. The authors found that as the number of answer choices increases, the models' accuracy drops significantly. Two main problems cause this: the model becomes less confident in the correct answer and early preferences for answers are hard to change. They tested new ways to organize and evaluate answers that helped improve accuracy with lots of choices.
Open → 2609.32809v1

Clinical concept embeddings improve hospital outcome predictions

HCOE: Hyperbolic Clinical Ontology Embeddings from Biomedical Language Models

Abstract: Biomedical language models (LMs) encode textual semantics but do not explicitly preserve medical code hierarchies. We present Hyperbolic Clinical Ontology Embeddings (HCOE) for hierarchy-aware clinical concept representation. HCOE maps frozen BioBERT embeddings into a Poincare ball, combining parent-side and child-side ontology-guided contrastive learning with coarse-to-fine ontology-path aggregation. It uses International Classification of Diseases (ICD) codes organized by Clinical Classifications Software (CCS) and Anatomical Therapeutic Chemical (ATC) medication hierarchies. Evaluations show that HCOE performs best on ICD/ATC clinical relation prediction and CCS-to-PheCode hierarchy transfer. On the MIMIC-IV dataset, HCOE also achieves the best performance on mortality prediction, readmission prediction, medication recommendation, and rare drug prediction.

Fri 25 SeptArtificial IntelligenceMachine Learning
The gist
Medical codes used in health records are organized in hierarchies, but many computer models do not use these relationships explicitly. The authors created a method called HCOE that combines medical language understanding with mathematical shapes that better reflect these hierarchies. This approach helps computers better predict patient outcomes like mortality and readmission, as well as recommend medications. The research shows that including code hierarchies makes clinical predictions more accurate.
Open → 2609.30763v1

Medgpt framework improves complex lung cancer decision making

An auditable conditional-strategy framework for open-ended decision-making in complex lung cancer

Abstract: Complex lung cancer decisions can involve several defensible pathways whose eligibility, sequencing and safety depend on unresolved information. Effective support must make explicit how patient conditions govern pathway eligibility, deferral and redirection. MedGPT Clinical Explorer (MCE) organizes alternatives, decision-changing unknowns, safety constraints and fallback into a conditional strategy for clinician review. To evaluate this representation in physician-authored strategies, multidisciplinary experts established case-specific references for 40 cases within a purposive 100-case corpus, and 250 physicians from 98 institutions produced 2,250 strategies under unaided, retrieval-reference and MCE-assisted conditions. MCE-assisted strategies expressed more applicable clinical requirements, measured by the Admissible Pathway Attainment Score (APAS; 0-100), than unaided strategies (adjusted difference, 12.87; 95% CI, 11.18-14.55) and retrieval-reference strategies (5.22; 3.52-6.93). With the same knowledge base available in the retrieval-reference and MCE-assisted conditions, the additional content centered on candidate pathways, decision-critical information and safety constraints. Physicians' whole-strategy acceptability judgments correlated with APAS (Spearman's rho = 0.671), while a complementary relationship audit assessed whether candidates, conditions and subsequent actions were coherently connected. Together, these findings identify two complementary dimensions of open-ended decision support: coverage of clinically relevant content and coherent links among pathways, conditions and subsequent actions. MCE provides a shared decision object that makes consequential omissions and pathway contingencies visible before action; prospective studies should evaluate its effects on clinical workflow and patient outcomes.

Thu 24 SeptArtificial Intelligence
The gist
Deciding the best treatment for lung cancer can be complicated because many options depend on specific patient information that is sometimes missing. The authors created a tool called MedGPT Clinical Explorer that helps doctors organize possible treatment pathways, unknown patient factors, and safety issues all in one place. In tests with hundreds of doctors, those using this tool made more complete and clearer treatment plans than those without it. The tool also makes sure all treatment decisions are logically connected and clearly explained before doctors act on them.
Open → 2609.29381v1

Radiomics and 3D image features improve kidney cancer subtype classification

Complementary Roles of Radiomics and Foundation Representations in Renal Cell Carcinoma Classification: A Comparative Study of 2D and 3D CT Encodings

Abstract: Accurate preoperative subtype classification of renal cell carcinoma (RCC) from contrast-enhanced computed tomography remains clinically challenging. Radiomics provides structured tumour descriptors, whereas foundation representations offer transferable image features. However, it remains unclear whether radiomics still adds value beyond pretrained representations, and how 2D and 3D MedVAE encoders compare in this setting. We compared handcrafted radiomics, 2D MedVAE, 3D MedVAE, and their fusion for binary clear-cell RCC versus non-clear-cell RCC classification on KiTS23 under a unified preprocessing pipeline. Concatenation, cross-attention, and gated fusion were evaluated as representative integration strategies, and radiomics feature importance was analysed to support decision-centric interpretability. Fusion consistently improved discrimination over image-only MedVAE branches. The best overall performance was achieved by 3D gated fusion, with an AUC of 82.7\%, outperforming the best 2D fusion model (79.6%), the radiomics baseline (74.4%), and the single-modality MedVAE branches. Ablation analysis further showed clear gains of the full fusion model over both image-only and radiomics-only variants, indicating complementary contributions from radiomics and image representations. These findings suggest that radiomics remains relevant for RCC CT classification in the presence of foundation representations, and that its integration with MedVAE is more effective in the 3D setting. More broadly, the study supports a complementary role for radiomics and foundation representations in clinically meaningful imaging decision support.

Tue 22 SeptComputer Vision and Pattern RecognitionArtificial Intelligence
The gist
It is hard to tell different types of kidney cancer just by looking at CT scans before surgery. This study shows that combining detailed, handcrafted features called radiomics with advanced 3D image analysis gives better results than using either alone. The researchers tested several ways to combine these sources and found that using 3D data with a fusion method worked best. Their work suggests these two approaches together help doctors classify kidney tumors more accurately.
Open → 2609.26463v1

Ultrasound and clinical data improve liver cancer invasion prediction

Preoperative Prediction of Microvascular Invasion in Hepatocellular Carcinoma by Integrating Multimodal Ultrasound and Clinical Data: A Multicenter Study

Abstract: Background: Microvascular invasion (MVI) predicts recurrence and survival in hepatocellular carcinoma (HCC) but requires postoperative histopathology for diagnosis. We developed and validated a model integrating multimodal ultrasound and clinical data for preoperative MVI prediction. Methods: This multicenter study included 489 patients with HCC from eight centers. All patients had B-mode ultrasound (BUS), color Doppler flow imaging (CDFI), dynamic contrast-enhanced ultrasound (DCE-US), and clinical information. Data from seven centers (n = 421) were used for model development with five-fold cross-validation; data from the remaining center (n = 68) formed an independent external validation cohort. The proposed multimodal information fusion network used modality-specific encoders, a hemodynamic temporal change module for bidirectional DCE-US perfusion changes, and a representation consistency learning module to align heterogeneous ultrasound representations before Transformer-based fusion. Results: In external validation, DCE-US achieved the highest single-modality area under the receiver operating characteristic curve (AUC; 0.8545+/-0.0198), versus clinical information (0.6715+/-0.0156), CDFI (0.6435+/-0.0344), and BUS (0.6087+/-0.0417). Pixel-difference sampling and the proposed temporal module outperformed alternative sampling and video representation methods. The full model achieved the best performance, with an AUC of 0.8953+/-0.0180, accuracy of 81.18%+/-2.83%, sensitivity of 86.40%+/-6.69%, and specificity of 78.14%+/-6.28. Conclusions: Integrating multimodal ultrasound and clinical information enabled promising preoperative MVI prediction in HCC. DCE-US was the main source of predictive information, while BUS, CDFI, and clinical information provided complementary value. The proposed framework may support preoperative risk stratification and individualized clinical decision-making.

Mon 21 SeptComputer Vision and Pattern Recognition
The gist
Microvascular invasion in liver cancer is important to know before surgery because it affects recovery and survival, but it is usually only found after surgery. The authors created a computer model that looks at different types of ultrasound images and clinical information to predict this invasion before surgery. They tested their model using data from multiple hospitals and found it made more accurate predictions than using any single type of information. This tool could help doctors better decide how to treat liver cancer patients before surgery.
Open → 2609.24524v1

Prototype purification improves multi disease detection in chest x rays

Purification and Regulation: Comorbidity-Aware Multi-Label Few-Shot Learning for Medical Image Classification

Abstract: Multi-label few-shot learning (MLFSL) remains a significant challenge in medical image analysis (MIA). Current metric-based meta-learning methods face two critical limitations in MIA. First, conventional prototype generation often entangles irrelevant disease information, leading to contaminated prototypes and degraded performance. Second, prior studies typically enforce inter-class separability in embedding space, largely neglecting the inherent correlations among diseases. To overcome these challenges, we propose Prototype Purification and Regulation (PPR), a novel MLFSL framework for MIA. PPR first performs prototype purification by leveraging sample-level comorbidity scores to emphasize disease-specific features, producing purified prototypes that better characterize each disease. Building upon these purified prototypes, PPR further addresses the underexplored problem of inter-class prototype distance in MIA by incorporating disease-level comorbidity statistics to adaptively regulate inter-class similarity, forming a comorbidity-aware embedding space. Overall, PPR sequentially enables the model to capture pure disease features and inter-class relationships for reliable MLFSL in MIA. Extensive experiments across four chest X-ray benchmark datasets, including cross-domain evaluation, show that PPR consistently outperforms state-of-the-art methods, significantly improving disease detection while demonstrating robust generalization and clinical applicability.

Fri 18 SeptComputer Vision and Pattern RecognitionMachine Learning
The gist
Identifying multiple diseases from medical images is hard, especially when only a few examples of each disease are available. The paper presents a new method called Prototype Purification and Regulation (PPR) that cleans up disease representations and respects how diseases often appear together. This helps the model recognize diseases more accurately and reliably. The authors tested PPR on chest X-ray tasks, showing it works better than existing methods even when applied to different datasets.
Open → 2609.21541v1

Routine blood tests predict infection type better than CRP in children

Routine Blood Tests Outperform CRP for Distinguishing Bacterial From Viral Infection in Children

Abstract: Acute infectious diseases are among the leading causes of medical consultations and hospitalizations in children worldwide. These infections are predominantly caused by viruses or bacteria, yet differentiating between the two remains a common clinical challenge. As a result, pediatricians often default to the safer option of prescribing antibiotics contributing to the growing problem of antimicrobial resistance. The objective is to assess the additional predictive value of CBC towards determining the current infection. This retrospective study used data from 906 pediatric patients aged between 2 and 14 years who were tested positive either for viral or bacterial infection between 2022 and 2026. Inclusion criteria further required availability of CBC results and CRP level measurements. These laboratory parameters as well as age were used as input features for several supervised classification models. Model performance was evaluated using AUC, sensitivity and specificity. The best performing model is XGBoost, which included all features, achieving out of-sample performance of AUC of 81.7% and sensitivity of 70.8%, specificity of 79.2%. All trained models outperform a CRP-based only decision-rule model in terms of AUC. We suggest that the decision to prescribe antibiotics should be based on a number of factors, including but not limited to CBC, some of which are not currently incorporated into routine practice.

Fri 18 SeptMachine Learning
The gist
Distinguishing whether a child's infection is caused by bacteria or viruses is hard, leading doctors to often prescribe antibiotics just in case. The authors found that routine blood tests, specifically complete blood count (CBC), combined with age and CRP levels, predict whether the infection is bacterial or viral more accurately than CRP levels alone. Their best method used a machine learning model that improved detection accuracy and could help reduce unnecessary antibiotic use. This suggests that using multiple blood test markers together is more helpful for diagnosis than relying on one marker.
Open → 2609.21332v1