Knowledge-Enriched Structured EHR Features for 30-Day Hospital Readmission Prediction on MIMIC-IV
该研究通过结合四种医学知识源增强结构化电子健康记录特征，不依赖临床笔记预测30天内再入院情况，减少计算成本并提高解释性。
0 citationsRead paper
Selected work
Representative Papers
When is 3D Worth It? A Resource-Performance Frontier for CNNs and Transformers in Lung CT
This study systematically evaluates whether the additional computational cost of 3D models is justified over 2D or 2.5D approaches in pulmonary CT analysis. Under a unified training protocol, the authors conduct controlled experiments comparing convolutional neural networks (CNNs) and Vision Transformers (ViTs) across 2D, 2.5D, and 3D input representations using the NLST (n=1,977) and LIDC-IDRI datasets, assessing performance, stability, and resource consumption. The work introduces the first joint dimension–architecture evaluation framework tailored for lung cancer screening, revealing that 3D CNNs suffer from threshold instability and ViTs are prone to degenerate predictions such as all-positive outputs. Results demonstrate that 2.5D CNNs achieve the optimal trade-off between discriminative capability and stability (ROC-AUC 0.682), highlighting the practical advantages of lower-dimensional models in real-world clinical deployment.
0 citationsRead paper
Integrating Machine Learning Ensembles and Large Language Models for Heart Disease Prediction Using Voting Fusion
This study addresses the limited reliability of single-model approaches in early cardiovascular disease prediction by proposing a hybrid architecture that integrates ensemble learning with large language models (LLMs). The framework combines mainstream ensemble methods—XGBoost, LightGBM, CatBoost, and Random Forest—with open-source LLMs such as Gemini 2.5 Flash through a voting mechanism, leveraging the OpenRouter API to enable zero-shot and few-shot inference. Evaluated under uncertainty, the proposed method achieves a prediction accuracy of 96.62% and an AUC of 0.97, substantially outperforming standalone ensemble models (95.78%) and LLMs alone (maximum 78.9%). This work presents the first empirical validation of an ML–LLM voting fusion strategy for clinical decision support, demonstrating its enhanced robustness and efficacy in real-world diagnostic scenarios.
0 citationsRead paper
Continuous monitoring of delayed outcomes in basket trials
This study addresses the challenge of delayed treatment effects in basket trials—particularly with immunotherapies—where conventional interim analyses struggle to timely discontinue ineffective arms, and existing Bayesian approaches are computationally intensive and ill-suited for delayed outcomes. The authors propose a computationally efficient, continuous monitoring framework that, for the first time, integrates Bayesian empirical methods with multiple imputation to adaptively select the optimal strategy for handling missing data, thereby enabling effective interim decision-making under outcome delay. Simulation results demonstrate that when accrual is slow and missingness is minimal, simple approaches suffice; however, in complex settings involving multiple baskets and agents, the proposed method substantially improves sample utilization and overall trial efficiency.
0 citationsRead paper
In-Context Learning for Label-Efficient Cancer Image Classification in Oncology
Clinical deployment of AI in oncology is hindered by scarcity of annotated data and high computational costs associated with model retraining. Method: This work presents the first systematic evaluation of multimodal vision-language models (VLMs)—including Paligemma, CLIP, ALIGN, and GPT-4o—in few-shot in-context learning (ICL) for tumor pathology image diagnosis, without parameter updates or fine-tuning. Experiments are conducted across multiple real-world tumor pathology datasets under zero-shot and few-shot ICL settings. Contribution/Results: GPT-4o achieves F1 scores of 0.81 (binary classification) and 0.60 (multiclass classification); notably, lightweight open-source VLMs (e.g., Paligemma) attain comparable performance, underscoring their viability in low-resource clinical settings. The study demonstrates that ICL enables expert-level tumor classification using only a handful of annotated examples—eliminating the need for retraining—and establishes an efficient, lightweight, and generalizable paradigm for diagnosing rare cancers.
0 citationsRead paper
Recent publications
Latest Papers
Knowledge-Enriched Structured EHR Features for 30-Day Hospital Readmission Prediction on MIMIC-IV
该研究通过结合四种医学知识源增强结构化电子健康记录特征，不依赖临床笔记预测30天内再入院情况，减少计算成本并提高解释性。
0 citationsRead paper
When is 3D Worth It? A Resource-Performance Frontier for CNNs and Transformers in Lung CT
This study systematically evaluates whether the additional computational cost of 3D models is justified over 2D or 2.5D approaches in pulmonary CT analysis. Under a unified training protocol, the authors conduct controlled experiments comparing convolutional neural networks (CNNs) and Vision Transformers (ViTs) across 2D, 2.5D, and 3D input representations using the NLST (n=1,977) and LIDC-IDRI datasets, assessing performance, stability, and resource consumption. The work introduces the first joint dimension–architecture evaluation framework tailored for lung cancer screening, revealing that 3D CNNs suffer from threshold instability and ViTs are prone to degenerate predictions such as all-positive outputs. Results demonstrate that 2.5D CNNs achieve the optimal trade-off between discriminative capability and stability (ROC-AUC 0.682), highlighting the practical advantages of lower-dimensional models in real-world clinical deployment.
0 citationsRead paper
Integrating Machine Learning Ensembles and Large Language Models for Heart Disease Prediction Using Voting Fusion
This study addresses the limited reliability of single-model approaches in early cardiovascular disease prediction by proposing a hybrid architecture that integrates ensemble learning with large language models (LLMs). The framework combines mainstream ensemble methods—XGBoost, LightGBM, CatBoost, and Random Forest—with open-source LLMs such as Gemini 2.5 Flash through a voting mechanism, leveraging the OpenRouter API to enable zero-shot and few-shot inference. Evaluated under uncertainty, the proposed method achieves a prediction accuracy of 96.62% and an AUC of 0.97, substantially outperforming standalone ensemble models (95.78%) and LLMs alone (maximum 78.9%). This work presents the first empirical validation of an ML–LLM voting fusion strategy for clinical decision support, demonstrating its enhanced robustness and efficacy in real-world diagnostic scenarios.
0 citationsRead paper
Continuous monitoring of delayed outcomes in basket trials
This study addresses the challenge of delayed treatment effects in basket trials—particularly with immunotherapies—where conventional interim analyses struggle to timely discontinue ineffective arms, and existing Bayesian approaches are computationally intensive and ill-suited for delayed outcomes. The authors propose a computationally efficient, continuous monitoring framework that, for the first time, integrates Bayesian empirical methods with multiple imputation to adaptively select the optimal strategy for handling missing data, thereby enabling effective interim decision-making under outcome delay. Simulation results demonstrate that when accrual is slow and missingness is minimal, simple approaches suffice; however, in complex settings involving multiple baskets and agents, the proposed method substantially improves sample utilization and overall trial efficiency.
0 citationsRead paper
In-Context Learning for Label-Efficient Cancer Image Classification in Oncology
Clinical deployment of AI in oncology is hindered by scarcity of annotated data and high computational costs associated with model retraining. Method: This work presents the first systematic evaluation of multimodal vision-language models (VLMs)—including Paligemma, CLIP, ALIGN, and GPT-4o—in few-shot in-context learning (ICL) for tumor pathology image diagnosis, without parameter updates or fine-tuning. Experiments are conducted across multiple real-world tumor pathology datasets under zero-shot and few-shot ICL settings. Contribution/Results: GPT-4o achieves F1 scores of 0.81 (binary classification) and 0.60 (multiclass classification); notably, lightweight open-source VLMs (e.g., Paligemma) attain comparable performance, underscoring their viability in low-resource clinical settings. The study demonstrates that ICL enables expert-level tumor classification using only a handful of annotated examples—eliminating the need for retraining—and establishes an efficient, lightweight, and generalizable paradigm for diagnosing rare cancers.
0 citationsRead paper