Retrieval-Augmented Generation (RAG) has shown significant potential in enhancing the performance of large language models (LLMs) in knowledge-intensive NLP tasks. However, in medical Q A, particularly in post-diagnosis diabetes management, it remains unclear whether RAG can improve the quality of responses. We developed the DiabMedQA dataset to evaluate the optimization performance of retrieval-augmented LLMs (RA-LLMs) in various Q A scenarios, encompassing three levels of question complexity and four health management tasks. Our experiments revealed that RA-LLMs achieved higher semantic similarity scores and were generally preferred by clinicians over standard LLMs. In the semantic evaluation, Qwen2.5-14B performed the best, with its BERTScore improving by an average of 10.56
Electrode array migration is a rare condition after cochlear implantation that can impair auditory outcomes but often remains undetected during routine follow-up. We evaluated the clinical utility of an electrical impedance-based, machine learning method for detecting postoperative electrode array migration using routinely acquired telemetry data. In this retrospective analysis, we evaluated impedance telemetry data from 32 cochlear implant cases (16 with radiologically confirmed electrode array migration and 16 matched stable controls). A random forest regression model was trained to estimate the longitudinal position of the most basal electrode based on transimpedance matrices, cochlear dimensions, and implant parameters. Model predictions were compared against radiological ground truth, and migration was detected using a defined threshold of 1 mm. Electrode array migration was detected with a sensitivity of 88
Medical documentation, including reply letters from specialty to primary care, imposes a substantial burden on clinicians. Recent advances in large language models (LLMs) may help alleviate this burden. This study aims to develop an instruction-tuned LLM to generate reply letters by integrating five types of electronic medical record documents: admission notes, progress notes, procedure notes, operative notes, and discharge summaries. A total of 8,527 hospitalization episodes from the cardiology and thoracic surgery departments of a tertiary hospital were used for training, and 105 episodes were allocated to the test set. Mistral-Nemo-Instruct-2407 was fine-tuned using the instruction-tuning dataset generated with GPT-4.1. Model performance was evaluated using a dual framework combining expert assessment and a GPT-based evaluation. Experts rated human-written letters higher in accuracy (4.77 vs. 4.21, p < 0.001) and appropriateness (4.90 vs. 4.63, p < 0.001), whereas GPT-based evaluations rated AI-generated letters higher across all dimensions. A hybrid scoring approach integrating expert and GPT-based evaluations improved score consistency among evaluators; in an exploratory comparison using a common weighting, the largest proportion of cases was classified as comparable between the two letter types across all quality dimensions. Expert ratings for workflow suitability (M = 4.30) indicated positive expectations for practical implementation. These findings suggest that LLM-based reply letter generation may serve as a supportive tool for drafting clinical correspondence, although its safety and impact on documentation burden require prospective evaluation in real-world clinical workflows.
To evaluate the clinical and practical trade-offs between domain adaptation and fine-tuning of large language models (LLMs) for radiology impression generation, with outcomes assessed through blinded expert radiologist review. In this retrospective institutional review board (IRB)-approved study 3,123,836 CT and MR reports were split into domain adaptation (80
