Image Search Series Part 4: Advancing Wound Care with Foundation Models and Context-Aware Retrieval
Introduction Wound assessment and management are central tasks in clinical practice, requiring accurate documentation and timely decision-making. Clinicians and nurses often rely on visual inspection to evaluate wound characteristics such as size, color, tissue composition, and healing progress. However, when seeking comparable cases (e.g., to inform treatment choices, validate assessments, or support education), existing search methods have significant limitations. Traditional keyword-based systems require precise terminology, which may not align with the way wounds are described in practice. Moreover, textual descriptors cannot fully capture the variability of visual wound features, resulting in incomplete or imprecise retrieval. Recent advances in computer vision offer new opportunities to address these challenges through both image classification and image retrieval. Automated classification of wound images into clinically meaningful categories (e.g., wound type, tissue condition, infection status) can support standardized documentation and assist clinicians in making more consistent assessments. In parallel, image retrieval systems enable search based on visual similarity rather than textual input alone, allowing clinicians to query databases directly with wound images and retrieve cases with similar characteristics. Together, these AI-based functionalities have the potential to improve case comparison, facilitate consistent monitoring, and enhance clinical training by providing immediate access to relevant examples and structured decision support. The Data The WoundcareVQA dataset is a new multimodal multilingual dataset for Wound Care Visual Question Answering. The WoundcareVQA dataset is available at https://osf.io/xsj5u/ [1] Table 1 summarizes dataset statistics. WoundcareVQA contains 748 images associated with 447 instances (each instance/query includes one or more images). The dataset is split into training (279 instances, 449 images), validation (105 instances, 147 images), and test (93 instances, 152 images). The training set was annotated by a single expert, the validation set by two annotators, and the test set by three medical doctors. Each query is also labeled with wound metadata, covering seven categories: anatomic location (41 classes), wound type (8), wound thickness (6), tissue color (6), drainage amount (6), drainage type (5), and infection status (3). Table 1: Statistics about the WoundcareVQA Dataset We selected two tasks with the highest inter-annotator agreement: Wound Type Classification and Infection Detection (cf. Table 2). Table 3 lists the classification labels for these tasks. Table 2: Inter-Annotator Agreement in the WoundcareVQA Dataset Table 3: Classification Labels for the Tasks: Infection Detection & Wound Type Classification Methods 1. Foundation-Model-based Image Search This approach relies on an image similarity-based retrieval mechanism using a medical foundation model, MedImageInsight [2-3]. Specifically, it employs a k-nearest neighbors (k-NN) search to identify the top k training images most visually similar to a given query image. The image search system operates in two phases: Index Construction: Embeddings are extracted from all training images using a pretrained vision encoder (MedImageInsight). These embeddings are then indexed to enable efficient and scalable similarity search during retrieval. Query and Retrieval: At inference time, the test image is encoded to produce a query embedding. The system computes the Euclidean distances between this query vector and all indexed embeddings, retrieving the k nearest neighbors with the smallest distances. To address the computational demands of large-scale image datasets, the method leverages FAISS (Facebook AI Similarity Search), an open-source library designed for fast and scalable similarity search and clustering of high-dimensional vectors. 2. Vision-Language Models (VLMs) & Retrieval-Augmented Generation (RAG) We leverage vision-language models (e.g., GPT-4o, GPT-4.1), a recent class of multimodal foundation models capable of jointly reasoning over visual and textual inputs. These models can be used for wound assessment tasks due to their ability to interpret complex visual patterns in medical images while simultaneously understanding medical terminology. We evaluate three settings: Zero-shot: The model predicts directly from the query input without additional examples. Few-shot Prompting: A small number of examples (5) from the training dataset are randomly selected and embedded into the input prompt. These paired images and labels provide contextual cues that guide the model's interpretation of new inputs. Retrieval-Augmented Generation (RAG): The system first retrieves the Top-k visually similar wound images using the MedImageInsight-based image search described above. The language model then reasons over the retrieved examples and their labels to generate the final prediction. The implementation of the MedImageInsight-based image search and the RAG method for the infection detection task is available in our Samples Repository: https://aka.ms/healthcare-ai-examples rag_infection_detection.ipynb Evaluation We computed accuracy scores to evaluate the image search methods (Top-1 and Top-5 with majority vote), GPT-4o and GPT-4.1 models (zero-shot), as well as 5-shot and RAG-based methods. Table 4 reports accuracy for wound type classification and infection detection. Figure 1 presents examples of correct and incorrection predictions. Accuracy Image Search Top-1 Image Search Top-5 + majority vote GPT-4o (2023-07-01) GPT-4o (2024-11-20) GPT4.1 (2025-04-14) GPT4.1 5-shot Prompting GPT-4.1- RAG-5 Wound Type 0.7933 0.8333 0.4671 0.4803 0.5066 0.6118 0.7533 Infection 0.6800 0.7267 0.3947 0.3882 0.375 0.7237 0.7697 Table 4: Accuracy Scores for Wound Type Classification & Infection Detection Figure 1: Examples of Correct and Incorrection Predictions (GPT-4.1-RAG-5 Method) For wound type classification, image search with MedImageInsight embeddings performs best, achieving 0.7933 (Top-1) and 0.8333 (Top-5 + majority vote). GPT models alone perform substantially worse (0.4671-0.6118), while GPT-4.1 with retrieval augmentation (RAG-5), which uses the same MedImageInsight-based image search method to retrieve the Top-5 similar cases, narrows the gap (0.7533) but does not surpass direct image search. This suggests that categorical wound type is more effectively captured by visual similarity than by case-based reasoning with vision-language models. For infection detection, the trend reverses. Image search reaches 0.7267 (Top-5 + majority vote), while RAG-5 achieves the highest accuracy at 0.7697. In this case, the combination of visually similar cases with VLM-based reasoning outperforms both standalone image search and GPT prompting. This indicates that infection assessment depends on contextual or clinical cues that may not be fully captured by visual similarity alone but can be better interpreted when enriched with contextual reasoning over retrieved cases and their associated labels. Overall, these findings highlight complementary strengths: foundation-model-based image search excels at categorical visual classification (wound type), while retrieval-augmented VLMs leverage both visual similarity and contextual reasoning to improve performance on more nuanced tasks (infection detection). A hybrid system integrating both approaches may provide the most robust clinical support. Conclusion This study demonstrates the complementary roles of vision-language models in wound assessment. Image search using foundation-model embeddings shows strong performance on categorical tasks such as wound type classification, where visual similarity is most informative. In contrast, retrieval-augmented generation (RAG-5), which combines image search with case-based reasoning by a vision-language model, achieves the best results for infection detection, highlighting the value of integrating contextual interpretation with visual features. These findings suggest that a hybrid approach, leveraging both direct image similarity and retrieval-augmented reasoning, provides the most robust pathway for clinical decision support in wound care. Image Search Series: Blog Posts & Jupyter Notebooks Image Search Series Part 1: Chest X-ray lookup with MedImageInsight | Microsoft Community Hub 2d_image_search.ipynb Image Search Series Part 2: AI Methods for the Automation of 3D Image Retrieval in Radiology | Microsoft Community Hub 3d_image_search.ipynb Image Search Series Part 3: Foundation Models and Retrieval-Augmented Generation in Dermatology | Microsoft Community Hub Image Search Series Part 4: Advancing Wound Care with Foundation Models and Context-Aware Retrieval | Microsoft Community Hub rag_infection_detection.ipynb Image Search Series Part V: Building Histopathology Image Search with Prov-GigaPath | Microsoft Community Hub 2d_pathology_image_search.ipynb The Microsoft healthcare AI models, including MedImageInsight, are intended for research and model development exploration. The models are not designed or intended to be deployed in clinical settings as-is nor for use in the diagnosis or treatment of any health or medical condition, and the individual models’ performances for such purposes have not been established. You bear sole responsibility and liability for any use of the healthcare AI models, including verification of outputs and incorporation into any product or service intended for a medical purpose or to inform clinical decision-making, compliance with applicable healthcare laws and regulations, and obtaining any necessary clearances or approvals. References Wen-wai Yim, Asma Ben Abacha, Robert Doerning, Chia-Yu Chen, Jiaying Xu, Anita Subbarao, Zixuan Yu, Fei Xia, M Kennedy Hall, Meliha Yetisgen. Woundcarevqa: A Multilingual Visual Question Answering Benchmark Dataset for Wound Care. Journal of Biomedical Informatics, 2025. Noel C. F. Codella, Ying Jin, Shrey Jain, Yu Gu, Ho Hin Lee, Asma Ben Abacha, Alberto Santamaría-Pang, Will Guyman, Naiteek Sangani, Sheng Zhang, Hoifung Poon, Stephanie L. Hyland, Shruthi Bannur, Javier Alvarez-Valle, Xue Li, John Garrett, Alan McMillan, Gaurav Rajguru, Madhu Maddi, Nilesh Vijayrania, Rehaan Bhimai, Nick Mecklenburg, Rupal Jain, Daniel Holstein, Naveen Gaur, Vijay Aski, Jenq-Neng Hwang, Thomas Lin, Ivan Tarapov, Matthew P. Lungren, Mu Wei: MedImageInsight: An Open-Source Embedding Model for General Domain Medical Imaging. CoRR abs/2410.06542 (2024) Model catalog and collections in Azure AI Foundry portal https://learn.microsoft.com/en-us/azure/ai-studio/how-to/model-catalog-overviewImage Search Series Part 2: AI Methods for the Automation of 3D Image Retrieval in Radiology
Introduction As the use of diagnostic 3D images increases, effective management and analysis of these large volumes of data grows in importance. Medical 3D image search systems can play a vital role by enabling clinicians to quickly retrieve relevant or similar images and cases based on the anatomical features and pathologies present in a query image. Unlike traditional 2D imaging, 3D imaging offers a more comprehensive view for examining anatomical structures from multiple planes with greater clarity and detail. This enhanced visualization has potential to assist doctors with improved diagnostic accuracy and more precise treatment planning. Moreover, advanced 3D image retrieval systems can support evidence-based and cohort-based diagnostics, demonstrating an opportunity for more accurate predictions and personalized treatment options. These systems also hold significant potential for advancing research, supporting medical education, and enhancing healthcare services. This blog offers guidance on using Azure AI Foundry and the recently launched healthcare AI models to design and test a 3D image search system that can retrieve similar radiology images from a large collection of 3D images. Along with this blog, we share a Jupyter Notebook with the the 3D image search system code, which you may use to reproduce the experiments presented here or start you own solution. 3D Image Search Notebook: http://aka.ms/healthcare-ai-examples-mi2-3d-image-search It is important to highlight that the models available on the AI Foundry Model Catalog are not designed to generate diagnostic-quality results. Developers are responsible for further developing, testing, and validating their appropriateness for specific tasks and eventually integrating these models into complete systems. The objective of this blog is to demonstrate how this can be achieved efficiently in terms of data and computational resources. The Problem Generally, the problem of 3D image search can be posed as retrieving cross-sectional (CS) imaging series (3D image results) that are similar to a given CS imaging series (query 3D image). Once posited this way, the key question becomes how to define such similarity? In the previous blog of this series, we worked with radiographs of the chest which constrained the notion of "similar" to the similarity between two 2D images, and a certain class of anatomy. In the case of 3D images, we are dealing with a volume of data, and a lot more variations of anatomy and pathologies, which expands the dimensions to consider for similarity; e.g., are we looking for similar anatomy? Similar pathology? Similar exam type? In this blog, we will discuss a technique to approximate the 3D similarity problem through a 2D image embedding model and some amount of supervision to constrain the problem to a certain class of pathologies (lesions) and cast it as "given cross-sectional MRI image , retrieve series with similar grade of lesions in similar anatomical regions". To build a search system for 3D radiology images using a foundation model (MedImageInsight) designed for 2D inputs, we explore the generation of representative 3D embedding vectors for the volumes with the foundation model embeddings of 2D slices to create a vector index from a large collection of 3D images. Retrieving relevant results for a given 3D image then consists in generating a representative 3D image embedding vector for the query image and searching for similar vectors in the index. An overview of this process is illustrated in Figure 1. Figure 1: Overview of the 3D image search process. The Data In the sample notebook that is provided alongside this blog, we use 3D CT images from the Medical Segmentation Decathlon (MSD) dataset [2-3] and annotations from the 3D-MIR benchmark [4]. The 3D-MIR benchmark offers four collections (Liver, Colon, Pancreas, and Lung) of positive and negative examples created from the MSD dataset with additional annotations related to the lesion flag (with/without lesion), and lesion group (1, 2, 3). The lesion grouping focuses on lesion morphology and distribution and considers the number, length, and volume of the lesions to define the three groups. It also adheres to the American Joint Committee on Cancer's Tumor, Node, Metastasis classification system’s recommendations for classifying cancer stages and provides a standardized framework for correlating lesion morphology with cancer stage. We selected the 3D-MIR Pancreas collection. 3D-MIR Benchmark: https://github.com/abachaa/3D-MIR Since the MSD collections only include unhealthy/positive volumes, each 3D-MIR collection was augmented with volumes randomly selected from the other datasets to integrate healthy/negative examples in the training and test splits. For instance, the Pancreas dataset was augmented using volumes from the Colon, Liver, and Lung datasets. The input images consist of CT volumes and associated 2D slices. The training set is used to create the index, and the test set is used to query and evaluate the 3D search system. 3D Image Retrieval Our search strategy, called volume-based retrieval, relies on aggregating the embeddings of the 2D slices of a volume to generate one representative 3D embedding vector for the whole volume. We describe additional search strategies in our 3D-MIR paper [4]. The 2D slice embeddings are generated using the MedImageInsight foundation model [5-6] from Azure AI Foundry model catalog [1]. In the search step, we generate the embeddings of the 3D query volumes according to the selected Aggregation method (Agg) and search for the top-k similar volumes/vectors in the corresponding 3D (Agg) index. We use the Median aggregation method to generate the 3D vectors and create the associated 3D index. We construct a 3D (Median) index using the training slices/volumes from the 3D-MIR Pancreas collection. Three other aggregation methods are available in the 3D image search notebook: Max Pooling, Average Pooling, and Standard Deviation. The search is performed following the k-Nearest Neighbors algorithm (or k-NN search) to find the k nearest neighbors of a given vector by calculating the distances between the query vector and all other vectors in the collection, then selecting the K vectors with the shortest distances. If the collection is large, the computation can be expensive, and it is recommended to use specific libraries for optimization. We use the FAISS (Facebook AI Similarity Search) library, an open-source library for efficient similarity search and clustering of high-dimensional vectors. Evaluation of the search results The 3D-MIR Pancreas test set consists of 32 volumes: 4 volumes with no lesion (lesion flag/group= -1) 3 volumes with lesion group 1 19 volumes with lesion group 2 6 volumes with lesion group 3 The training set consists of 269 volumes (with and without lesions) and was used to create the index. We evaluate the 3D search system by comparing the lesion group/category of the query volume and the top 10 retrieved volumes. We then compute Precision@k (P@k). Table 1 presents the P@1, P@3, P@5, P@10, and overall Precision. Table 1: Evaluation results on the 3D-MIR Pancreas test set The system accurately recognizes Healthy cases, consistently retrieving the correct label in test scenarios involving non-lesion pancreas images. However, performance varies for different lesion groups, reflecting challenges in precisely identifying smaller lesions (Group 1) or more advanced lesions (Group 3). This discrepancy highlights the complexity of lesion detection and underscores the importance of carefully tuning embeddings or adjusting the vector index to improve retrieval accuracy for specific lesion sizes. Visualization Figure 2 presents four different test queries from the Pancreas test set and the top 5 nearest neighbors retrieved by the volume-based search method. In each row, the first image is the query, followed by the retrieved images ranked by similarity. The visual overlays help in assessing retrieval accuracy; Blue indicates the pancreas organ boundaries, and Red highlights the mark regions corresponding to the pancreas tumor. Figure 2: Top 5 results for different queries from the Pancreas test set Table 2 presents additional results of the volume-based retrieval system [4] on other 3D-MIR datasets/organs (Liver, Colon, and Lung) using additional foundation models: BiomedCLIP [7], Med-Flamingo [8], and BiomedGPT [9]. When considering the macro-average across all datasets, MedImageInsight-based retrieval outperforms substantially other foundation models. Table 2: Evaluation Results on the 3D-MIR benchmark (Liver, Colon, Pancreas, and Lung) These results mirror a use case akin to lesion detection and severity measurement in a clinical context. In real-world applications—such as diagnostic support or treatment planning—it may be necessary to optimize the model to account for particular goals (e.g., detecting critical lesions early) or accommodate different imaging protocols. By refining search criteria, integrating more domain-specific data, or adjusting embedding methods, practitioners can enhance retrieval precision and better meet clinical requirements. Conclusion The integration of 3D image search systems in clinical environment can enhance and accelerate the retrieval of similar cases and provide better context to clinicians and researchers for accurate complex diagnoses, cohort selection, and personalized patient care. This 3D radiology image search blog and related notebook offers a solution based on 3D embedding generation for building and evaluating a 3D image search system using the MedImageInsight foundation model from Azure AI Foundry model catalog. Image Search Series: Blog Posts & Jupyter Notebooks Image Search Series Part 1: Chest X-ray lookup with MedImageInsight | Microsoft Community Hub 2d_image_search.ipynb Image Search Series Part 2: AI Methods for the Automation of 3D Image Retrieval in Radiology | Microsoft Community Hub 3d_image_search.ipynb Image Search Series Part 3: Foundation Models and Retrieval-Augmented Generation in Dermatology | Microsoft Community Hub Image Search Series Part 4: Advancing Wound Care with Foundation Models and Context-Aware Retrieval | Microsoft Community Hub rag_infection_detection.ipynb Image Search Series Part V: Building Histopathology Image Search with Prov-GigaPath | Microsoft Community Hub 2d_pathology_image_search.ipynb The Microsoft healthcare AI models, including MedImageInsight, are intended for research and model development exploration. The models are not designed or intended to be deployed in clinical settings as-is nor for use in the diagnosis or treatment of any health or medical condition, and the individual models’ performances for such purposes have not been established. You bear sole responsibility and liability for any use of the healthcare AI models, including verification of outputs and incorporation into any product or service intended for a medical purpose or to inform clinical decision-making, compliance with applicable healthcare laws and regulations, and obtaining any necessary clearances or approvals. References Model catalog and collections in Azure AI Foundry portal https://learn.microsoft.com/en-us/azure/ai-studio/how-to/model-catalog-overview Michela Antonelli et al. The medical segmentation decathlon. Nature Communications, 13(4128), 2022 https://www.nature.com/articles/s41467-022-30695-9 MSD: http://medicaldecathlon.com/ Asma Ben Abacha, Alberto Santamaría-Pang, Ho Hin Lee, Jameson Merkow, Qin Cai, Surya Teja Devarakonda, Abdullah Islam, Julia Gong, Matthew P. Lungren, Thomas Lin, Noel C. F. Codella, Ivan Tarapov: 3D-MIR: A Benchmark and Empirical Study on 3D Medical Image Retrieval in Radiology. CoRR abs/2311.13752, 2023 https://arxiv.org/abs/2311.13752 Noel C. F. Codella, Ying Jin, Shrey Jain, Yu Gu, Ho Hin Lee, Asma Ben Abacha, Alberto Santamaría-Pang, Will Guyman, Naiteek Sangani, Sheng Zhang, Hoifung Poon, Stephanie Hyland, Shruthi Bannur, Javier Alvarez-Valle, Xue Li, John Garrett, Alan McMillan, Gaurav Rajguru, Madhu Maddi, Nilesh Vijayrania, Rehaan Bhimai, Nick Mecklenburg, Rupal Jain, Daniel Holstein, Naveen Gaur, Vijay Aski, Jenq-Neng Hwang, Thomas Lin, Ivan Tarapov, Matthew P. Lungren, Mu Wei: MedImageInsight: An Open-Source Embedding Model for General Domain Medical Imaging. CoRR abs/2410.06542, 2024 https://arxiv.org/abs/2410.06542 MedImageInsight: https://aka.ms/mi2modelcard Sheng Zhang, Yanbo Xu, Naoto Usuyama, Hanwen Xu, Jaspreet Bagga, Robert Tinn, Sam Preston, Rajesh Rao, Mu Wei, Naveen Valluri, Cliff Wong, Andrea Tupini, Yu Wang, Matt Mazzola, Swadheen Shukla, Lars Liden, Jianfeng Gao, Angela Crabtree, Brian Piening, Carlo Bifulco, Matthew P. Lungren, Tristan Naumann, Sheng Wang, Hoifung Poon. BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image-text pairs. NEJM AI 2025; 2(1) https://ai.nejm.org/doi/full/10.1056/AIoa2400640 Moor, M., Huang, Q., Wu, S., Yasunaga, M., Dalmia, Y., Leskovec, J., Zakka, C., Reis, E.P., Rajpurkar, P.: Med-flamingo: a multimodal medical few-shot learner. Machine Learning for Health, ML4H@NeurIPS 2023, 10 December 2023, New Orleans, Louisiana, USA. Proceedings of Machine Learning Research, vol. 225, pp. 353–367. PMLR, (2023) https://proceedings.mlr.press/v225/moor23a.html Zhang, K., Zhou, R., Adhikarla, E., Yan, Z., Liu, Y., Yu, J., Liu, Z., Chen, X., Davison, B.D., Ren, H., et al.: A generalist vision–language foundation model for diverse biomedical tasks. Nature Medicine, 1–13 (2024) https://www.nature.com/articles/s41591-024-03185-2Image Search Series Part 3: Foundation Models and Retrieval-Augmented Generation in Dermatology
Introduction Dermatology is inherently visual, with diagnosis often relying on morphological features such as color, texture, shape, and spatial distribution of skin lesions. However, the diagnostic process is complicated by the large number of dermatologic conditions, with over 3,000 identified entities, and the substantial variability in their presentation across different anatomical sites, age groups, and skin tones. This phenotypic diversity presents significant challenges, even for experienced clinicians, and can lead to diagnostic uncertainty in both routine and complex cases. Image-based retrieval systems represent a promising approach to address these challenges. By enabling users to query large-scale image databases using a visual example, these systems can return semantically or visually similar cases, offering useful reference points for clinical decision support. However, dermatology image search is uniquely demanding. Systems must exhibit robustness to variations in image quality, lighting, and skin pigmentation while maintaining high retrieval precision across heterogeneous datasets. Beyond clinical applications, scalable and efficient image search frameworks provide valuable support for research, education, and dataset curation. They enable automated exploration of large image repositories, assist in selecting challenging examples to enhance model robustness, and promote better generalization of machine learning models across diverse populations. In this post, we continue our series on using healthcare AI models in Azure AI Foundry to create efficient image search systems. We explore the design and implementation of such a system for dermatology applications. As a baseline, we first present an adapter-based classification framework for dermatology images by leveraging fixed embeddings from the MedImageInsight foundation model, available in the Azure AI Foundry model catalog. We then introduce a Retrieval-Augmented Generation (RAG) method that enhances vision-language models through similarity-based in-context prompting. We use the MedImageInsight foundation model to generate image embeddings and retrieve the top-k visually similar training examples via FAISS. The retrieved image-label pairs are included in the Vision-LLM prompt as in-context examples. This targeted prompting guides the model using visually and semantically aligned references, enhancing prediction quality on fine-grained dermatological tasks. It is important to highlight that the models available on the AI Foundry Model Catalog are not designed to generate diagnostic-quality results. Developers are responsible for further developing, testing, and validating their appropriateness for specific tasks and eventually integrating these models into complete systems. The objective of this blog is to demonstrate how this can be achieved efficiently in terms of data and computational resources. The Data The DermaVQA-IIYI [2] dermatology image dataset is a de-identified, diverse collection of nearly 1,000 patient records and nearly 3,000 dermatological images, created to support research in skin condition recognition, classification, and visual question answering. DermaVQA-IIYI dataset: https://osf.io/72rp3/files/osfstorage (data/iiyi) The dataset is split into three subsets: Training Set: 2,474 images associated with 842 patient cases Validation Set: 157 images associated with 56 cases Test Set: 314 images associated with 100 cases Total Records: 2,945 images (998 patient cases) Patient Demographics: Out of 998 patient cases: Sex – F: 218, M: 239, UNK: 541 Age (available for 398 patients): Mean: 31 yrs | Min: 0.08 yrs | Max: 92 yrs This wide range supports studies across all age groups, from infants to the elderly. A total of 2,945 images are associated with the patient records, with an average of 2.9 images per patient. This multiplicity enables the study of skin conditions from different perspectives and at various stages. Image Count per Entry: 1 image: 225 patients 2 images: 285 patients 3 images: 200 patients 4 or more images: 288 patients The dataset includes additional annotations for anatomic location, comprising 39 distinct labels (e.g., back, fingers, fingernail, lower leg, forearm, eye region, unidentifiable). Each image is associated with one or multiple labels. We use these annotations to evaluate the performance of various methods across different anatomical regions. Image Embeddings We generate image embeddings using the MedImageInsight foundation model [1] from the Azure AI Foundry model catalog [3]. We apply Uniform Manifold Approximation and Projection (UMAP) to project high-dimensional image embeddings produced by the MedImageInsight model into two dimensions. The visualization is generated using embeddings extracted from both the DermaVQA training and test sets, which covers 39 anatomical regions. For clarity, only the most frequent anatomical labels are displayed in the projection. Figure 1. UMAP projection of image embeddings produced by the MedImageInsight Model on the DermaVQA dataset. The resulting projection reveals that the MedImageInsight model captures meaningful anatomical distinctions: visually distinct regions such as fingers, face, fingernail, and foot form well-separated clusters, indicating high intra-class consistency and inter-class separability. Other anatomically adjacent or visually similar regions, such as back, arm, and abdomen, show moderate overlap, which is expected due to shared visual features or potential labeling ambiguity. Overall, the embeddings exhibit a coherent and interpretable organization, suggesting that the model has learned to encode both local and global anatomical structures. This supports the model’s effectiveness in capturing anatomy-specific representations suitable for downstream tasks such as classification and retrieval. Enhancing Visual Understanding We explore two strategies for enhancing visual understanding through foundation models. I. Training an Adapter-based Classifier We build an adapter-based classification framework designed for efficient adaptation to medical imaging tasks (see our prior posts for introduction into the topic of adapters: Unlocking the Magic of Embedding Models: Practical Patterns for Healthcare AI | Microsoft Community Hub). The proposed adapter model builds upon fixed visual features extracted from the MedImageInsight foundation model, enabling task-specific fine-tuning without requiring full model retraining. The architecture consists of three main components: MLP Adapter: A two-layer feedforward network that projects 1024-dimensional embeddings (generated by the MedImageInsight model) into a 512-dimensional latent space. This module utilizes GELU activation and Layer Normalization to enhance training stability and representational capacity. As a bottleneck adapter, it facilitates parameter-efficient transfer learning. Convolutional Retrieval Module: A sequence of two 1D convolutional layers with GELU activation, applied to the output of the MLP adapter. This component refines the representations by modeling local dependencies within the transformed feature space. Prediction Head: A linear classifier that maps the 512-dimensional refined features to the task-specific output space (e.g., 39 dermatology classes). The classifier is trained for 10 epochs (approximately 48 seconds) using only CPU resources. Built on fixed image embeddings extracted from the MedImageInsight model, the adapter efficiently tailors these representations for downstream classification tasks with minimal computational overhead. By updating only the adapter components, while keeping the MedImageInsight backbone frozen, the model significantly reduces computational and memory overhead. This design also mitigates overfitting, making it particularly effective in medical imaging scenarios with limited or imbalanced labeled data. A Jupyter Notebook detailing the construction and training of an MedImageInsight -based adapter model is available in our Samples Repository: https://aka.ms/healthcare-ai-examples-mi2-adapter Figure 3: MedImageInsight-based Adapter Model II. Boosting Vision-Language Models with in-Context Prompting We leverage vision-language models (e.g., GPT-4o, GPT-4.1), which represent a recent class of multimodal foundation models capable of jointly reasoning over visual and textual inputs. These models are particularly promising for dermatology tasks due to their ability to interpret complex visual patterns in medical images while simultaneously understanding domain-specific medical terminology. 1. Few-shot Prompting In this setting, a small number of examples from the training dataset are randomly selected and embedded into the input prompt. These examples, consisting of paired images and corresponding labels, are intended to guide the model's interpretation of new inputs by providing contextual cues and examples of relevant dermatological features. 2. MedImageInsight-based Retrieval-Augmented Generation (RAG) This approach enhances vision-language model performance by integrating a similarity-based retrieval mechanism rooted in MedImageInsight (Medical Image-to-Image) comparison. Specifically, it employs a k-nearest neighbors (k-NN) search to identify the top k dermatological training images that are most visually similar to a given query image. The retrieved examples, consisting of dermatological images and their corresponding labels, are then used as in-context examples in the Vision-LLM prompt. By presenting visually similar cases, this approach provides the model with more targeted contextual references, enabling it to generate predictions grounded in relevant visual patterns and associated clinical semantics. As illustrated in Figure 2, the system operates in two phases: Index Construction: Embeddings are extracted from all training images using a pretrained vision encoder (MedImageInsight). These embeddings are then indexed to enable efficient and scalable similarity search during retrieval. Query and Retrieval: At inference time, the test image is encoded similarly to produce a query embedding. The system computes the Euclidean distance between this query vector and all indexed embeddings, retrieving the k nearest neighbors with the smallest distances. To handle the computational demands of large-scale image datasets, the method leverages FAISS (Facebook AI Similarity Search), an open-source library designed for fast and scalable similarity search and clustering of high-dimensional vectors. The implementation of the image search method is available in our Samples Repository: https://aka.ms/healthcare-ai-examples-mi2-2d-image-search Figure 2: MedImageInsight-based Retrieval-Augmented Generation Evaluation Table 1 presents accuracy scores for anatomic location prediction on the DermaVQA-iiyi test set using the proposed modeling approaches. The adapter model achieves a baseline accuracy of 31.73%. Vision-language models perform better, with GPT-4o (2024-11-20) achieving an accuracy of 47.11%, and GPT-4.1 (2025-04-14) improving to 50%. However, incorporating few-shot prompting with five randomly selected in-context examples (5-shot) slightly reduces GPT-4.1’s performance to 48.72%. This decline suggests that unguided example selection may introduce irrelevant or low-quality context, potentially reducing the effectiveness of the model’s predictions for this specialized task. The best performance among the vision-language approaches is achieved using the retrieval-augmented generation (RAG) strategy. In this setup, GPT-4.1 is prompted with five nearest-neighbor examples retrieved using the MedImageInsight-based search method (RAG-5), leading to a notable accuracy increase to 51.60%. This improvement over GPT-4.1’s 50% accuracy without retrieval showcases the relevance of the MedImageInsight-based RAG method. We expect larger performance gains when using a more extensive dermatology dataset, compared to the relatively small dataset used in this example -- a collection of 2,474 images associated with 842 patient cases which served as the basis for selecting relevant cases and similar images. Dermatology is a particularly challenging domain, marked by a high number of distinct conditions and significant variability in skin tone, texture, and lesion appearance. This diversity makes robust and representative example retrieval especially critical for enhancing model performance. The results underscore the importance of example relevance in few-shot prompting, demonstrating that similarity-based retrieval can effectively guide the model toward more accurate predictions in complex visual reasoning tasks. Table 1: Comparative Accuracy of Anatomic Location Prediction on DermaVQA-iiyi Figure 2: Confusion Matrix of Anatomical Location Predictions by the trained MLP adapter: The matrix illustrates the model's performance in classifying wound images across 39 anatomical regions. Strong diagonal values indicate correct classifications, while off-diagonal entries highlight common misclassifications, particularly among anatomically adjacent or visually similar regions such as 'lowerback' vs. 'back' and 'hand' vs. 'fingers'. Figure 3. Examples of correct anatomical predictions by the RAG approach. Each image depicts a case where the model's predicted anatomical region exactly matches the ground truth. Shown are examples from visually and anatomically distinct areas including the eye region, lips, lower leg, and neck. Figure 4. Examples of misclassifications by the RAG approach. Each image displays a case where the predicted anatomical label differs from the ground truth. In several examples, predictions are anatomically close to the correct regions (e.g., hand vs. hand-back, lower leg vs. foot, palm vs. fingers), suggesting that misclassifications often occur between adjacent or visually similar areas. These cases highlight the challenge of precise localization in fine-grained anatomical classification and the importance of accounting for anatomical ambiguity in both modeling and evaluation. Conclusion Our exploration of scalable image retrieval and advanced prompting strategies demonstrates the growing potential of vision-language models in dermatology. A particularly challenging task we address is anatomic location prediction, which involves 39 fine-grained classes of dermatology images, imbalanced training data, and frequent misclassifications between adjacent or visually similar regions. By leveraging Retrieval-Augmented Generation (RAG) with similarity-based example selection using image embeddings from the MedImageInsight foundation model, we show that relevant contextual guidance can significantly improve model performance in such complex settings. These findings underscore the importance of intelligent image retrieval and prompt construction for enhancing prediction accuracy in fine-grained medical tasks. As vision-language models continue to evolve, their integration with retrieval mechanisms and foundation models holds substantial promise for advancing clinical decision support, medical research, and education at scale. In the next blog of this series, we will shift focus to the wound care subdomain of dermatology, and we will release accompanying Jupyter notebooks for the adapter-based and RAG-based methods to provide a reproducible reference implementation for researchers and practitioners. The Microsoft healthcare AI models, including MedImageInsight, are intended for research and model development exploration. The models are not designed or intended to be deployed in clinical settings as-is nor for use in the diagnosis or treatment of any health or medical condition, and the individual models’ performances for such purposes have not been established. You bear sole responsibility and liability for any use of the healthcare AI models, including verification of outputs and incorporation into any product or service intended for a medical purpose or to inform clinical decision-making, compliance with applicable healthcare laws and regulations, and obtaining any necessary clearances or approvals. Image Search Series: Blog Posts & Jupyter Notebooks Image Search Series Part 1: Chest X-ray lookup with MedImageInsight | Microsoft Community Hub 2d_image_search.ipynb Image Search Series Part 2: AI Methods for the Automation of 3D Image Retrieval in Radiology | Microsoft Community Hub 3d_image_search.ipynb Image Search Series Part 3: Foundation Models and Retrieval-Augmented Generation in Dermatology | Microsoft Community Hub Image Search Series Part 4: Advancing Wound Care with Foundation Models and Context-Aware Retrieval | Microsoft Community Hub rag_infection_detection.ipynb Image Search Series Part V: Building Histopathology Image Search with Prov-GigaPath | Microsoft Community Hub 2d_pathology_image_search.ipynb The Microsoft healthcare AI models, including MedImageInsight, available in the Microsoft Foundry model catalog, are intended for research and model development exploration. The models are not designed or intended to be deployed in clinical settings as-is nor for use in the diagnosis or treatment of any health or medical condition, and the individual models’ performances for such purposes have not been established. You bear sole responsibility and liability for any use of the healthcare AI models, including verification of outputs and incorporation into any product or service intended for a medical purpose or to inform clinical decision-making, compliance with applicable healthcare laws and regulations, and obtaining any necessary clearances or approvals. References Noel C. F. Codella, Ying Jin, Shrey Jain, Yu Gu, Ho Hin Lee, Asma Ben Abacha, Alberto Santamaría-Pang, Will Guyman, Naiteek Sangani, Sheng Zhang, Hoifung Poon, Stephanie L. Hyland, Shruthi Bannur, Javier Alvarez-Valle, Xue Li, John Garrett, Alan McMillan, Gaurav Rajguru, Madhu Maddi, Nilesh Vijayrania, Rehaan Bhimai, Nick Mecklenburg, Rupal Jain, Daniel Holstein, Naveen Gaur, Vijay Aski, Jenq-Neng Hwang, Thomas Lin, Ivan Tarapov, Matthew P. Lungren, Mu Wei: MedImageInsight: An Open-Source Embedding Model for General Domain Medical Imaging. CoRR abs/2410.06542 (2024) Wen-wai Yim, Yujuan Fu, Zhaoyi Sun, Asma Ben Abacha, Meliha Yetisgen, Fei Xia: DermaVQA: A Multilingual Visual Question Answering Dataset for Dermatology. MICCAI (5) 2024: 209-219 Model catalog and collections in Azure AI Foundry portal https://learn.microsoft.com/en-us/azure/ai-studio/how-to/model-catalog-overview
Security Copilot Clinic: AI‑Driven Agentic Defense for Healthcare
Healthcare security teams are operating under unprecedented pressure. Ransomware continues to target clinical environments, identity‑based attacks are increasing in sophistication, and the risk of PHI exposure remains a constant concern — all while SOC teams face chronic staffing shortages. Microsoft Security Copilot is now available for organizations using Microsoft 365 E5, bringing generative AI assistance directly into the security tools healthcare teams already rely on. This clinic series is designed to show how Security Copilot changes day‑one operations — turning noisy alerts into clear, actionable investigations and faster containment. Why attend this clinic For healthcare CISOs, SOC leaders, and security architects, Security Copilot represents more than an AI assistant — it’s a shift in how investigations are conducted across endpoint, identity, email, data, and cloud workloads. In this session, you’ll see how Security Copilot helps healthcare security teams: Move faster with confidence by summarizing complex evidence across security signals Reduce investigation fatigue by standardizing analyst workflows Communicate risk clearly by translating technical findings into leadership‑ready insights Protect patient data without adding new tools or headcount All examples and demonstrations are grounded in real healthcare security scenarios. What we’ll explore See the full incident picture in one place Microsoft‑built Security Copilot agents embedded across Defender, Entra, Intune, and Purview automatically correlate signals from endpoint, identity, email, data, and cloud applications into a single investigation view — eliminating manual pivoting between tools. Move from alert to action faster Embedded agents analyze related signals in real time and surface prioritized investigation paths along with recommended containment actions directly in the analyst workflow. Standardize investigations and reduce noise Agent‑driven prompts and investigation structure help standardize analyst response, reduce alert fatigue, and create repeatable workflows that scale in lean SOC environments. Protect PHI and communicate risk with confidence Security Copilot uses embedded data and threat intelligence to produce leadership‑ready summaries that clearly articulate potential PHI exposure, attack progression, and business impact. Session format and audience Format 60‑minute live session End‑to‑end demo Interactive Q&A Who should attend CISOs and Security Leaders SOC Managers and Analysts Security and Cloud Architects Clinical IT and Infrastructure Leaders Upcoming sessions Date Time (ET) Registration March 13, 2026 12:00 – 1:00 PM Session #1 March 20, 2026 12:00 – 1:00 PM Session #2 March 27, 2026 12:00 – 1:00 PM Session #3 Secure healthcare — together Security Copilot enables healthcare organizations to respond faster, investigate smarter, and communicate risk more effectively — all within the Microsoft security ecosystem teams already trust. If you’re evaluating how AI‑driven, agentic defense can support your healthcare SOC, this clinic will give you practical insight you can apply immediately.Dragon Copilot centralizes trusted medical content and relevant contextual information in-workflow
This blog is co-authored by Bert Hoorne, Principal Program Manager & Ksenya Kveler, Principle Medical Science Manager Dragon Copilot delivers medical intelligence from trusted sources directly within clinical workflows for healthcare organizations in one solution. We are pleased to announce that we are expanding those knowledge sources with additional best‑in‑class content providers and enabling broader access to your organization’s internal sources with Microsoft 365 Copilot integration. Access information from new credible medical content providers Dragon Copilot users will gain access to an additional robust collection of trusted clinical content from leading evidence-based resources. We are partnering with renowned publishers to bring you the best, most trusted content, safely and securely, within clinician’s workflows while helping to reduce the use of unauthorized AI tools and applications, commonly referred to, as “shadow AI.” Access content from Wolters Kluwer UpToDate We’ve partnered with Wolters Kluwer UpToDate to bring trusted, evidence-based clinical guidance directly into Dragon Copilot. Customers with an active Wolters Kluwer UpToDate license will be able to access UpToDate content in Dragon Copilot, within the context of their clinical workflows. This integration allows clinicians to ask both general questions and patient specific questions and receive answers grounded in UpToDate evidence, with clear references to supporting sources. Over time, it will also introduce contextual links to UpToDate concepts layered on top of Dragon Copilot–generated notes, further enhancing clinical insight at the point of care. “Clinicians need reliable guidance that supports fast, confident decision-making without disrupting care delivery. We are excited to partner with Microsoft to bring UpToDate’s gold standard evidence and expertise-based clinical insights to Dragon Copilot, helping clinicians quickly access, actionable answers that reduce cognitive burden and support better patient care.” Yaw Fellin, Senior Vice President and General Manager, UpToDate Clinical Decision Support and Provider Solutions Wolters Kluwer Health Here’s an example of UpToDate content embedded in the Dragon Copilot workflow: Obtain trusted clinical evidence with Elsevier ClinicalKey AI Elsevier’s ClinicalKey AI will be available in Dragon Copilot. This integration enables customers with an active Elsevier ClinicalKey AI license to surface trusted medical literature and clinical evidence directly within clinicians’ workflows. “Clinicians are navigating a complex and rapidly changing healthcare landscape and need solutions they can trust. The ClinicalKey AI extension for Dragon Copilot transforms how clinicians interact with trusted medical literature and clinical answers. The conversational interface makes evidence discovery faster and more intuitive.” Jukka Valimaki, SVP Clinical Solutions Elsevier Here’s an example of ClinicalKey AI content embedded in the Dragon Copilot workflow: Support clinical decisions with EBMcalc With the integration of EBMcalc medical calculators, Dragon Copilot enables clinicians to use evidence-based calculators directly within their workflows—applied in context to the patient they’re caring for. “Clinicians need trusted, evidence-based insights exactly at the point of care. By integrating EBMcalc’s rigorously curated clinical calculators and references into Dragon Copilot, we’re helping make high quality medical evidence more accessible, more actionable, and easier to use within everyday clinical workflows”. Louis Leff, MD, MACP, Founder and CEO EBMcalc Access independent evidence in Dragon Copilot with Wiley and Cochrane Wiley and Microsoft are partnering to bring scientific literature and clinical evidence directly into the healthcare workflow, starting with the Cochrane Library. Through this integration, customers with an active Cochrane Library AI license will be able to access Cochrane’s high-quality, independent evidence, systematic reviews, and clinical answers, to inform more reliable and efficient decision-making. This includes the Cochrane Database of Systematic Reviews (CDSR), the home of gold-standard evidence syntheses, widely used to inform clinical guidelines worldwide. "Working with Microsoft to bring the Cochrane Library into Dragon Copilot reflects a shared commitment to meeting researchers and clinicians where they are. Healthcare Institutions can now access independent, peer-reviewed evidence— right within their clinical workflow” Josh Jarrett, SVP & GM of AI Growth Wiley Access work context with Microsoft 365 Copilot in Dragon Copilot With the Microsoft 365 Copilot integration, Dragon Copilot enables clinicians to seamlessly access information from their emails, chats, OneDrive and SharePoint, within the flow of their clinical work. Clinicians can combine this information with additional questions and actions, all governed by existing organizational and user access controls. Use of this data within Dragon Copilot workflow remains fully at the user’s discretion. Here’s an example of content from an email surfaced by Microsoft 365 Copilot accessible through the Dragon Copilot workflow: Read more for a deeper dive on how Dragon Copilot enables work context access with Microsoft 365 Copilot integration. Safe web search Dragon Copilot safe web search delivers trusted, evidence linked answers when curated sources are unavailable—ensuring clinicians continue to receive timely support without disrupting their workflow. The goal of safe web search is to prevent broken workflows and eliminate unsafe external browsing. Clinicians remain within their clinical context, focused on the patient—without tab hopping or the risk of landing on unreliable or unverified websites. Safe web search eliminates “no response” dead ends by maintaining a seamless conversational experience in Dragon Copilot and reducing unanswered prompts. This capability is enabled by using verified, secure, and responsible mechanisms designed for safe clinical experiences. It enforces multilayer protection through evidence validation, provenance linked responses, content filtering, and regulated search with built in safeguards. Here’s an example of content from a safe web search in the Dragon Copilot workflow: Conclusion These advancements represent an important step forward in how Dragon Copilot delivers trusted medical intelligence - bringing together best‑in‑class clinical evidence, organizational knowledge, and safe web access in one governed, in‑workflow experience. We will continue to expand our partner ecosystem, deepen integrations with leading evidence providers, and evolve Dragon Copilot conversational extensibility to meet clinicians where they work.Bringing Organizational Knowledge into the Clinical Workflow
This blog is co-authored by Hadas Bitran, Partner GM, Health AI, Microsoft Health & Life Sciences Every day, clinicians spend valuable time looking for information that lives in different places. An email thread from a specialist colleague. A Microsoft Teams discussion about a complex case. Updated organizational processes buried in SharePoint or OneDrive. This information provides context that could be critical to their workflows or help inform their decisions. But that context is not part of their clinical workflow. The result? Clinicians are forced to break their clinical workflow, searching manually across organizational resources, and mentally combining scattered data points, all while a patient is waiting. This isn't a knowledge problem. It's a retrieval problem. And it's costing time, focus, cognitive burden and clinical confidence every single day. That's exactly the gap we're closing by bringing clinical intelligence and your organization's knowledge into one seamless, workflow-native experience. Clinical workflow, now with your organizational context Within Dragon Copilot, clinicians will be able to securely surface relevant information across Microsoft 365, without leaving the clinical workflow: Email: retrieve relevant information that was exchanged with patients, colleagues or from specialist correspondence, referral communications, or care coordination threads. find me the email from Dr. Ting that mentioned the latest research about this mutation. In this example, the chat functionality in Dragon Copilot uses the patient and encounter context to resolve the referenced mutation, then leverages Microsoft 365 Copilot behind the scenes to locate the email from Dr. Ting that mentions it. Microsoft Teams: surface information from Microsoft Teams chats that the clinician had with colleagues, discussions or group chat conversations. The patient is traveling to Florida. Identify dialysis centers near the patient’s destination based on information shared by Dr. Salomon in Microsoft Teams and provide practical travel guidelines I can share with the patient. In this example, Dragon Copilot uses trusted sources for travel guidelines and Microsoft 365 Copilot to retrieve relevant Microsoft Teams messages from Dr. Salomon, identifying nearby dialysis centers in Florida. SharePoint and OneDrive: access organizational knowledge on demand: HR policies, facility procedures, compliance guidelines, shift schedules, and more Who is on call for nephrology tonight and who is covering tomorrow morning? In this example, Dragon Copilot leverages Microsoft 365 Copilot behind the scenes to locate the most up‑to‑date Excel file with upcoming shift and coverage information from the hospital’s SharePoint, and surfaces the answer directly in the conversation, without disrupting the clinician’s workflow. With Microsoft 365 Copilot, work context is available directly inside Dragon Copilot, clinicians can choose if, and when to access their work information. Within Dragon Copilot, they can ask questions in natural language and receive the most relevant information, grounded in patient context, from trusted clinical sources and their Microsoft 365 data. One conversational flow. Full clinical and work context. No tab switching, no manual searching, no lost focus. Trusted by design, built for healthcare Security and privacy are built in from the ground up. Information is always accessed on behalf of the individual user, fully respecting existing Microsoft 365 identity and access management, compliance, and privacy controls, meaning clinicians see only what they're authorized to see, and that Dragon Copilot will only use their work context if the clinician consented to it. This also means no new security risks to manage, and no changes to how your organization governs access to information. For healthcare organizations where data sensitivity, regulatory compliance, and patient privacy are non-negotiable, this better-together experience is designed to meet that bar from day one. Join the Private Preview If you're a Dragon Copilot customer, and your organization is using Microsoft 365 Copilot, we invite you to be among the first to experience this new capability. Register now for early access to the private preview and play a role in shaping the future of clinical workflow intelligence. Register for private preview
Why nursing needs a different kind of AI—and how Dragon Copilot delivers
The Dragon Copilot experience for nurses was made generally available (GA) in December 2025 with a clear mission: help nursing staff focus on care, not the computer. From the start, the goal was to create a comprehensive AI clinical assistant—one that works alongside nurses throughout their shift, reduces cognitive load, captures the full scope of care delivered, and translates real clinical work into automated next steps, including documentation—fundamentally transforming workflows to keep patient care at the center. Microsoft has continued to execute on that vision. Recent enhancements include extended mobile access with Android support—enabling nurses to record care in Epic Rover on Android devices—as well as significant expansion in ambient documentation coverage. Together, these advances reflect a consistent approach: adoption follows when technology aligns with how nurses work. Expansive nursing documentation coverage Nursing work spans multiple flowsheet templates, assessments, state changes, and, at times, narrative notes. When solutions support only a subset of this work, nurses are left filling gaps after the fact—reintroducing cognitive load and eroding the value of this technology. Microsoft has expanded Dragon Copilot’s ambient documentation capabilities by broadening the range of supported nursing value types—and by extending it to deliver complete coverage across all flowsheet templates in supported departments and settings. The result is comprehensive documentation generated from each recording including: Lines, Drains, Airways, and Wounds (LDAs) documentation, including assessments, additions, and removals Nurse notes, automatically generated from natural nurse-patient conversations and voice memos captured on the go Full flowsheet template coverage—not just a subset—including admission and discharge flowsheets, blood administration, CIWA-Ar, and care plan-related flowsheets Adaptations to each organizations charting philosophy, including macros support, chart-by-exception, pertinent positives, and more This breadth matters because nursing work is rarely captured within only a narrow set of flowsheets—nor does it typically result in just narrative notes. Yet many solutions labeled “for nurses” prioritize what is easiest to automate, rather than what nurses need. The result can be a false sense of completeness, with nurses still managing gaps across their shift. Why nursing ambient documentation is hard—and what makes Dragon Copilot unique Achieving comprehensive, high‑quality nursing documentation has required specialized technology designed to address the structural, workflow, and feedback challenges unique to nursing—challenges that general narrative ambient models and physician‑oriented solutions are not built to solve: Flowsheets are messy, complex, and frequently changing Flowsheets are large, hospital-specific, internally ambiguous, and constantly evolving under governance. Complex logic—such as cascading rows, documentation‑by‑exception patterns, and duplicative or overlapping rows—makes it far from straightforward to accurately map a clinical observation to the correct field and value. Microsoft works directly with real hospital schemas, handling hierarchy, ambiguity, and multiple valid documentation destinations—without requiring flowsheet redesign or sacrificing quality. Nurses don’t speak for documentation Bedside language is optimized for care delivery, not chart completeness. Critical details are often implied or never spoken aloud. Microsoft’s technology translates natural nursing communication into accurate documentation without changing nurse behavior. Built on industry‑leading transcription accuracy and decades of speech recognition expertise, Dragon Copilot is informed by real‑world integration across diverse EHR environments, preserving accurate translation and clinical intent that directly impact downstream documentation accuracy. Nursing audio is diverse Recordings mix shorthand, dialogue, monologue, and unit-specific language. Dragon Copilot accounts for mixed speaking modes instead of flattening audio through a generic pipeline or requiring nurses to speak in constrained ways. Feedback loops are noisy Nurse corrections to AI output often reflect hindsight or personal preferences rather than model error. Microsoft’s approach analyzes correction patterns with clinical context, enabling calibration at the institution, department, and even individual user level. Bedside workflows demand predictability Baseline LLMs are not suited for real-world nursing accuracy, latency, and cost requirements — especially with tens-of-thousands of possible flowsheet values. Dragon Copilot is optimized for consistent performance across real nursing environments, exceeding the reliability and latency characteristics of baseline models. Beyond specialized nursing architecture, Dragon Copilot enforces strict quality and safety gates for new documentation outputs—including oversight by Microsoft’s internal, nurse-led Clinical Integrity team, phased validation, and Responsible AI review—ensuring new documentation covered meets defined nursing standards before being introduced at scale. Dragon Copilot represents a fundamental shift in how nursing work is supported by AI by meeting the full complexity of bedside care head-on. By delivering comprehensive ambient documentation across live inpatient care environments, Dragon Copilot helps ensure that the care nurses provide is accurately captured, trusted, and usable downstream. The result is an AI clinical assistant that keeps nurses focused on what matters most: their patients.Can you use AI to implement an Enterprise Master Patient Index (EMPI)?
The Short Answer: Yes. And It's Better Than You Think. If you've worked in healthcare IT for any length of time, you've dealt with this problem. Patient A shows up at Hospital 1 as "Jonathan Smith, DOB 03/15/1985." Patient B shows up at Hospital 2 as "Jon Smith, DOB 03/15/1985." Patient C shows up at a clinic as "John Smythe, DOB 03/15/1985." Same person? Probably. But how do you prove it at scale — across millions of records, dozens of source systems, and data quality that ranges from pristine to "someone fat-fingered a birth year"? That's the problem an Enterprise Master Patient Index (EMPI) solves. And traditionally, it's been solved with expensive commercial products, rigid rule engines, and a lot of manual review. We built one with AI. On Azure. With open-source tooling. And the results are genuinely impressive. This post walks through how it works, what the architecture looks like, and why the combination of deterministic matching, probabilistic algorithms, and AI-enhanced scoring produces better results than any single approach alone. 1. Why EMPI Still Matters (More Than Ever) Healthcare organizations don't have a "patient data problem." They have a patient identity problem. Every EHR, lab system, pharmacy platform, and claims processor creates its own patient record. When those systems exchange data via FHIR, HL7, or flat files, there's no universal patient identifier in the U.S. — Congress has blocked funding for one since 1998. The result: Duplicate records inflate costs and fragment care history Missed matches mean clinicians don't see a patient's full medical picture False positives can merge two different patients into one record — a patient safety risk Traditional EMPI solutions use deterministic matching (exact field comparisons) and sometimes probabilistic scoring (fuzzy string matching). They work. But they leave a significant gray zone of records that require human review — and that queue grows faster than teams can process it. What if AI could shrink that gray zone? 2. The Architecture: Three Layers of Matching Here's the core insight: no single matching technique is sufficient. Exact matches miss typos. Fuzzy matches produce false positives. AI alone hallucinates. But layer them together with calibrated weights, and you get something remarkably accurate. Let's break each layer down. 3. Layer 1: Deterministic Matching — The Foundation Deterministic matching is the bedrock. If two records share an Enterprise ID, they're the same person. Full stop. The system assigns trust levels to each identifier type: Identifier Weight Why Enterprise ID 1.0 Explicitly assigned by an authority SSN 0.9 Highly reliable when present and accurate MRN 0.8 System-dependent — only valid within the same healthcare system Date of Birth 0.35 Common but not unique — 0.3% of the population shares any given birthday Phone 0.3 Useful signal but changes frequently Email 0.3 Same — supportive evidence, not proof The key implementation detail here is MRN system validation. An MRN of "12345" at Hospital A is completely unrelated to MRN "12345" at Hospital B. The system checks the identifier's source system URI before considering it a match. Without this, you'd get a flood of false positives from coincidental MRN collisions. If an Enterprise ID match is found, the system short-circuits — no need for probabilistic or AI scoring. It's a guaranteed match. 4. Layer 2: Probabilistic Matching — Where It Gets Interesting This is where the system earns its keep. Probabilistic matching handles the messy reality of healthcare data: typos, nicknames, transposed digits, abbreviations, and inconsistent formatting. Name Similarity The system uses a multi-algorithm ensemble for name matching: Jaro-Winkler (60% weight): Optimized for short strings like names. Gives extra credit when strings share a common prefix — so "Jonathan" vs "Jon" scores higher than you'd expect. Soundex / Metaphone (phonetic boost): Catches "Smith" vs "Smythe," "Jon" vs "John," and other sound-alike variations that string distance alone would miss. Levenshtein distance (typo detection): Handles single-character errors — "Johanson" vs "Johansn." These scores are blended, and first name and last name are scored independently before combining. This prevents a matching last name from compensating for a wildly different first name. Date of Birth — Smarter Than You'd Think DOB matching goes beyond exact comparison. The system detects month/day transposition — one of the most common data entry errors in healthcare: Scenario Score Exact match 1.0 Month and day swapped (e.g., 03/15 vs 15/03) 0.8 Off by 1 day 0.9 Off by 2–30 days 0.5–0.8 (scaled) Different year 0.0 This alone catches a category of mismatches that pure deterministic systems miss entirely. Address Similarity Address matching uses a hybrid approach: Jaro-Winkler on the normalized full address (70% weight) Token-based Jaccard similarity (30% weight) to handle word reordering Bonus scoring for matching postal codes, city, and state Abbreviation expansion — "St" becomes "Street," "Ave" becomes "Avenue" 5. Layer 3: AI-Enhanced Matching — The Game Changer This is where the architecture diverges from traditional EMPI solutions. OpenAI Embeddings (Semantic Similarity) The system generates a text embedding for each patient's complete demographic profile using OpenAI's text-embedding-3-small model. Then it computes cosine similarity between patient pairs. Why does this work? Because embeddings capture semantic relationships that string-matching can't. "123 Main Street, Apt 4B, Springfield, IL" and "123 Main St #4B, Springfield, Illinois" are semantically identical even though they differ character-by-character. The embedding score carries only 10% of the total weight — it's a signal, not a verdict. But in ambiguous cases, it's the signal that tips the scale. GPT-5.2 LLM Analysis (Intelligent Reasoning) For matches that land in the human review zone (0.65–0.85), the system optionally invokes GPT-5.2 to analyze the patient pair and provide structured reasoning: { "match_score": 0.92, "confidence": "high", "reasoning": "Multiple strong signals: identical last name, DOB matches exactly, same city. First name 'Jon' is a common nickname for 'Jonathan'.", "name_analysis": "First name variation is a known nickname pattern.", "potential_issues": [], "recommendation": "merge" } The LLM doesn't just produce a number — it explains why it thinks two records match. This is enormously valuable for the human reviewers who make final decisions on ambiguous cases. Instead of staring at two records and guessing, they get AI-generated reasoning they can evaluate. When LLM analysis is enabled, the final score blends traditional and LLM scores: Final Score = (Traditional Score × 0.8) + (LLM Score × 0.2) The LLM temperature is set to 0.1 for consistency — you want deterministic outputs from your matching engine, not creative ones. 6. The Graph Database: Modeling Patient Relationships Records and scores are only half the story. The real power comes from how the system stores and traverses relationships. We use Azure Cosmos DB with the Gremlin API — a graph database that models patients, identifiers, addresses, and clinical data as vertices connected by typed edges. (:Patient)──[:HAS_IDENTIFIER]──▶(:Identifier) │ ├──[:HAS_ADDRESS]──▶(:Address) │ ├──[:HAS_CONTACT]──▶(:ContactPoint) │ ├──[:LINKED_TO]──▶(:EmpiRecord) ← Golden Record │ ├──[:POTENTIAL_MATCH {score, confidence}]──▶(:Patient) │ └──[:HAS_ENCOUNTER]──▶(:Encounter) └──[:HAS_OBSERVATION]──▶(:Observation) Why a Graph? Three reasons: Candidate retrieval is a graph traversal problem. "Find all patients who share an identifier with Patient X" is a natural graph query — traverse from the patient to their identifiers, then back to other patients who share those same identifiers. In Gremlin, this is a few lines. In SQL, it's a multi-table join with performance that degrades as data grows. Relationships are first-class citizens. A POTENTIAL_MATCH edge stores the match score, confidence level, and detailed breakdown directly on the relationship. You can query "show me all high-confidence matches" without any joins. EMPI records are naturally hierarchical. A golden record (EmpiRecord) links to multiple source patients via LINKED_TO edges. When you merge two patients, you're adding an edge — not rewriting rows in a relational table. Performance at Scale Cosmos DB's partition strategy uses source_system as the partition key, providing logical isolation between healthcare systems. The system handles Azure's 429 rate-limiting with automatic retry and exponential backoff, and uses batch operations for bulk loads to avoid RU exhaustion. 7. FHIR-Native Data Ingestion The system ingests HL7 FHIR R4 Bundles — the emerging interoperability standard for healthcare data exchange. Each FHIR Bundle is a JSON file containing a complete patient record: demographics, encounters, observations, conditions, procedures, immunizations, medication requests, and diagnostic reports. The FHIR loader: Maps FHIR identifier systems to internal types (SSN, MRN, Enterprise ID) Handles all three FHIR date formats (YYYY, YYYY-MM, YYYY-MM-DD) Extracts clinical data for comprehensive patient profiles Uses an iterator pattern for memory-efficient processing of thousands of patients Tracks source system provenance for audit compliance This means the service can ingest data directly from any FHIR-compliant EHR — Epic, Cerner, MEDITECH, or Synthea-generated test data — without custom integration work. 8. The Conversational Agent: Matching via Natural Language Here's where it gets fun. The system includes a conversational AI agent built on the Azure AI Foundry Agent Service. It's deployed as a GPT-5.2-powered agent with OpenAPI tools that call the matching service's REST API. Instead of navigating a complex UI to find matches, a data steward can simply ask: "Search patients named Aaron" "Compare patient abc-123 with patient xyz-456" "What matches are pending review?" "Approve the match between patient A and patient B" The agent is integrated directly into the Streamlit dashboard's Agent Chat tab, so users never leave their workflow. Under the hood, when the agent decides to call a tool (like "search patients"), Azure AI Foundry makes an HTTP request directly to the Container App API — no local function execution required. Available Agent Tools Tool What It Does searchPatients Search patients by name, DOB, or identifier getPatientDetails Get detailed patient demographics and history findPatientMatches Find potential duplicates for a patient compareTwoPatients Side-by-side comparison with detailed scoring getPendingReviews List matches awaiting human decision submitReviewDecision Approve or reject a match getServiceStatistics MPI dashboard metrics This same tool set is also exposed via a Model Context Protocol (MCP) server, making the matching engine accessible from AI-powered IDEs and coding assistants. 9. The Dashboard: Putting It All Together The Patient Matching Service includes a full-featured Streamlit dashboard for operational management. Page What You See Dashboard Key metrics, score distribution charts, recent match activity Match Results Filterable list with score breakdowns — deterministic, probabilistic, AI, and LLM tabs Patients Browse and search all loaded patients with clinical data Patient Graph Interactive graph visualization of patient relationships using streamlit-agraph Review Queue Pending matches with approve/reject actions Agent Chat Conversational AI for natural language queries Settings Configure match weights, thresholds, and display preferences The match detail view provides six tabs that walk reviewers through every scoring component: Summary, Deterministic, Probabilistic, AI/Embeddings, LLM Analysis, and Raw Data. Reviewers don't just see a number — they see exactly why the system scored a match the way it did. 10. Azure Architecture The full solution runs on Azure: Service Role Azure Cosmos DB (Gremlin + NoSQL) Patient graph storage and match result persistence Azure OpenAI (GPT-5.2 + text-embedding-3-small) LLM analysis and semantic embeddings Azure Container Apps Hosts the FastAPI REST API Azure AI Foundry Agent Service Conversational agent with OpenAPI tools Azure Log Analytics Centralized logging and monitoring The separation between Cosmos DB's Gremlin API (graph traversal) and NoSQL API (match result documents) is intentional. Graph queries excel at relationship traversal — "find all patients connected to this identifier." Document queries excel at filtering and aggregation — "show me all auto-merge matches from the last 24 hours." 11. What We Learned AI doesn't replace deterministic matching. It augments it. The three-layer approach works because each layer compensates for the others' weaknesses: Deterministic handles the easy cases quickly and with certainty Probabilistic catches the typos, nicknames, and formatting differences that exact matching misses AI provides semantic understanding and human-readable reasoning for the ambiguous middle ground The LLM is most valuable as a reviewer's assistant, not a decision-maker. We deliberately keep the LLM weight at 20% of the final score. Its real value is the structured reasoning it produces — the "why" behind a match score. Human reviewers process cases faster when they have AI-generated analysis explaining the matching signals. Graph databases are naturally suited for patient identity. Patient matching is fundamentally a relationship problem. "Who shares identifiers with whom?" "Which patients are linked to this golden record?" "Show me the cluster of records that might all be the same person." These are graph traversal queries. Trying to model this in relational tables works, but you're fighting the data model instead of leveraging it. FHIR interoperability reduces integration friction to near zero. By accepting FHIR R4 Bundles as the input format, the service can ingest data from any modern EHR without custom connectors. This is a massive practical advantage — the hardest part of any EMPI project is usually getting the data in, not matching it. 12. Try It Yourself The Patient Matching Service is built entirely on Azure services and open-source tooling https://github.com/dondinulos/patient-matching-service : Python with FastAPI, Streamlit, and the Azure AI SDKs Azure Cosmos DB (Gremlin API) for graph storage Azure OpenAI for embeddings and LLM analysis Azure AI Foundry for the conversational agent Azure Container Apps for deployment Synthea for FHIR test data generation The matching algorithms (Jaro-Winkler, Soundex, Metaphone, Levenshtein) use pure Python implementations — no proprietary matching engines required. Whether you're building a new EMPI from scratch or augmenting an existing one with AI capabilities, the three-layer approach gives you the best of all worlds: the certainty of deterministic matching, the flexibility of probabilistic scoring, and the intelligence of AI-enhanced analysis. Final Thoughts Can you use AI to implement an EMPI? Yes. And the answer isn't "replace everything with an LLM." It's "use AI where it adds the most value — semantic understanding, natural language reasoning, and augmenting human reviewers — while keeping deterministic and probabilistic matching as the foundation." The combination is more accurate than any single approach. The graph database makes relationships queryable. The conversational agent makes the system accessible. And the whole thing runs on Azure with FHIR-native data ingestion. Patient matching isn't a solved problem. But with AI in the stack, it's a much more manageable one. Tags: Healthcare, Azure, AI, EMPI, FHIR, Patient Matching, Azure Cosmos DB, Azure OpenAI, Graph Database, Interoperability
