MedTech software for diagnostic platforms

Exciting Innovation in Healthcare AI: MedRAG I just came across a groundbreaking paper introducing MedRAG - a novel approach that enhances Retrieval-Augmented Generation (RAG) with Knowledge Graph-Elicited Reasoning for healthcare applications! Diagnostic errors are a serious problem in healthcare, with approximately 795,000 individuals suffering permanent disability or death annually due to misdiagnosis in the US alone. MedRAG addresses this challenge by significantly improving the accuracy and specificity of AI-powered diagnostic support. >> How MedRAG Works: The system combines RAG with a comprehensive four-tier hierarchical diagnostic knowledge graph to enhance reasoning capabilities. Here's the technical breakdown: 1. Diagnostic Knowledge Graph Construction: MedRAG systematically builds a four-tier hierarchical diagnostic KG through disease clustering, hierarchical aggregation, and LLM augmentation. This captures critical diagnostic differences between diseases with similar manifestations. 2. Diagnostic Differences KG Searching: When a patient's manifestations are input, the system decomposes them into clinical features, embeds them, and matches them with relevant diagnostic differences through multi-level matching and upward traversal within the KG. 3. KG-elicited Reasoning RAG: The system retrieves relevant Electronic Health Records (EHRs) and integrates them with the identified diagnostic differences KG to trigger reasoning in a large language model, generating precise diagnoses and treatment recommendations. 4. Proactive Diagnostic Questioning: MedRAG can identify when patient information is insufficient and proactively suggest follow-up questions based on discriminability scores of features in the knowledge graph. The researchers evaluated MedRAG on both a public dataset (DDXPlus) and a private chronic pain diagnostic dataset from Tan Tock Seng Hospital. It outperformed state-of-the-art RAG models, achieving up to 11.32% improvement in diagnostic accuracy for diseases with similar manifestations. What's particularly impressive is MedRAG's compatibility across various backbone LLMs, including open-source models like Mixtral-8x7B and Llama-3.1-Instruct, as well as closed-source models like GPT-4o. This technology has tremendous potential to reduce misdiagnosis rates and improve healthcare outcomes by providing more accurate, specific diagnostic support and personalized treatment recommendations.

Check out recently released healthcare foundation models from Microsoft: MedImageInsight - Embedding model for advanced image analysis, including classification and similarity search in medical imaging. - Streamlines workflows across radiology, pathology, ophthalmology, dermatology, and other modalities. - Researchers can use embeddings directly or build adapters for specific tasks. - Enables tools to automatically route imaging scans to specialists or flag potential abnormalities. - Enhances efficiency and patient outcomes. - Supports Responsible AI safeguards like out-of-distribution detection and drift monitoring to maintain stability and reliability. CXRReportGen - Multimodal AI model for generating detailed, structured reports from chest X-rays. - Incorporates current and prior images along with key patient information. - Highlights AI-generated findings directly on images to align with human-in-the-loop workflows. - Accelerates turnaround times while enhancing diagnostic precision. - Supports diagnosis of a wide range of conditions—from lung infections to heart problems. - Addresses the most common radiology procedure globally. MedImageParse - Precise image segmentation model covering X-rays, CT scans, MRIs, ultrasounds, dermatology images, and pathology slides. - Can be fine-tuned for specific applications like tumor segmentation or organ delineation. - Enables developers to build AI tools for sophisticated medical image analysis.

Microsoft has developed the 𝐌𝐢𝐜𝐫𝐨𝐬𝐨𝐟𝐭 𝐀𝐈 𝐃𝐢𝐚𝐠𝐧𝐨𝐬𝐭𝐢𝐜 𝐎𝐫𝐜𝐡𝐞𝐬𝐭𝐫𝐚𝐭𝐨𝐫 (𝐌𝐀𝐈-𝐃𝐱𝐎), a breakthrough system that achieves 85% diagnostic accuracy on complex medical cases from the New England Journal of Medicine - more than four times higher than experienced physicians who averaged only 20% accuracy. 𝐊𝐞𝐲 𝐁𝐫𝐞𝐚𝐤𝐭𝐡𝐫𝐨𝐮𝐠𝐡 - MAI-DxO functions as a virtual panel of physicians, using sequential diagnosis to iteratively ask questions and order tests rather than relying on simple multiple-choice answers. - System was tested against 304 complex NEJM cases using their new Sequential Diagnosis Benchmark (SD Bench). 𝐒𝐮𝐩𝐞𝐫𝐢𝐨𝐫 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 - Diagnostic accuracy: 85.5% when paired with OpenAI's o3 model vs. 20% for human physicians - Cost efficiency: Achieves higher accuracy while spending less on diagnostic tests - Broad capability: Combines both specialist depth and generalist breadth that no single physician can match 𝐑𝐞𝐚𝐥-𝐖𝐨𝐫𝐥𝐝 𝐈𝐦𝐩𝐚𝐜𝐭 - With over 50 million health-related AI sessions daily across Microsoft products, this technology addresses critical healthcare challenges including rising costs (approaching 20% of US GDP) and diagnostic delays. - System could reduce the estimated 25% waste in healthcare spending. 𝐂𝐮𝐫𝐫𝐞𝐧𝐭 𝐒𝐭𝐚𝐭𝐮𝐬 - Remains research-only and requires rigorous safety testing, clinical validation, and regulatory approval before clinical deployment. - Microsoft emphasizes AI will complement, not replace, physicians by automating routine tasks while doctors focus on patient relationships and complex decision-making. - Technology represents a significant step toward medical AI that can handle medicine's most challenging diagnostic cases more accurately and cost-effectively than current approaches.

Artificial intelligence makes strides in specialized diagnostics but faces challenges in complex clinical scenarios, such as rare disease diagnosis and emergency condition identification. To address these limitations, we develop Meta General Practitioner (MetaGP), a 32-billion-parameter generative foundation model trained on extensive datasets, including over 8 million electronic health records, biomedical literature, and medical textbooks. MetaGP demonstrates robust diagnostic capabilities, achieving accuracy comparable to experienced clinicians. In rare disease cases, it achieves an average diagnostic score of 1.57, surpassing GPT-4’s 0.93. For emergency conditions, it improves diagnostic accuracy for junior and mid-level clinicians by 53% and 46%, respectively. MetaGP also excels in generating medical imaging reports, producing high-quality outputs for chest X-rays and computed tomography, often rated comparable to or superior to physician-authored reports. These findings highlight MetaGP’s potential to transform clinical decision-making across diverse medical contexts. Interesting medical NLP publication by @Fei Liu and larger team: https://lnkd.in/ent52p5d Additional information on companies in the medical NLP space for those interested: Hippocratic AI John Snow Labs Insilico Medicine AION Labs GenBio AI DeepSeek AI Verily Life Sciences Cerebras Systems Movano Health Microsoft BioNTech SE and InstaDeep OpenAI Anthropic Perplexity Google DeepMind Genentech 

Last month, a founder called us with a surprise: their AI-based diagnostic tool had just been classified as an In Vitro Diagnostic (IVD) under the IVDR. Their first reaction: “But it’s just software!” It’s a moment we’ve seen often. The line between software and diagnostics is disappearing fast, and for many innovators, this means their app, algorithm, or data platform is now subject to one of Europe’s most stringent regulations. IVDR doesn’t just redefine compliance; it redefines accountability. Software that interprets biological or diagnostic data must now show clinical performance, cybersecurity robustness, and traceability, just like traditional test kits or analyzers. This shift is catching companies off guard, not because the rules are unclear, but because they assume “code” operates outside the lab. In our reviews, three mistakes keep showing up: ❌ Treating software validation like IT testing and not clinical evidence. ❌ Weak traceability between code logic, intended use, and risk controls. ❌ Missing cybersecurity and PMS plans for continuous updates. Here’s how we guide teams stepping into this “Software as IVD” frontier: ✅ Clearly define the intended purpose and diagnostic claim. ✅ Map algorithm outputs to clinical relevance. ✅ Validate analytical + clinical performance, not just functionality. ✅ Align lifecycle with IEC 62304 and IVDR Annex XIII. ✅ Set up PMS to capture real-world software performance. As the IVDR era unfolds, every diagnostic startup must ask: Is our software just functioning, or is it clinically proven to perform? #IVDR #SaMD #DigitalHealth #IVDSoftware #RegulatoryStrategy #MedTechLeadership #Elexes #AI #SAMD #IVD #Regulatory #Medicaldevice