🚨 Pulse of Innovation 🚨

Breaking news in the healthcare AI

Power of the CHIP: How NVIDIA is powering Healthcare AI developments

The NVIDIA is ranked 3rd in the world by valuation and has dedicated NVIDIA Healthcare division that focuses on the AI powered Healthcare solutions that not only include the powerful chips but also involve data processing, software development and many more things. The common theme in all these is NVIDIA’s proactive strategy in collaborating with big and small company’s to bring AI innovations at the forefront that really benefit the patients. We cover several of these initiatives by NVIDIA today.

Big Tech Partnership

NVIDIA has extended the partnership with Microsoft to better integrate its MONAI (Medical Open Network for AI), and the Nuance Precision Imaging Network (PIN) to enable the development, validation, deployment of various services such as clinical research and drug discovery, enhance medical image-based diagnostic technology, and increase access to precision medicine.

Meta’s Llama open-source LLM has been optimized using NVIDIA to Llama NIM platform that is available as open-source and utilized by healthcare start-ups sucha as QuantiPhi, Mendel AI, Activ Surgical etc. to enhance their work in clinical research, augmented reality and documentation.

Big Pharma Partnership

NVIDIA has partnered with Johnson and Johnson MedTech to provide AI powered solutions to the operating room. This includes better staffing, work-flow improvements as well as automated documentation intra and post procedure.

NVIDIA has also partnered with the GE Healthcare and created SonoSAMtrack, an ultrasound platform for better segmentation of anatomies, lesion and other important features.

Startup partnership

Hippocratic AI and Abridge AI are focused on NVIDIA powered AI platform to generate virtual GenAI health agent that engages patients in multiple areas to better understand and respond to their healthcare needs, improving the health work force burden. Hippocratic and Abridge are different in their areas of focus and patient engagement but use NVIDIA powered GenAI platform.

Summary

Healthcare workforce burnouts is critical and GenAI can help resolve several of these by reducing time spent on documentation, improving diagnostic tools, enhancing patient safety infrastructure.

🧑🏼‍🔬 Bench to Bedside👨🏽‍🔬

Developments in healthcare AI research and innovations

Enhancing Clinical Trials: A Deep Learning Approach to Histopathology-Based Patient Prescreening- The FGFR Device.

Overview:

• Fibroblast Growth Factor Receptor (FGFR) alteration testing identifies patients who may benefit from FGFR-targeted therapies like erdafitinib or pemigatinib for metastatic or locally advanced cancers.

• The recent Nature article details the creation and implementation of an AI-based algorithm - FGFR device, aimed at improving patient prescreening in clinical trials. This algorithm was developed using over 3000 H&E-stained whole slide images from patients with advanced urothelial cancers. It is designed to detect biomarkers with high sensitivity to ensure that eligible patients are not excluded from trials prematurely.

  Key Findings

• Algorithm Performance: The deep learning algorithm demonstrated an area under the curve (AUC) of 0.75, with a sensitivity of 88.7% and specificity of 31.8%. This indicates a strong ability to identify patients who are likely to benefit from further molecular testing.

• Impact on Molecular Testing: The deployment of this algorithm is projected to reduce the need for molecular testing by approximately 28.7%, which can significantly cut down on costs and expedite patient recruitment for clinical trials.

• Validation and Deployment: The algorithm was validated on a dataset of 350 patients and successfully deployed in a non-interventional study across 89 global clinical sites. This deployment highlights the algorithm’s potential in real-world clinical settings to prioritize or deprioritize molecular testing resources.

Conclusion:

•  The software for histopathology-based deep learning in clinical trial prescreening shows promise in accelerating patient selection. Its potential lies in swiftly analyzing large datasets, reducing pathologists’ workload, and expediting trial enrolment.

• Limitations include the necessity for rigorous validation against traditional diagnostic methods in diverse group of patients to ensure reliability and safety.

👩‍⚕️AI in the Clinic 🏥

Real-world and practical use of healthcare AI in clinic

Application of AI in Rural Hospital for improved documentation and faster decision making

• Mike Bluff Hospital is a rural Wisconsin health center and get its patients’ health information from multiple sources and the available data is in variety formats which are structured, unstructured and even scanned copies at times. This makes the compilation of records time consuming and cumbersome.

• The hospital was in a need of a solution that can streamline the process of the data compilation in patient charts smoother and faster. The hospital then implemented the Meditech Expanse AI platform that will sort and compile the patient data from various formats into the hospital EHR system. It also provides Google health search and summarization platform that allows accurate and faster search of the data.

• After initial roll out to physician and advanced practitioners the results showed a decrease in time spent on the documentation by 40%. In addition it reduced the burden on work force by 16 hours per week. This is a classic example of practical use of AI in Healthcare in resource limited environment.

🩺 Start-Up Stethoscope 💵 

Trending startups and technologies impacting clinical practice

Qure.ai : Revolutionizing Healthcare with Deep Learning AI

AI-driven medical imaging solutions that enhance the diagnostic accuracy and speed help to improve patient outcomes globally. Qure.ai is one such company from Mumbai, India on a mission to make healthcare more accessible and affordable using the power of deep learning.

How It Works

• qXR: An AI tool for Chest X- Ray interpretation, capable of assistance in 20 seconds with findings across lungs ( Tuberculosis, lung cancer, nodules, consolidation) heart, pleura, and mediastinum.

  

  qXR for Tuberculosis

  

  qXR for Lung Nodules

• qMSK:  Qure.ai’s AI-powered solution designed for musculoskeletal (MSK) imaging. It focuses on analysing musculoskeletal X-rays to detect and evaluate a wide range of bone and joint abnormalities.

  

  QMSK utility

• qCT: An AI solution for analysing NCCT findings in acute stroke which include intracerebral hemorrhage (ICH) and subtypes, cranial fractures, mass effect, midline shift, early ischemic changes (EIC), acute infarcts, and automated Alberta Stroke Program Early CT Score (ASPECTS) assessments.

🤖Patients First, AI Second🤖

Ethical and regulatory landscape of healthcare AI

AI in the Provision of Health care: Key points from American College of Physicians Policy Position Paper

• The ACP Medical Informatics Committee, The ACP Ethics, Professionalism and Human Rights Committee and The Board of Regents developed the position paper based on the basis of a review of laws, regulations, ethics and broad review of studies and literature about the uses of AI in Healthcare.

Recommendations:

To ensure the ethical and effective integration of AI in healthcare, the following are the recommendations:

• Clinician Engagement: Clinicians should always be involved in the AI development and implementation process to ensure alignment with clinical needs and practices.

• Prioritise Clinical Safety, Effectiveness, and Health Equity: A continuous improvement process must be established, which includes a robust feedback mechanism, peer review processes, and risk assessment. This ensures AI tools are safe, effective, and equitable.

• Minimise Health Disparities:

  1. Maintain diversity and inclusion by using data that represents all populations to ensure AI tools do not exacerbate health disparities.

  2. Support from Congress, the Department of Health and Human Services, academia, and the federal government is crucial for research and investment in inclusive AI technologies.

• Accountability: AI developers must be held accountable through a coordinated federal AI strategy. This strategy should be built on a unified governance framework that covers development, deployment, enforcement, and reporting of adverse events.

• Reduce Clinical Burden: AI tools should be designed to reduce the clinical workload and enhance patient care.

• Training for Healthcare Professionals: Healthcare professionals need proper training to understand and effectively use AI-enabled healthcare systems.

• Environmental Considerations: The environmental impact of AI technologies should be monitored and mitigated throughout their lifecycle.