Last Month In Healthcare AI (August 2024)

Contents: ?? 1) Prediction & Diagnosis | ??2) Treatment & Care | ?? 3) Latest Approvals | ?? 4) Generative AI | ??5) AI Assistants | ??6) Governance & Ethics

?? 1) Prediction & Diagnosis

?? GPs use AI to boost cancer detection rates in England by 8%.? The C the Signs artificial intelligence program scans medical records, test results and other personal characteristics to increase the likelihood of spotting cancers and has helped doctors detect significantly more cases (study )

?? Healthcare AI platform Ubie and the American Kidney Fund (AKF) will collaborate to fine-tune Ubie's AI-enabled symptom checker for kidney disease detection aiming to shorten a patient's time to treatment. Ubie's AI is trained on clinical research papers, and then its physician review panel analyzes the prediction pathways and symptoms associated with a disease

?? An AI tool from the University of Cambridge is over 80% accurate at predicting if early dementia signs will progress to Alzheimer’s, using non-invasive cognitive tests and MRI scans. It outperforms standard methods, distinguishing stable cases from those likely to develop Alzheimer’s, and predicts symptom progression speed

?? Tulane University developed a deep learning model that outperforms existing methods in predicting osteoporosis risk, using key factors like age, grip strength and weight. Tested on over 8,000 participants, it could enable earlier diagnosis and more accessible risk assessments

?? Fujifilm and Juntendo Hospital in Japan developed an AI that predicts fall risk in outpatients with 96% accuracy using data from over 70,000 patients. The AI analyzes over 500 features, including age and prescription history, from Fujifilm's CITA Clinical Finder.

?? AI and imaging advancements are improving multiple sclerosis diagnosis. A new model combining optical coherence tomography and infrared imaging achieves 92% accuracy, outperforming traditional methods. This approach detects subtle optic nerve changes that may be missed by humans, offering a less invasive, cost-effective alternative to MRI.

?? AI can now detect early signs of metabolic diseases by analyzing facial temperature patterns. Researchers found that colder noses and warmer cheeks may indicate conditions like high blood pressure and diabetes. These temperature differences, identified using thermal cameras and AI, correlate with aging and metabolic health

??2) Treatment & Care

?? The new AI tool, DeepPT, predicts a cancer patient's response to treatment using histopathology images of tumor tissue, which are quick and cost-effective to process. Developed by the National Cancer Institute and Pangea Biomed, it helps select the most suitable treatment. If the tool predicts a positive response to therapy, patients are 2-5 times more likely to respond

?? A study by Digbi Health found that personalized dietary and lifestyle programs, enhanced by AI, can influence the gut microbiome and support weight loss. Analyzing data from 103 participants, the AI identified specific gut bacteria and pathways linked to BMI reduction. The program, which uses genetic and microbiome profiling, showed that 80% of participants lost weight

?? PocketHealth uses AI to help patients understand medical scans.? It allows users to access and share their medical images and reports quickly. The AI feature generates prompt questions for patients to ask their doctors, helping them better prepare for appointments

??3) Latest Approvals

?? Philips' IntelliSite Pathology Solution 5.1, now FDA cleared , uses AI-driven diagnostics and cloud-based collaboration to streamline pathology workflows and enhance diagnostic accuracy.

??Qure.ai has received Health Canada's Class III medical device license for its AI-powered medical imaging solutions. The approval covers applications for lung cancer detection, neurocritical CT triage, and more, enhancing diagnostic accuracy and patient care. This adds to Qure.ai's existing FDA and EU approvals,

??4) Generative AI

?? GE HealthCare partners with Amazon Web Services (AWS) to build generative AI for medical uses.? The pair will build models that clinicians can use to leverage data more efficiently across health-care operations, including screenings, diagnoses, decision support and workflows like scheduling

?? Microsoft is partnering with Mass General Brigham to develop radiology-focused multimodal AI foundation models. The aim is to develop, test and validate AI algorithms and applications that improve the accuracy and consistency of medical image analysis and enable healthcare organizations?to build medical imaging AI copilots

?? A new AI tool tested at NYU Langone Health shows promise in responding to patient questions in electronic health records (EHRs). The AI-generated responses were found to be as accurate, relevant, and complete as those by human providers, with higher scores for empathy and tone

?? Scientists at Weill Cornell Medicine and Dana-Farber Cancer Institute have customized ChatGPT for digital pathology, creating an AI tool that provides accurate, literature-backed answers to complex questions. Another tool integrates ChatGPT with PathML software, offering step-by-step coding guidance for pathologists, making advanced image analysis more accessible.

?? Abridge, in collaboration with Mayo Clinic and Epic, is developing AI-powered tools to assist nurses with medical documentation. This new generative AI product integrates into Epic's inpatient nursing workflows, helping nurses manage complex tasks like patient care notes, vital signs, and safety documentation

??5) AI Assistants

?? OpenAI and Thrive Global launched Thrive AI Health to create a genAI-enabled health coach for personalized behavioral change. The AI coach focuses on connection, sleep, fitness, stress management, and food, aiming to improve health outcomes and reduce healthcare costs

?? Research from Nottingham Trent University highlights the potential of chatbots in mental health care. A review of 31 studies found chatbots effective in managing conditions like depression and anxiety, with users responding well to their empathetic interactions. However, usability issues, especially for older adults, need improvement

?? A new AI chatbot trained on drag queens by the AIDS Healthcare Foundation helps patients with sexual health concerns. The chatbot, called Glitter Byte, offers non-judgmental advice, manages appointments, and provides STI information, particularly aiding underserved communities.

??6) Governance & Ethics

?? The EU's new AI Act , set to be effective by 2026, imposes strict regulations on AI, especially in clinical research. High-risk AI systems must meet standards for transparency, data governance, and human oversight. This law affects global companies using AI in the EU, requiring immediate preparation to ensure compliance and maintain safety and ethics in AI deployment