Feature
Products that incorporate artificial intelligence (AI) are making inroads into the health care industry, with major medical equipment manufacturers and software giants being joined by a range of startups. The Food and Drug Administration has granted approval to 691 medical devices based on AI and machine learning, as of last fall. Many of the devices draw on AI’s capabilities in image identification, enhancing the work of radiologists, while others are using large language models (LLMs) to provide assistance and suggest diagnostic insights, based on patient records and health care literature. Here, we examine some of the top products employing AI for health care applications.
1. IQ3
Butterfly Networks combines a handheld probe, a smartphone and AI software to create a portable ultrasound instrument that can be used even by technicians who have little experience with ultrasound. The IQ3 probe is built around semiconductor-based ultrasound-on-a-chip technology that captures high-resolution images and displays them on a smartphone. Its AI algorithms, hosted in the cloud on AWS, help identify issues of concern in the images. For instance, an algorithm trained on thousands of ultrasound lung images can automatically count B-lines — sonographic artifacts that can indicate breathing problems — based on a six-second clip. Counting them manually takes much longer.2. Tempus One
The Tempus One virtual assistant gives clinicians quick access to a patient’s complete clinical and molecular profile. It also has access to an array of other data sets that the clinician can query to help them make clinical decisions. The software is a voice and text virtual assistant that uses generative AI built on large language models to answer questions and provide clinically supported guidance. It is designed to help clinicians deal with genomic testing that can inform decisions for personalized medicine. The assistant draws on Tempus’s health care dataset, which contains more than 100 petabytes of curated data.
3. Aidoc
Aidoc’s software for cardiovascular disease uses AI to examine medical scans and consolidate data to flag suspicious findings for human radiologists. The system provides triage to highlight high-priority cases and streamlines a radiologist’s workflow. It helps to coordinate recommended procedures for patients and can also alert clinical trial coordinators if a patient might be a good candidate for their trial. The software also manages patient follow-up for clinics and primary care physicians.
4. CT-3500
The CT-5300 scanner by Philips uses AI-based image reconstruction to allow imaging with an 80% lower radiation dose, 85% lower noise and a 60% improvement in low-contrast detectability. AI also drives a smart positioning camera, reducing the time it takes to position a patient by up to 23%, while improving manual centering accuracy by up to 50% and increasing consistency from user to user by up to 70%. A suite of AI-enhanced workflow tools help improve dose, speed and image quality in scanning for various conditions, from cardiac imaging to trauma. 5. Navina
The Navina generative AI assistant by the company Navina is designed to streamline the handling of large amounts of patient data, turning patient charts and information on separate computer screens into natural language interactions. The AI assistant helps doctors to understand a patient’s health status, handle administrative tasks, such as generating progress notes and referral documents, and get recommendations for care based on a patient’s data and up-to-date clinical guidelines.
See more: 10 Top AI Health Care Companies
6. Sonic DL
Sonic DL by GE Healthcare is a deep learning technology that allows magnetic resonance images to be captured up to 12 times faster than with conventional techniques. That’s fast enough to capture a high-quality cardiac image in the space of a single heartbeat, the company says. It reduces overall scan time by as much as 83% and doesn’t require patients to repeatedly hold their breath. Not only is holding breath tiring for patients who can do it, it prevents accurate scans for those who can’t. And prolonged scanning can also reduce image quality due to the increased chance of movement. 7. AI-Rad Companion
The AI-Rad Companion by Siemens Healthineers uses AI algorithms to automate the post-processing of imaging data sets. It aims to automate routine tasks in radiology to improve a radiologist’s workflow and handle high volumes of cases. It contains different modules for different modalities and regions of the body. For instance, for lung CT, it highlights nodules in the lungs and calculates volume, diameter and tumor burden. For brain MR, it automatically segments different brain structures and provides individual volumetric analysis. 8. Aquillon One/Genesis SP
The Aquillon One/Genesis SP by Siemens is a CT scanning platform that relies on deep learning to enhance imaging. It can image various aspects of the heart in the space of a single heartbeat as well as image multiple anatomical areas during one held breath and with one contrast injection. It uses adaptive iterative dose reduction to reduce radiation exposure while achieving resolution of 0.5 mm. The platform’s neural network was trained with high-quality imaging data of the whole body, including the brain, lungs, heart and musculoskeletal system. 9. Microsoft Fabric
Microsoft Fabric is a general-purpose analytics platform, and Microsoft Cloud for Healthcare has created tailored offerings within it to handle health care analytics and AI workloads. It allows organizations to bring together previously separate data, including electronic health records, picture archiving and communications systems, lab systems, claims systems and medical devices. With that, they can use text analytics for health, a Microsoft Azure AI language service that allows them to get patient care insights from unstructured data in seven languages. The platform also allows them to use Azure’s AI Health Bot to customize an assistant for managing clinical and administrative workloads and AI Health Insights to assist doctors in decision making. 10. Curie
Curie by Enlitic is an AI-powered data management framework for medical images. It contains applications that allow for data standardization, including consistent and clinically relevant labeling. It allows images to be analyzed in real-time as well as historically and creates an imaging database to enhance research. Modules offer data anonymization and de-identification. The framework speeds up workflow and reduces repetitive tasks and can also reduce billing errors due to consistent labeling, the company says.
See more: 5 AI Case Studies in Health Care
