.webp)
Harrison.ai launches its AI platform, Harrison.rad.1, for radiology to enhance diagnostic accuracy and patient care.
An Australian med-tech company, Harrison.ai, has announced the availability of its generative AI platform called Harrison.rad.1. It is a radiology-specific vision language model designed to detect and localize radiological findings from medical images. Amazingly built, Harrison.rad.1 is very much like OpenAI's ChatGPT, where users are invited to upload an X-ray image and instruct the model to state what is found, getting results almost immediately.
The AI model, built on millions of clinical images, reports, and radiology studies, utilizes a proprietary dataset that has been annotated by a large group of medical specialists to ensure clinically accurate training signals. As said by Aengus Tran, the co-founder and CEO of Harrison.ai, data quality, modeling intensity, and the ethical trajectory influence the success both in practice and clinical adoption for AI technology.
The previous radiology solution offered by the company, Annalise.ai, is cleared for clinical use in over 40 countries and is in clinical use in hospitals worldwide. The chest X-ray solution from Annalise.ai has been undergoing trials in Scotland since 2022 with some success, particularly in cutting the time to lung cancer treatment by nine days.
Harrison.rad.1 did remarkably well with the Fellowship of the Royal College of Radiologists (FRCR) 2B Rapids exam, scoring 51.4 out of 60. This score is on par with experienced radiologists but leaves behind other AI models like OpenAI's GPT-4 and Google's Gemini-1.5 Pro, whose score was below 30.
Harrison.ai intends to cooperatively work with industry partners, health professionals, and regulators to assure the delivery of AI per safe, fair, and responsible methods of deployment in healthcare. The objective is to widen the ambit of its AI technology across different medical image areas such as CT brain and chest scans, able to impact ourselves in hospitals in the future for as much as 80% of radiology scans.
With an investment of around $160 million, Harrison.ai will continue on its pathway to enable enhanced patient care, leveraging the best AI that currently exists.
.webp)
