Top international medical AI company Lunit has unveiled a new medical imaging AI platform that enables nodule detection and consolidation in seconds, with near-perfect accuracy.
Chest x-rays and mammography are among the most used diagnostic imaging tests conducted and are vital to the detection of lung and breast cancer. However, the accuracy of traditional methods often falls short of modern expectations. Over 25 percent of lung cancers and 15 percent of breast cancer cases are missed by chest x-ray.
Even the most experienced radiographers stand to benefit from the machine learning capabilities of AI image analysis software. These AIs are given access to a wealth of training data, spotting trends and minute abnormalities that the human eye is simply unable to detect. This sort of complex visual pattern recognition is the perfect candidate for AI assistance, thanks to advancements in deep learning.
Founded in 2013, in South Korea, Lunit develops advanced medical image analytics and novel imaging biomarkers. Their INSIGHT platform claims to be the first ever real-time imaging AI analytics on the web.
How Lunit’s medical imaging AI platform works
Lunit INSIGHT has been trained with a huge collection of anonymous clinical data from 18 partner hospitals, totalling over 1 million case images.
Using this data, the AI algorithms have learned to detect target diseases and significant radiologic findings, including lung cancer, tuberculosis, pneumonia, pneumothorax, and breast cancer for chest x-ray and mammograms.
“Lunit’s vision is to develop advanced software for medical data analysis and interpretation that goes beyond the level of human vision,” said Anthony Paek, CEO of Lunit. “In presenting Lunit Insight, we hope to contribute in opening a new era of medical practice, by helping and empowering healthcare professionals to make more accurate, consistent, and efficient clinical decisions for the patients.”
The benefits of AI in medical imaging
When even the latest image recognition AIs can mistake tortoises for guns, the thought of relying on AI to diagnose life-threatening diseases can sound unnerving. Yet, Lunit’s new cloud based AI medical imaging software boasts 97 percent standalone accuracy in nodule detection and 99 percent for consolidation and pneumothorax (the presence of air or gas in the cavity between the lungs and the chest wall).
The Lunit Insight platform is publicly accessible. Anyone can log in and upload x-ray images, which the AI will assess in seconds and return its findings, including the level of abnormality and a visualisation of the AI’s attention map.
Additionally, Lunit solutions can be easily integrated in SaaS API and library incensing format to merge the technology with existing workflows. Nuance, EnvoyAI and Infinitt Healthcare are among the companies set to integrate the software.
Future uses of AI in medical imaging
The chest radiography tool is already available, but the mammography component is timetabled for release early next year. It will detect suspicious breast cancer lesions. Lunit is also looking to develop other medical imaging AI solutions for digital breast tomosynthesis, chest CT, and coronary CT angiography.
Lunit’s AI solutions have been proven to increase the diagnostic performance of its users by up to 20 percent, with non-radiology physicians benefitting from the software significantly. The technology isn’t designed to replace the diagnosis of medical professionals, merely to augment their conclusions – particularly in less definitive cases.
The medical imaging AI company hopes to obtain clinical validation in the coming months. “Large-scale multi-center reader studies are set to be conducted in early 2018,” said Suh, Chief Medical Officer of Lunit. “These are the studies with multiple leading hospitals in Korea and the US for Lunit’s chest x-ray and mammography solutions; publication of the results is targeted for late 2018.”
Likewise, Food and Drug Administration (FDA) approval for Lunit’s chest x-ray and mammography solutions are expected around the same period.
The platform has been ranked in the ‘Top 100 AI Start-ups’ by CB Insights and listed in the ‘Top 5 AI start-ups for Social Impact’ by Nvidia, so it’s gaining attention in the healthcare world and the wider technology space, and justifiably so. The solution’s capability and ease of use has been demonstrated and, with over 1 billion chest x-rays performed worldwide each year, its potential to impact patients’ lives is immense.