Taiwan-based start-up, JelloX Biotech, says its products will help doctors diagnose and treat cancer tumors with speed and accuracy without an added strain on budgets. The company made its debut at Intel Innovation in September.
JelloX has gained the attention of healthcare experts and pharmaceutical brands for its potential for the future of cancer pathology using 3D imaging technology.
The company believes that digital pathology, particularly using 3D imaging, is currently viewed by the healthcare industry as prohibitively expensive due to the need for costly graphics equipment, but JelloX says it plans to fill in this gap in the market with their artificial intelligence (AI)-powered software platform MetaLite combined with an edge solution.
The company is currently finalizing preparations to begin the FDA approval process for MetaLite next year through the software as a medical device path.
Because of a study released in The Harvard Gazette that shows that cancer risks are rising with each new generation, and — while cancer mortality rates have been declining — the COVID-19 pandemic has caused delayed cancer screening, diagnosis, and treatment, exacerbating the problem and reversing this trend, JelloX says its work is timely.
MetaLite, it says, is a cost-effective software platform that leverages AI to analyze 3D-rendered tissue samples rapidly across various parameters and then pinpoint and label tumors and biomarkers.
JelloX Biotech CEO, Yen-Yin Lin, said: “JelloX Biotech envisions a world where AI-driven 3D pathology empowers healthcare providers by making cancer diagnosis more precise, faster, and less costly.
“MetaLite plays a leading role in this future. Cancer diagnosis and treatment will be built around the platform’s federated learning capabilities to grow a singular knowledge pool accessible to doctors around the world — a cancer pathology database that fully protects patient privacy.”
MetaLite, Lin added, elevates efficiency and effectiveness by addressing the extensive labor costs, painstaking processes, and myriad technical expenses associated with cancer pathology.
Speed and accuracy
JelloX notes that the current common practice of manual diagnosis of physical biopsy samples is a complex process that takes many hours of exacting work by pathologists. MetaLite expedites this by rapidly analyzing the many parameters of a sample, identifying cancerous tissue and biomarkers, and then labelling them.
In fact, JelloX has developed a total solution prototype with MetaLite software run on an Intel device that can analyze every parameter in just a few minutes, compared with an hour on a standard computer and several hours manually. This solution also utilizes custom AI algorithms deployed through the Intel Distribution of OpenVINO Toolkit that further accelerates the process.
Besides being a lengthy process, studies have shown that when different doctors manually analyze the same sample, there is a 20% to 30% variation among the diagnoses, which has clear implications for treatment and the potential for misdiagnosis to occur.
MetaLite, the company says, is not only capable of pinpointing tumors and biomarkers with high precision, but its AI enables it to continuously learn and optimize its capabilities — and these will grow exponentially with federated learning added to a future version.
Furthermore, before 3D analysis can even begin, a physical tissue sample must be rendered into 3D format via a scanner. MetaLite, JelloX claims, offers broad compatibility across scanner brands and formats.
Based in Hsinchu, Taiwan, JelloX Biotech Inc. is a startup founded by researchers from National Tsing Hua University that focuses on advancing cancer pathology through digital and 3D imaging and AI technology.