(26 Jan 2020) LEAD IN:
Researchers in Taiwan are combining machine learning through artificial intelligence (AI) with medical imaging to help diagnose tumors more quickly.
They say  AI computing reduces of the time required to diagnose a tumor, from days to under ten minutes. This allows patients to be treated earlier, improving their chances of recovery.  

STORY-LINE
These AI Labs are working with Taipei Medical University Hospital to develop tools which cuts the time it takes doctors to diagnose cancers.
An advance which could benefit patients.
Ethan Tu is the founder of Taiwan AI Labs.
He explains how the new AI platform learns to read the medical images: "We actually learn how to segment a liver out of the CT scan slides from the open data or only a few data set. Then we use an interactive way to automatically learn from the doctor how to segment the tumor out of the CT scan. And through this interactive approach, we don't need a huge amount of medical images. We only need a small amount of medical images to start with. Then using the interactive way, the computer will learn from the doctor over  time."
Medical imaging is a technique of creating images of organs and tissues inside a body allowing doctors to investigate whether medical intervention is necessary.
Machine learning is when a set of techniques and algorithms enable computers to discover complicated patterns in large data sets.
They team says the old method of using CT scans to detect tumours meant a delay of five to seven working days for the results. Now doctors can have the results in less than ten minutes.
The collaboration with Taipei Medical University Hospital began in June 2019.
Taiwan AI Labs is also working on similar technology with Taipei's Veteran Hospital.
The company develops customized systems and platforms to match the capabilities of the different scanning machines used by the hospital.
Chung Yi-hsiang of Taipei Medical University Hospital says: "I am indeed very positive about AI's future. Artificial intelligence currently helps in interpreting the medical images, but in the future, I hope it will be combined with gene mapping, automatic calculation and machine learning so that it could learn by itself after it is fed with enough data. Then it will be even more accurate when interpreting medical images."
He believes AI is a welcome development which should improve patients' outcomes.
"Artificial intelligence can help a doctor shorten the time to figure out if there is any early stage tumor. With the saved time, the doctor can discuss and explain the disease to the patient. Take our unit as an example, we can detect the disease within five to ten minutes and the doctor will have enough time to speak with the patient about it," he says.
Training a computer to correctly identify a tumour means it must process thousands of images to enable the machine learning algorithms to build up an understanding of what it needs to identify. .
Lin Ying-hung is the director in charge of Taiwan AI Labs' development of the Taiwan Precision Medical Imaging Platform which is what the computer platform is called.
He believes the precision offered by the technology prevents wasted resources.
Artificial intelligence applied in medicine also changes traditional doctor-patient relationships.
Chen says artificial intelligence still has room for improvement when it comes to making a patient comfortable with accepting a diagnosis based on AI.

Find out more about AP Archive: http://www.aparchive.com/HowWeWork
Twitter: https://twitter.com/AP_Archive
Facebook: https://www.facebook.com/APArchives
Google+: https://plus.google.com/b/102011028589719587178/+APArchive​
Tumblr: https://aparchives.tumblr.com/​​
Instagram: https://www.instagram.com/APNews/

You can license this story through AP Archive: http://www.aparchive.com/metadata/youtube/a102871afbc5412c833ff0a12918eff9