To cut it short, it works just the way human doctors detect cancer; recognizing patterns and interpreting X-ray images, microscope slides, M.R.I.s and other medical scans.

We can train computers by feeding a large number of medical images and data into the Artifical Neural Network Systems which can recognize patterns to a specific disease such as wrist fracture, cancer or pneumonia at an early stage which might be hard for a person to detect. Such a system works just like human beings, the more we study or practice, the more we learn; likewise, the more data it receives, the better it detects.

In another study, the researchers applied AI to read CT scans which are used to screen people for lung cancer which killed 1.7 million people last year. Studies have found that screening can reduce the risk of dying from lung cancer. Additionally, the scan can help to identify spots which might later become cancer, so that radiologists can sort patients into risk groups and decide whether they need frequent follow-up scans or biopsies.

According to Dr Tse, one of the researchers involved said that using large data sets for training is just like school students, who are given quizzes and lessons so they can learn and ace the test. Likewise, after training the computers for a long period of time, the result that they received was commendable.  

Tested against 6716 cases with known diagnoses, the system was 94 per cent accurate. Pitted against six expert radiologists, when no prior scan was available, the AI beat the doctors: It had fewer false positives and false negatives. When an earlier scan was available, the system and the doctors were neck and neck. This is a result of superior computational power that makes the machines process a vast amount of data for AI to identify patterns that humans cannot see.