Artificial intelligence: New prospects in treating colon cancer

Thanks to the new technologies and diagnostic instruments, identifying colon cancer in real time has become possible

Written Thursday, by Emanuele Mortarotti

The new discoveries in the field of medical research are going hand in hand with the exponential development of artificial intelligence (AI) and the development of new medical devices linked to it. Diagnostic tests for cancer now use AI in order to be more specialised in identifying tumour growth. The study we are discussing today, which appeared only days ago on the Futurism website, shows how these tests are effectively useful and revolutionary in their category.

A necessary change

A new computerised endoscopic system that checks for the presence of cancer cells or their growth could be the future of cancer detection. Closely linked to the use of Artificial Intelligence, this new diagnostic system is able to determine if a group of cells called colorectal polyps, which grow attached to the walls of the colon, are benign, that is, definable as colorectal adenomas.

This new intelligent equipment was tested using more than 30,000 sample images of colorectal polyps, each magnified 500 times, and functions using the machine’s memory. Artificial Intelligence is able to verify approximately 300 different characteristics (which it compares with the knowledge already existing on the subject) of each polyp in less than a second. The initial success found in these preliminary studies was followed by a series of prospective studies: the results, the absolute first regarding endoscopy together with AI in clinical spheres, were presented at the 25th UEG (United European Gastroenterology) WEEK in Barcelona between the end of October and beginning of November 2017.

The innovative study was conducted by a team led by Dr. Yuichi Mori from the Showa University of Yokohama, Japan. Mori and his colleagues tested their new system on 250 patients who had previously been diagnosed with colorectal polyps. Artificial intelligence “predicted” the pathology of each polyp, comparing it subsequently with other pathological reports taken from previously removed samples. The results were more than encouraging: the system examined 306 polyps in real time, with 94% susceptibility, 79% specificity and 86% accuracy, and also registered excellent results in identifying the growth of irregular tissue that is potentially dangerous for the colon.

Artificial Intelligence applied to medicine

We have therefore seen that AI is able to identify, quickly and with absolute reliability, the abnormal growth of cells that are almost certainly carcinogenic. “The most decisive and noteworthy discovery of this system is that artificial intelligence allows an optical biopsy to be done in real time on colorectal polyps during a coloscopy, regardless of the skills of the endoscopy technician”, affirmed Mori during his opening speech at UEG WEEK. “This allows the adenomatous polyp to be completely removed and prevents a useless polypectomy on non-neoplastic polyps”. Moreover, researchers presented the results from their study to prove that their system was ready for clinical experimentation.  “We believe these results are more than acceptable to begin considering it a medical application; our primary objective is to obtain regulated approval for this diagnostic system of ours”, added Mori. While this type of biopsy can be saluted by the medical community as the first AI-friendly one, it is not the first time that artificial intelligence has been used to improve medical diagnoses and, in general, to contribute to research in this field. For example, there is an instrument that is equipped with artificial intelligence able to identify skin cancer, and the NVIDIA company is working on a device that is able to accelerate cancer research using technologies that are connected to what is known as deep learning, the brand new branch of AI that relies on study, development and testing of neural networks. Watson, the intelligent system created by IBM, has also obtained surprising results, reaching in some cases 99% accuracy in recommending the same therapy recommended by an actual physician.

In conclusion, it is obvious that researching and improving medical devices and instruments that identify cancer will become fundamental for saving the lives of millions of people. We anticipate an interesting future and will be following the evolutions and possible applications in the field of Medical Devices with great attention. If you would like further information, contact Dispotech, Your Disposable Excellence.

Emanuele Mortarotti
Author Emanuele Mortarotti


write a comment