Artificial Intelligence May Detect Overlooked Lung Tumors

Artificial intelligence holds great promise, particularly for rapid and accurate cancer diagnosis. Turkish Scientist Ulaş Bağcı and his team's artificial intelligence system successfully diagnosed small lung tumors with % 95 accuracy. The radiologists used for comparison achieved only % 65.

The research was conducted at the University of Florida by a team led by Turkish scientist Ulaş Bağcı. The workings of the brain were used as a model to create the system. This artificial intelligence model inspired by brain function is called multilayer artificial neural networks(convolutional neural network); this is also known as deep learning as a method.

In the figure above,you can see the definitions and differences of 3 widely used, closely related new computing terms: artificial intelligence, machine learning and deep learning.

Using multilayer artificial neural networks, the system was taught what to look for in computed tomography (CT) images. More than 1000 CT images from the American National Research Institute and Mayo Clinic were entered into the software, enabling the computer to find tumors; it successfully diagnosed tumors with % 95 accuracy.

You May Be Interested:

• What Imaging (Radiology) Tests Are Used in Cancer?

What Does This Study Mean for Us?

• The research was presented in Spain at the 21st Medical Image Computing and Computer Assisted Intervention (MICAI 2018) congress.
• This system closely resembles the facial recognition systems familiar from social media (for example, Facebook recognizing and tagging people in photos by their faces). This artificial intelligence software, using a multilayer artificial neural network model, examines different lung CT images to recognize tumors.
• In the future, artificial intelligence systems are considered certain to be used in cancer diagnosis with high success. However, we are still at a very early stage when it comes to artificial intelligence making treatment decisions.
• Doctors Win the First Round Against Machines
• IBM Watson's Treatment Recommendations Are Not Accurate or Safe
• Our Other News on Artificial Intelligence in Medicine