I want to try the impossible to show it can be done – Artificial Intelligence (AI)

Lung cancer results in over 1.7 million graves per year, causing it the deadliest of all cancers worldwide more than breast, prostate, and colorectal cancers connected and it’s the sixth most frequent cause of death globally, according to the World Health Organization. (1) While lung cancer has one of the poorest survival rates among all cancers, medications are much more successful when the cancer is detected early. Unfortunately, the statistics are sobering because the powerful majority of cancers are not found until later stages.

Over the past three decades, doctors have searched ways to screen people at high danger for lung cancer. Though moderate dose CT screening has been confirmed to reduce mortality, there are still difficulties that lead to an unclear diagnosis, following unnecessary procedures, economic costs, and more. (2)

Google is decided to change that, and with its new AI-based tool, it expects to make lung cancer prognostication more accurate and more accessible. Google has been using AI to solve problems in healthcare from diagnosing eye infections to predicting patient results in medical records. (3) Today they are sharing new research explaining how Artificial intelligence (AI) can predict lung cancer in methods that could increase the chances of survival for many people at danger around the world.

Radiologists typically look into hundreds of 2D images within a single CT scan and cancer can be minuscule and difficult to spot. They created a prototype that can not only create the overall lung cancer malignancy prediction which is viewed in 3D volume but also recognize subtle malignant tissue in the lungs (lung nodules). (4)

Testing AI with certified radiologists:

The machine learning algorithm on more than 45,856 chest CT scans, taken from the National Health Institute and Northwestern University, some of which highlighted cancer in various stages. It gained an impressive 94.4% accuracy in this cycle of testing.

The algorithm was later put to work using a single CT scan for diagnosis, with the correctness of the algorithm compared to that of six board-certified radiologists. Google’s AI identified five percent more cancer cases than the radiologists. It also decreased false-positive exams by more than 11 percent. (5)

In addition to determining a patient’s overall lung malignancy, the model can identify subtle malignant tissue in the lungs. The deep learning algorithm can also factor in data from previous scans, which is helpful in predicting lung cancer risk because the growth rate of different lung nodules can be indicative of malignancy. The researchers think their findings show that deep learning and AI can significantly develop lung cancer screenings. (6)

Despite the value of lung cancer screenings, only 2-4 percent of qualified patients are screened presently. This work shows the potential for AI to improve both accuracy and consistency, which could further accelerate the adoption of lung cancer screening worldwide.

Google has also employed machine learning and imaging analytics to recognize breast cancer. Researchers discovered that the machine learning algorithms were apt to flag breast cancer cells that had grown to nearby lymph nodes, which is a risk factor when deciding how to properly treat patients. The machine learning models beat other automated methods and challenged human clinicians. (7)

These initial results are encouraging, but further, they will value the impact and use in clinical practice. They are co-operating with Google Cloud Healthcare and Life Sciences crew to serve this model through the Cloud Healthcare API and are in direct conversations around the world to advance additional clinical validation research and deployment. (8)

To know more about Artificial Intelligence, refer:

Artificial Intelligence Part 1

Artificial Intelligence Part 2

Artificial Intelligence Part 3