AI’s 82% Accuracy in Cancer Detection Offers Hope to Patients


A new study suggests that artificial intelligence is much better at assessing the aggressiveness of a rare type of cancer from scans compared to the current method. AI, by detecting subtle details that are not visible to the naked eye, achieved an 82% accuracy rate, while traditional lab analysis scored only 44%.

Researchers from the Royal Marsden Hospital and the Institute of Cancer Research believe that this AI technology could significantly enhance cancer treatment and potentially help thousands of patients each year. They are also excited about the AI’s potential to detect other types of cancer at an early stage.

AI is already demonstrating great promise in diagnosing breast cancer and reducing the time it takes to start treatment. Computers can process vast amounts of data and be trained to recognize patterns, make predictions, solve problems, and even learn from their own mistakes.

“We’re incredibly excited by the potential of this state-of-the-art technology,” stated Professor Christina Messiou, a consultant radiologist at The Royal Marsden NHS Foundation Trust and a professor specializing in personalized oncology imaging at The Institute of Cancer Research in London.

“It could lead to patients having better outcomes, through faster diagnosis and more effectively personalised treatment.”

In their study published in Lancet Oncology, the researchers employed a method known as radiomics to detect subtle indicators, imperceptible to the naked eye, of retroperitoneal sarcoma. This type of cancer develops in the connective tissue located at the back of the abdomen, and the researchers analyzed scans from 170 patients.

Using this dataset, the AI algorithm was able to assess the aggressiveness of tumors in an additional 89 patients from hospitals in Europe and the United States more accurately than traditional biopsies, which involve analyzing a small portion of cancerous tissue under a microscope.

Dental nurse Tina McLaughlan received a diagnosis of a sarcoma located at the back of her abdomen in June last year, following complaints of stomach pain. To pinpoint the issue, doctors used computerized-tomography (CT) scan images as they deemed a needle biopsy too risky.

Although Tina, a 65-year-old from Bedfordshire, was not a participant in the AI trial, she shared with BBC News that she believes this technology will be beneficial for other patients. After having the tumor removed, she now undergoes scans every three months at the Royal Marsden.

“Hopefully, it would lead to a quicker diagnosis.”

About the personalised treatment

The researchers’ vision for this technology is to provide personalized treatment for individuals diagnosed with this type of cancer, which affects approximately 4,300 people in England annually.

Professor Christina Messiou envisions that this technology could eventually be implemented worldwide, enabling high-risk patients to receive customised treatments while sparing low-risk individuals from unnecessary treatments and follow-up scans.

Dr. Paul Huang from the Institute of Cancer Research in London expressed his enthusiasm, stating that this technology has the potential to bring about a transformation in the lives of individuals with sarcoma. It could allow for the development of treatment plans customized to the specific biology of their cancer.

“It’s great to see such promising findings.”

It’s much better at finding out how aggressive a rare cancer is from scans than the methods we use now. Researchers from The Royal Marsden Hospital and The Institute of Cancer Research in London used AI to uncover hidden clues in scans.

This means we can catch cancer earlier and make personalized treatment plans. AI can help us make more accurate diagnoses and prevent people from getting treatments they don’t need. This is a big step toward a future where people get the right treatment, improving healthcare and saving lives all over the world.

Read more

Recommended For You