advertisement

Lung cancer

30 April 2019

AI takes aim at lung cancer screening

Artificial intelligence is being used to make lung cancer screening more accurate, considerably reducing false positive findings.

The term artificial intelligence (AI) might bring to mind robots or self-driving cars. But one group of researchers is using a type of AI to improve lung cancer screening.

Screening is important for early diagnosis and improved survival odds, but the current lung cancer screening method has a 96% false positive rate.

Procedures mostly unnecessary

But in the new study, investigators were able to reduce false findings of lung cancer without missing any actual cases.

A low-dose CT scan is the standard diagnostic test for people at high risk of lung cancer. In the United States, about one-quarter of these scans reveal shadows indicating nodules in the lung. Despite that positive result, fewer than 4% of those patients actually have cancer.

"A positive test creates anxiety, increases health care costs, and the follow-up tests are not risk-free," said study co-author Panayiotis Benos. He's vice chairman of computational and systems biology at the University of Pittsburgh.

"For the 96% of people who have benign nodules, these procedures are unnecessary. So, we try to mine the data to tell which are benign and which are malignant," Benos explained in a university news release.

Benos and his colleagues entered CT scan data from 218 high-risk patients into a machine learning algorithm – a form of artificial intelligence – to create a model that calculates the probability of cancer.

Leading cause of cancer death

The researchers then compared the model's results against the patients' actual diagnoses. The model would have spared 30% of patients with benign nodules from further, unnecessary tests, without missing a single case of cancer, the team reported.

According to the study authors, they were able to rule out cancer in about one-third of the patients. This means these people wouldn't need biopsies, PET scans or short-interval CT scans. They could just make an appointment to come back in a year.

This is the first time artificial intelligence has been used in this way, the researchers said.

The study was published in the journal Thorax.

According to the American Cancer Society, lung cancer is the leading cause of cancer death among both men and women. This year, in the United States, the society estimates that there will be about 228 000 new cases of lung cancer and about 143 000 lung cancer deaths.

Image credit: iStock

 

Ask the Expert

Cancer expert

CANSA’s purpose is to lead the fight against cancer in South Africa. Its mission is to be the preferred non-profit organisation that enables research, educates the public and provides support to all people affected by cancer. Questions are answered by CANSA’s Head of Health Professor Michael Herbst. For more information, visit cansa.org.za.

Still have a question?

Get free advice from our panel of experts

The information provided does not constitute a diagnosis of your condition. You should consult a medical practitioner or other appropriate health care professional for a physical exmanication, diagnosis and formal advice. Health24 and the expert accept no responsibility or liability for any damage or personal harm you may suffer resulting from making use of this content.

* You must accept our condition

Forum Rules