AI and Lung Cancer
From smart watches to clinical trials, throughout the world medical researchers are making great strides in using artificial intelligence (AI) to diagnosis illnesses and custom-fit individual medical treatments. One example is how one group of researchers is using a certain type of AI to improve the way health professionals conduct lung cancer screening.
It is a well-accepted medical truth that regular screening is critical to early diagnosis and improving survival odds for patients, however the customary lung cancer screening method used today has a 96 percent false positive rate. The standard method uses a low-dose CT scan on people with a high risk of lung cancer. These scans show shadows which may indicate nodules in the lung – a typical sign of lung cancer. Contrary to these positive results, fewer than 4 percent of those patients actually have cancer.
But in a new study out of the University of Pittsburg, researchers were successful in reducing the rate of “false findings” of lung cancer without leaving out any true positive cases. In the study, CT scan data from over 200 high-risk patients were computed via a machine learning algorithm -- a form of artificial intelligence -- to create a model that calculates the probability of cancer.
The researchers compared the model's results against the patients' actual diagnoses. The algorithm model determined that about a third of the patients with benign nodules should not have undergone further, unnecessary tests. These labor-intensive follow-up tests include invasive biopsies, PET scans or short-interval CT scans, which all cost time, money, resources, and for the patient, unneeded stress. Remarkably, the model did not miss a single case of cancer according to the study authors.
Although the study was restricted to lung cancer screening, it clearly demonstrates another way that AI is positively changing the way scientists and medical doctors work together to improve the diagnosis and treatment possibilities for all patients. Studies like these save hospitals and patients time, money – and lives.