Researchers at the University of Missouri tested artificial intelligence to detect melanoma by evaluating images of suspicious skin abnormalities. The work aims to help doctors find cases faster so patients can get medical attention sooner.
The technology is a decision-support tool and not a replacement for medical experts. It could help patients with limited access to dermatologists. The team trained models using a very large set of images from three-dimensional total body photography, a method that maps the skin.
The researchers compared three existing AI models. Individual models reached high accuracy and combining them improved results. The study shows promise but the system is not yet ready for routine clinical use.
Difficult words
- researcher — People who do scientific studies.Researchers
- detect — Find or notice something that is not easy to see.
- melanoma — A serious type of skin cancer.
- abnormality — Something on the body that looks different from normal.abnormalities
- dermatologist — A doctor who treats skin problems.dermatologists
- accuracy — How often a test gives correct results.
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- How could this technology help patients who live far from dermatologists?
- Would you trust a decision-support tool to help your doctor? Why or why not?
- What data did the researchers use to train the AI models?
Related articles
Milk glands in many mammals have receptors for H5N1
A study found that milk gland tissue from pigs, sheep, goats, beef cattle, alpacas and humans contains sialic acid receptors that can let H5N1 attach to cells. Researchers warn of surveillance and concerns about raw milk from mammals.