trending Market Intelligence /marketintelligence/en/news-insights/trending/uz0tt1drbdl6hltr-wzemq2 content esgSubNav
In This List

AI beats dermatologists in diagnosing skin cancer

Blog

A Pharmaceutical Company Capitalizes on M&A Activity with Brokerage Research

Blog

2021 Year in Review: Highlighting Key Investment Banking Trends

Blog

Insight Weekly: US stock performance; banks' M&A risk; COVID-19 vaccine makers' earnings

Blog

Global M&A By the Numbers: Q3 2021


AI beats dermatologists in diagnosing skin cancer

Researchers in Germany, France, and the U.S. found artificial intelligence to be more capable than humans at detecting skin cancer.

According to a study published in the Annals of Oncology cancer journal, a so-called deep learning convolutional neural network, or CNN, was found to be better than experienced dermatologists in detecting malignant melanomas.

CNN is an artificial neural network that mimics the human brain and is capable of learning quickly from images and teaching itself from what it has learned to improve performance, a process known as machine learning.

More than 100,000 images of malignant and benign skin cancers and moles were shown to the CNN as part of its machine learning process.

"With each training image, the CNN improved its ability to differentiate between benign and malignant lesions," said Holger Haenssle, the first author of the study and the senior managing physician of dermatology at the University of Heidelberg in Germany.

Head to head

The AI network was matched up against 58 dermatologists from around the world.

During the first part of the comparison, the dermatologists and the CNN were given 100 images of the skin and asked to diagnose accordingly.

The dermatologists identified 86.6% of melanomas and correctly identified an average of 71.3% of lesions that were not cancerous. However, the CNN was able to identify 95% of melanomas correctly.

During the second part of the study, the dermatologists were provided clinical information about the patient and close-up images of the same 100 cases and asked for diagnoses and management decisions again. The dermatologists improved their performance by accurately diagnosing 88.9% of malignant melanomas and correctly identifying 75.7% that were not cancerous but did not beat the CNN.

The researchers believe that the CNN is not a replacement for dermatologists, but it can be used as an additional aiding tool. The team concluded that there is no substitute for a thorough clinical examination, but in the future, an automated diagnostics program using machine learning AI could "change the diagnostic paradigm in dermatology."

The incidence of cancerous melanoma is increasing, with an estimated 232,000 new cases worldwide and about 55,500 deaths from the disease each year. It can be cured if detected early, but many cases are only diagnosed when the cancer is more advanced and harder to treat.