Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Micrograph of lung tissue affected by cancer

Lung tissue (brown) is invaded by cells (purple) of a type of cancer called adenocarcinoma, which can be detected by a machine-learning model. Credit: Eye of Science/SPL

Cancer

Pathologists meet their match in tumour-spotting algorithm

Deep-learning model picks out two kinds of lung cancer by studying images of tissue.

Artificial-intelligence technology that examines images of lung tissue can identify two common lung cancers just as well as pathologists do.

A team led by Narges Razavian and Aristotelis Tsirigos at the New York University School of Medicine trained and tested a convolutional neural network — a deep-learning algorithm that is adept at processing images — on 1,634 images of cancerous and healthy lung tissue. The algorithm identified healthy cases and distinguished as accurately as three pathologists between two common types of lung cancer: adenocarcinoma and squamous-cell carcinoma.

The researchers also trained the network on adenocarcinoma images that had been labelled with the mutations underlying the cancer. After this training, the algorithm was able to accurately predict the mutations associated with some unlabelled images.

If the algorithm were trained on further labelled images, it might be able to identify adenocarcinomas’ mutations with greater accuracy, the researchers say. That could improve this tumour’s treatment, which is often tailored to the underlying mutation.

More Research Highlights...

Coloured transmission electron micrograph of two Streptococcus sanguinis bacteria

Genomic analysis identified starch-loving Streptococcus sanguinis bacteria (artificially coloured) in the mouths of modern humans and Neanderthals, but not in chimpanzees’ mouths. Credit: National Infection Service/Science Photo Library

Microbiome

Microbes in Neanderthals’ mouths reveal their carb-laden diet

Gunk on ancient teeth yields bacterial DNA, allowing scientists to trace the oral microbiome’s evolution.
Artist's concept of NASA's Voyager 1 spacecraft entering interstellar space

Data collected by the Voyager 1 spacecraft, which launched in 1977, has helped scientists to calculate the density of the interstellar plasma. Credit: NASA/JPL-Caltech

Astronomy and astrophysics

Voyager 1 captures faint ripples in the stuff between the stars

The first spacecraft to visit interstellar space has now become the first to make continuous measurements of waves in that remote realm.
Light micrograph of a human egg cell during fertilisation

As a human egg cell is fertilized, two chromosome-containing cellular structures (dotted circles, centre) merge into one — a process that often goes wrong. Credit: Pascal Goetgheluck/Science Photo Library

Developmental biology

The error-prone step at the heart of making an embryo

High-resolution imaging shows why the union between two sets of chromosomes goes awry as least as often as not.
Satellite image of broken iceberg B-44.

Dark water borders chunks of iceberg broken off a West Antarctica glacier. The melting of the region’s ice sheet could allow the bedrock to rise, sloughing water into the ocean. Credit: NASA

Climate change

Antarctic rocks on the rebound could raise sea level much more than expected

When the ice covering the west of the continent disappears, the bedrock could rise up and shove extra water into the ocean.
Monteverde Cloud Forest Preserve, Costa Rica

Mist wafts through the trees at the Monteverde Cloud Forest Biological Preserve in Costa Rica. Cloud forests around the world are threatened by development, wood collection and climate change. Credit: Stefano Paterna/Alamy

Conservation biology

Forests that float in the clouds are drifting away

Tropical cloud forests are safe havens for a vast range of creatures and plants, but they are under siege around the globe.
Illustration of a brown dwarf

A rapidly spinning brown dwarf (pictured, artist’s impression) tends to have narrow atmospheric bands; the faster the spin, the thinner the bands. Credit: NASA/JPL-Caltech

Astronomy and astrophysics

Dim stars that have failed at fusion are masters of spin

Three brown dwarfs whirl on their axes at a dizzying rate that might be close to the celestial speed limit for these bodies.
Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing

Search

Quick links