The idea of using dogs’ noses to detect the scent of human cancer has been around for decades, and researchers have even tried building machines to do the job, but none have been able to match the sensitivity of man’s best friend. Rather than try to compete with evolution, scientists decided to bring in the experts. In a proof-of-concept study, researchers used dogs to train a machine learning algorithm to detect prostate cancer, the second most common cancer in men. Because the study included a small number of samples—from 50 men, of whom only 12 had a positive biopsy—the dogs didn’t achieve the sensitivity and specificity that other canines have been able to, but it’s a promising first step, according to Marc Aubreville, a professor of image understanding and medical application of artificial intelligence at Germany’s Technische Hochschule Ingolstadt, who wasn’t involved in the work. “It’s really interdisciplinary what they’re ...
Our Favorite Cancer Stories of 2021
This year revealed just how much scientists have learned about the disease, from how animals become naturally cancer-resistant to how tumor cells harness extracellular DNA to develop rapid drug resistance.


NEIL POLLOCK, MEDICAL DETECTION DOGS; © MURPHYSCIENCEART.COM; KRISTEN TURNER; UNIVERSITY OF ROCHESTER; O. SAYDAM ET AL., FRONT CELL DEV BIOL, 2021.

The Scientist ARCHIVES
Become a Member of
Keywords
Meet the Author
Amanda Heidt
Amanda was an associate editor at The Scientist, where she oversaw the Scientist to Watch, Foundations, and Short Lit columns. When not editing, she produced original reporting for the magazine and website. Amanda has a master's in marine science from Moss Landing Marine Laboratories and a master's in science communication from UC Santa Cruz.View full profile