These snoots are made for more than just booping.

SpotitEarly, a biotech startup working to transform cancer screening, just released a new study, pointing to the successful results of using both artificial intelligence and canines to detect cancer early in breath samples.

The researchers call this a “bio-AI hybrid” cancer screening approach.  By using AI to assist cancer-sniffing dogs — which have long been utilized by various laboratories to detect disease — cancers including breast, lung, prostate, and colorectal cancers were able to be detected with high rates of success.

The research, dubbed “The Rainbow Study” has just been published in Nature and marks the largest clinical trial to date for cancer detection using both AI and trained canines.

SpotitEarly’s team analyzed breath samples from over 1,400 individuals between the ages of 22 and 94. The participants provided a non-invasive breath sample in a specialized mask, which was then sent to the lab (which may or may not have been staffed by labradors).

The sniffer dogs did their jobs, but this time, with the help of dynamic sensors, AI-based models, and machine learning algorithms, which, according to SpotitEarly, could mark the presence or absence of cancer odor signatures in just fractions of a second.

AI tech decodes hundreds of real-time signals from the sniffer dogs, analyzing their physical and behavioral responses and integrating it with medical data to deliver an accurate cancer screening.

“Discovering cancer early is critical to recovery. I am pleased to see these promising results and that this new research supports and validates our non-invasive, self-administered cancer screening method,” Len Lichtenfeld, MD, chief medical officer of SpotitEarly and former deputy chief medical officer for the American Cancer Society, said in a statement.

“In combining the unique, scent-based biometrics of dogs and the power of AI, we have successfully unlocked the technology to detect different common cancers, in its early stages, with a single test. I expect that these results will pave the way for the development of a new generation cancer screening test to enhance cancer screening capabilities.”

The test demonstrated exceptional accuracy, achieving high sensitivity markers across the four targeted cancer types, including:

• Breast cancer: 94% sensitivity
• Lung cancer: 97% sensitivity
• Prostate cancer: 97% sensitivity
• Colorectal cancer: 86% sensitivity

The test also showed efficacy in identifying other cancers as well, according to SpotitEarly.

“Taken together, these results are a very encouraging step toward improving cancer survival rates, especially given the ease of obtaining the breath samples used to perform the test,” Lichtenfeld added.

Considering the invasive methods used to screen for a variety of cancers, the success of breath samples signals exciting potential in medicine.

The tests conducted by SpotitEarly are based on the detection of “cancer odor signatures,” which include unique patterns of volatile organic compounds that secrete into the respiratory system in the early stages of cancer. 

While it was already clear that sniffer dogs could help in this arena, integrating their superior snouts with cutting-edge AI could take breath sample screenings to a new, scalable frontier.

Following the success of this study, SpotitEarly plans to collaborate with leading U.S. medical centers in two upcoming clinical trials, which aim to strengthen the evidence for these screening tests.

“SpotitEarly’s solution demonstrates effectiveness in the early detection of cancer,” Professor Nadir Arber, principal investigator and director of the Integrated Cancer Prevention Center at Sourasky Medical Center, said in a statement.

“After rigorously reviewing the clinical validation, I am deeply impressed by its potential to revolutionize cancer screening and set a new standard in early detection.”

‍

Header image courtesy of SpotitEarly