Beagles with supercharged noses are helping a health startup screen people for cancer using breath samples, an approach that mixes animal talent with machine learning to catch disease earlier. The test asks patients to breathe into a mask, sends that mask to a lab, and relies on trained dogs—and the data they generate—to flag potential cases for follow-up. Company leaders highlight low routine screening rates and hope this option nudges more people toward early detection. Clinical trials claim high accuracy and the possibility of spotting cancer at very early stages.
Iceman, like other beagles, has an extraordinary sense of smell and is part of this new screening effort. Spotitearly built a system around that canine talent to create a simple at-home pre-screening option and to increase the odds of spotting cancer sooner. “Only 14% of the newly diagnosed cancers in the U.S. are done so through the routine screening. So not a lot of people are actually getting those tests on time,” says CEO Shlomi Madar. The company positions the test as a way to reach people who miss traditional screening windows.
The sampling routine is deliberately low friction. “All we ask from the patient is to wear a mask for a few minutes, breathe into it, put it in a capsule and send it back to the lab.” Once the lab has the sample, trained beagles evaluate the scent for markers tied to the four most common cancers: breast, colorectal, prostate, and lung. Early trial results reported by the company showed dogs identifying cancer in samples with about 94% accuracy, a figure that grabbed attention in the medical tech space.
The team says the dogs are picked because their noses are vastly more sensitive than ours, capable of detecting quantities humans would miss by orders of magnitude. In practice, that canine sensitivity is paired with AI to interpret behavior and physiological cues while dogs work. “We monitor the dogs’ behavior and physiological signals. So, we look at things like their heart rate, their acceleration,” Madar explains, “And all that plethora of information is being fed into a machine learning algorithm. So, it actually gets better with time.”
Training is rigorous: only beagles already skilled as sniffer dogs are considered, and each goes through targeted conditioning to recognize scent signatures linked to specific cancers. To guard against error, the lab runs each sample through several dogs rather than relying on a single response. “We use a lot of redundancy. So, every sample is being sniffed not by one dog, but the entire pack, and more than once,” Madar notes. That layered approach aims to reduce false positives and boost confidence in flagged results.
When a sample is flagged, a healthcare professional reviews the finding and recommends next steps, typically more definitive diagnostic testing. The company stresses this screening is not a replacement for medical care but a triage tool to prompt timely doctor visits. While still in U.S. clinical studies, the test kit is available for pre-order and is expected to roll out commercially next year, offering another route for people who avoid or miss standard screens.
Spotitearly also says the model can expand: train new packs on different cancer types and the scope of detectable disease widens. The program carries an unexpected social benefit for dogs that don’t complete training; those animals are donated to families with special needs. The approach blends biology and tech in a pragmatic, human-centered package that aims to convert canine capability into actionable health signals.
