Dogs may soon play a bigger role in cancer screening, with researchers combining their powerful sense of smell with artificial intelligence (AI) to detect possible signs of disease from breath samples.

According to a report by pharmaphorum, health technology company SpotitEarly is developing a non-invasive screening system that uses trained beagles to identify volatile organic compounds, or VOCs, associated with cancer.

Participants breathe into a medical-grade N95-style mask for about three minutes. The mask captures VOCs produced by the body’s metabolic processes before being sealed and sent to a laboratory for analysis.

At the lab, trained dogs sniff samples derived from the masks. Their reactions are monitored using cameras, accelerometers and heart-rate sensors, with the resulting data processed by a machine-learning system.

SpotitEarly CEO Shlomi Madar described the approach as “double AI,” combining animal intelligence with artificial intelligence.

The AI system analyzes subtle changes in the dogs’ posture, movement, timing and heart rate that may be difficult for human handlers to interpret consistently. Each breath sample is also assessed multiple times by different dogs to reduce individual bias.

The system is not intended to replace conventional cancer diagnostics. Instead, SpotitEarly is exploring its use as a pre-screening tool that could identify people who may need further testing.

A multicenter, double-blind study involving 1,400 participants in Israel reported 94.8 percent sensitivity and specificity, according to pharmaphorum. The company is now preparing to expand trials in the United States.

Beagles are primarily used because of their scenting ability and temperament. The dogs undergo four to six months of training and work only for limited periods each day.

SpotitEarly said a single laboratory using its system could potentially process up to 1 million samples annually, although pharmaphorum noted that such capacity has yet to be independently validated at scale.

Researchers are also studying whether the same approach could eventually be used for other diseases that produce detectable VOC signatures.

For now, Madar said the dogs remain central to the system because artificial sensors have yet to match the sensitivity of canine smell.