SpotitEarly, an Israeli biotech firm founded in 2020, is building a novel at-home cancer screening service that combines trained beagles’ sense of smell with artificial intelligence to detect volatile organic compounds in human breath associated with cancer. The company published a double-blind clinical study in Nature’s Scientific Reports involving 1,400 participants, showing about 94 % accuracy for detecting breast, colorectal, prostate, and lung cancers. It plans to launch consumer access through physician networks next year at a price point (≈ $250 per cancer) well below competitors like Galleri, and is now partnering with Hackensack Meridian Health in the U.S. to run a 2,000-participant breast cancer trial (the PINK Study).
Sources: FierceHealthcare, ScienceNews
Key Takeaways
– The SpotitEarly system leverages a hybrid bio-AI model: trained dogs detect odor cues in breath samples, and AI monitors their behavioral/physiological signals to validate cancer presence.
– Its published clinical results (1,400 participants, 94 % accuracy) are focused on four common cancers and will be expanded via further trials, including in collaboration with U.S. health systems.
– If successful and scalable, this kind of non-invasive screening could lower cost and increase accessibility relative to existing multi-cancer detection tests, but adoption hinges on further validation and regulatory approval.
In-Depth
In the quest to catch cancer earlier—when treatment outcomes tend to be better—SpotitEarly is betting on a curious but scientifically grounded alliance: dogs and artificial intelligence. The startup’s approach rests on the biological fact that some cancers appear to emit volatile organic compounds (VOCs)—tiny chemical signatures that carry in breath or exhaled air. Dogs, with their sophisticated olfactory systems, have long been suspected of being able to detect disease signals humans cannot. But reading dog behavior alone is inconsistent and subjective. That’s where AI comes in.
In its recently published Rainbow study (in Nature’s Scientific Reports), SpotitEarly processed breath samples from about 1,400 people in a double-blind fashion. The beagles are trained to sit when they detect cancer odors; meanwhile, cameras, microphones, and sensors record the dogs’ heart rate, breathing, and microbehaviors. A machine learning model integrates those signals to confirm or override the dogs’ behavior. The study reported ~94 % accuracy (sensitivity and specificity) across breast, colorectal, prostate, and lung cancers—including early stages.
That’s a promising start, but by no means proof of broad viability. SpotitEarly is now gearing toward U.S. trials, notably the PINK Study with Hackensack Meridian Health, which is recruiting up to 2,000 participants focused initially on breast cancer detection. The trials aim to further test sensitivity, specificity, scalability, and integration into medical workflows.
In comparison, other multi-cancer early detection (MCED) strategies—such as blood-based tests that look for circulating tumor DNA fragments—are also advancing, but often with higher cost and regulatory uncertainty. The hybrid canine-AI approach offers a potentially lower-cost, non-invasive alternative, but it will need to overcome challenges of standardization, reproducibility, variable dog performance, and regulatory scrutiny.
If SpotitEarly’s model works at scale, it could shift cancer screening toward greater accessibility, earlier detection, and lower burden on patients and health systems. Still, skeptics will insist on independent replication, long-term outcomes, and validation across diverse populations before it can be adopted in mainstream practice.