This algorithm has been designed to spot these genetic indicators across many dog breeds with different forms of cancer: say, a bloodhound with lymphoma or a golden retriever with hemangiosarcoma. When the PetDx team began their study, they used blood from 224 of the dogs—ones known to either have cancer or not—to refine their algorithm. This “training set” helped PetDx determine a threshold for each genomic variation, defining what they should call signal rather than just noise.
Then they unleashed the algorithm onto data from the 876 other dogs. For each, it would render a binary answer: yes cancer, or no cancer. (For most cancers, it wouldn’t identify which form it was.) Pet owners and vets already knew whether their dogs had a cancer diagnosis, but the PetDx researchers did not, so as to not influence their analysis. The team then compared their results to the vets’ previous diagnoses.
Overall, the algorithm caught about 55 percent of the total cases, representing 30 varieties of cancer. It was most accurate when identifying late-stage cancers, or any stage of aggressive types. For the three most aggressive cancers, OncoK9 gave the right answer 85 percent of the time. For all metastatic cancers, the figure was over 80 percent. For the eight most lethal cancers, it was 62 percent. “Any of these numbers,” says Flory, “are a huge improvement over what the current paradigm is.” She estimates that’s between 3 and 10 percent for early-stage cancers.
Cici Pepperoni turned out to be one of the positive tests—a surprise to Inserra and their vet. The company notified the vet, who scheduled scans in May 2021 that revealed growths in the dog’s lungs and a mass in her heart. “We realized, oh my gosh, not only does she have cancer, but she has a lot of cancer,” Inserra says.
It turned out to be hemangiosarcoma—a death sentence. Hemangiosarcoma generally makes itself known as an emergency, when a tumor suddenly bleeds into the space around the dog’s heart. “They're collapsed. They have anemia. They're bleeding,” says Flory. Their family must rush to get help. “All of a sudden, they're at the ER clinic having to really make this life or death decision.”
But since OncoK9 caught Cici’s case many months early, Inserra’s family had options. They could try palliative care to lower the risk of bleeding, or chemo to maybe shrink the tumor. They chose the former to keep Cici more comfortable. By July, she had unmistakable symptoms. Weight loss. Coughing. “She started getting a little bit more spacey,” Inserra says. The vet noticed fluid in her chest. Inserra’s family prepared to say goodbye.
They brought Cici to San Diego’s Dog Beach so she could run around and have fun as best she could. They brought the most desirable treats they could find, including, of course, pepperoni pizza. She wasn’t eating much anymore, but Cici didn’t disappoint: “She took one little bite,” Inserra says. “That was our happy last moment.”
Cici’s case illustrates the thorny ethical debate over early cancer diagnoses for dogs. Dogs don’t have as many treatment options as people do. And blood tests still need to be confirmed by imaging and biopsies, so pet owners who can’t afford more testing may opt to just wait and see. “If it works, it's going to be fantastic for the dogs,” says Elinor Karlsson, a computational biologist with the University of Massachusetts Medical School who studies the genomics of dog diseases. (Karlsson is not involved with PetDx and is an unpaid scientific adviser for canine cancer analysis startup FidoCure.) But we have to be careful, Karlsson says: “Are you just stressing a lot of people out without being able to help?”