The latest technology for detecting COVID-19 could involve cold, wet noses and a few quick sniffs—but not from the people being screened.

This week, the Miami Heat is debuting a vanguard of coronavirus-attuned canines at the entrance to AmericanAirlines Arena. They'll walk up a line of socially distanced spectators waiting to get in, and, if all goes according to plan, sit down whenever they encounter anyone who is infected. These COVID sniffers have also been deployed at airports in Europe.

“If you think about it, detection dogs are not new,” Matthew Jafarian, the Heat’s executive vice president for business strategy, told the Associated Press. “You’ve seen them in airports, they’ve been used in mission critical situations by the police and the military. We’ve used them at the arena for years to detect explosives.”

Representatives from the New York Knicks and Brooklyn Nets did not return calls to discuss potential dog deployment at Madison Square Garden and the Barclays Center.

But they would certainly be right to wonder about the effectiveness of these doggie detectors, Dr. Audrey Odom John told Gothamist. She leads an infectious disease team at the Children's Hospital of Philadelphia that is exploring how dogs can be used to identify people with COVID-19, malaria, and other diseases. Dr. John laid out the pros, cons, and long-term promise of nostril-based screening.

Many people are aware of dogs' superpowers in screening people for drugs and explosives, and they might have heard that some dogs can sniff out diabetes. Where we are with infectious diseases and, in particular, the SARS-CoV-2 coronavirus?

It turns out that infectious diseases themselves give off their own odor. We had previously done some work in malaria and found that malaria parasites make volatile odors—their own perfume, if you will.

But in the case of viruses, it probably has more to do with the host immune response. These reactions by the body generate specific compounds that dogs can detect through smell. Our work involves trying to identify what those compounds are.

How do you do that?

We have been working with an investigator named Cynthia Otto, who directs the Penn Vet Working Dog Center. She and her team were able to use samples from children and adults with and without SARS-CoV-2 infection and train the dogs on those samples...with really high sensitivity and specificity.

That’s because sensitivity is the rate of true positives, while specificity is their ability to distinguish SARS-CoV-2 from other germs, right? How did the dogs do? What was their batting average?

There were eight dogs -- seven Labrador retrievers and one Belgian Malinois. Most of them were over 90 percent accurate, and one of them never missed a beat and approached 100% accuracy.

That's quite a good dog! But you used saliva and urine samples in a lab. Do you think they could do as well outdoors, sniffing people wearing pants and coats, cologne, and perfume—with the smell of stale beer on the pavement?

There's testing in a controlled setting, and then there's the real world experience. So, if we have a dog who does around 100% in the lab, can she do this in a field setting? It's certainly likely to be a harder thing for them to do when there's distractible odors and other distractions in the environment.

Given that, do you think the sports arena setup is likely to be accurate?

Without seeing their data and what tests that they've done to show their own accuracy in a complex environment like a stadium, it's hard to know how good they're going to do. But I do think it's very likely the dogs can do it. They do it with ovarian cancer, for instance, which would seem to be an even more complicated thing to sniff out.

You're not an ethicist, but I wonder what you think of putting dogs in this situation. Here you've got spectators spending hundreds of dollars on tickets, and you've got dogs who maybe don't bat .1000. Say they miss five or ten percent: They're still doing pretty well, but that's not much comfort if you've got courtside tickets. Are we asking too much of the dogs?

I would see using these animals as a preliminary test, a little bit like a thermometer. If the dog screens positively on you, you would then go to the little room, and you'd get one of the rapid antigen tests that in 15 minutes will tell you Yes or No. That's how I would imagine using it, not with the dog as the only thing.

You know, this internet meme-type thing went around a while back about how to appear intelligent in meetings, and one thing they suggested was that no matter what the issue is, always ask, "But will it scale?" So I'm asking you: will it scale—this system with dogs here, there, and everywhere, screening people for COVID?

I don't think it's too likely. There are only so many dogs and only so many trainers and only so many handlers. The scenario that we think is most viable in terms of putting this out at scale is doing more of a breathalyzer-type test. There's on-the-shelf detection technology that can rapidly translate to point-of-care sensors. We've identified specific [blood] biomarkers for COVID, and the next step is to validate those with new children and adults.

The big question is whether you can identify individuals who are asymptomatically infected. Being able to catch people before they're infectious would be a real benefit.

So the dogs are just doing the leg work for you—they're a means to an end. These breathalyzer tests are the real goal?

We've been essentially approaching this two ways, both trying to use mass spectrometry, which is a technique to look at all the different compounds that are present in breath, to see if the machine can 'smell' the difference between people with and without infection.

And at the same time, we're also using dogs. They actually do all the math and the mass spectrometry by themselves just with their noses.

And the dogs tell us right away that this is a viable way to make a new diagnostic test. If dogs can tell the difference, we should be able to make a machine that can tell the difference.