What the new lameness tech can and can’t tell your vet

AI gait apps, standing CT and PET scanners are changing how vets track down lameness. A new Utah State University guide runs through the options, and we looked at the research behind each one to see how much it really adds.

Lameness is one of the most common and costly problems in horses, and the hardest cases are the mild ones. A horse that is clearly lame on one leg is easy to spot. A horse that is slightly short behind, or just not jumping the way he used to, is far less so.

This difficulty is well documented, and very familiar to most owners. In a 2010 University of Missouri study, 131 horses were each assessed by between two and five experienced vets. When the lameness was moderate or severe, the vets mostly agreed. When it was mild, agreement was poor. Later work by the same group found vets agreed on a lame foreleg about three times in four, but on a lame hind leg only about two times in three.

There’s a second problem with this. When Swedish researchers fitted motion sensors to 222 riding horses that were in regular work and that their owners believed were sound, 72.5% showed movement asymmetries above the thresholds normally used to flag lameness. Not all of those horses are in pain, as some asymmetry may simply be how that horse moves, but it shows how much the eye can miss, and why an objective measure is useful.

Utah State University Extension’s fact sheet, New Frontiers in Diagnosing Equine Lameness, by Hyrum Hoopes and Dr Karl Hoopes, sets out the tools now available to vets. Here’s what each one does, and what the evidence behind it shows.

Nerve blocks still come first

The USU guide focuses on technology, but one step comes before anything else. After watching the horse move, a vet will usually numb parts of the leg with local anaesthetic, working up from the foot, to find where the pain is. If the lameness improves after a block, the pain is likely to be in the area that was numbed. Vet references describe these diagnostic nerve blocks as the gold standard for localising pain.

However, they aren’t foolproof. Anaesthetic can spread to nearby structures, so results sometimes need checking with a second block on another day. Yet, the step does matter, because imaging only answers the question you point it at. Scanning a leg before the pain has been localised can reveal changes unrelated to the lameness and lead to the wrong diagnosis.

AI gait analysis

Systems such as the Lameness Locator, which uses sensors attached to the horse, and Sleip, a smartphone app that analyses video, measure how evenly the head and pelvis move. They then flag which leg is likely to be lame and by how much.

The evidence from reviews of this technology is encouraging but still in the early stages. A 2022 German study tested an AI system on smartphone videos of 22 lame horses and eight sound ones on a circle. Compared with the vets’ clinical assessment, it identified foreleg lameness in every case and hind leg lameness about 90% of the time, using the stifle as a reference point. It struggled in direct sunlight, in darkness and with very evenly coloured coats, and the authors said it needed a much larger dataset.

A 2026 study compared two vets with Sleip and two sensor systems on 25 horses. The objective systems agreed with each other more closely than the vets agreed with each other, scoring 0.84 against 0.73 on a scale where 1 is perfect agreement. Both the vets and the machines found hind legs harder than front legs. But a study of 22 reining Quarter Horses that compared Sleip with a validated sensor system found the gap between the two systems’ readings was bigger than the thresholds used to decide which leg is lame. In other words, two systems can give different answers on the same horse.

The USU guide’s own caution is the key one: these tools show that a horse is moving unevenly, not why. A vet still has to find the cause. One practical use is comparing the same horse before and after a nerve block, where a number takes some of the guesswork out of judging whether the lameness has improved.

X-rays and ultrasound

Digital X-rays are quick, relatively cheap and portable. They are the usual first look for fractures, arthritis, laminitis and hoof balance, and the images are easy to share with your farrier. They don’t, however, show soft tissue such as tendons and ligaments.

Ultrasound is the everyday tool for tendons and ligaments. It shows the surface of bones but can’t see inside them, and it has blind spots inside the hoof. In a 2023 Colorado State University study of 70 ultrasound and MRI exams from 45 horses, ultrasound caught 85% of the deep digital flexor tendon tears that MRI found, but it also flagged problems MRI didn’t confirm, and its accuracy for the navicular bursa and the collateral sesamoidean ligament was only 61–67%. The authors concluded that damage further down the tendon and to other tissues inside the hoof is unlikely to be found without MRI.

Standing CT and MRI

CT uses X-rays to build a three-dimensional image and is best for bone. Until recently it meant a general anaesthetic for the horse. Standing cone-beam CT systems, available since 2021, allow it to be done under sedation, although cost and space still limit how many clinics have one.

MRI is the tool for soft tissue inside the foot and fetlock, and it also shows bone bruising. High-field MRI gives the sharpest images but needs a general anaesthetic. Low-field MRI can be done with the horse standing. The USU guide says both give images good enough for clinical decisions, although a 2020 UK study, in which ten experienced observers graded foot scans from different systems, found image quality differed significantly between them.

CT and MRI often complement each other. Research at the University of Cambridge comparing the two in fetlocks found CT was better at picking up bone lesions, while MRI was better for soft tissue injury and the fluid build-up that comes with bone damage. For suspected bone injury, the best picture came from using both. A separate study at the University of Liège found standing CT added diagnostic information to standing MRI in 71% of cases.

Bone scans and PET

Nuclear scintigraphy, or a bone scan, involves injecting a small amount of radioactive tracer, which collects in areas of active bone change and is picked up by a gamma camera. The images are less detailed than CT or MRI, but a bone scan is good at showing where a problem is, particularly in areas buried under muscle such as the pelvis. As with any imaging, the findings need to match the clinical picture. A hot spot showing on nuclear scintigraphy is not a reason to treat without further supporting evidence.

PET (positron emission tomography) is a more precise form of nuclear imaging. UC Davis became the first veterinary facility to use it in horses in 2016. At Santa Anita Park in California, a standing PET scanner performed 188 scans on more than 100 racehorses in its first year. When the fetlocks of 32 horses were scanned with both PET and a bone scan, seven independent observers detected and described abnormalities more consistently on PET, especially in the sesamoid bones, which are often involved in catastrophic fetlock fractures. PET findings have helped vets there decide when to rest or retire higher-risk horses. Availability of this technology is still very limited, and as far as we know, not available yet in South Africa.

What it means for your horse

• The basics still come first.A thorough lameness exam and nerve blocks tell your vet where to look. Imaging is then aimed at that area.
• A gait app is not a diagnosis.Many sound-looking horses move asymmetrically. If an app flags something, talk to your vet before panicking or acting on it.
• X-rays and ultrasound can miss problems inside the hoof.If a foot lameness has no clear answer, ask whether MRI or CT referral is worthwhile.
• No single scan shows everything.Different tools show different tissues, which is why vets sometimes recommend more than one.

The USU guide argues that earlier, more precise diagnosis reduces pain, shortens recovery and stops small problems from becoming career-ending injuries. That’s a reasonable assumption to make, but the guide sadly doesn’t cite studies that measure those outcomes directly.

---

Sources

• Utah State University Extension, New Frontiers in Diagnosing Equine Lameness (August 2026): https://extension.usu.edu/equine/research/new-frontiers-in-diagnosing-equine-lameness
• Keegan et al. 2010 and 2013, summarised in Vetmeduni Vienna thesis: https://phaidra.vetmeduni.ac.at/api/object/o:783/download
• Rhodin et al. 2017, 222 owner-sound horses (cited in Equine Veterinary Journal): https://onlinelibrary.wiley.com/doi/full/10.1111/evj.12715
• Vet Times, First opinion approach to equine lameness diagnosis: https://www.vettimes.com/news/vets/equine/first-opinion-approach-to-equine-lameness-diagnosis
• Vet Times, An approach to diagnosing lameness in equine patients: https://www.vettimes.com/news/vets/equine/an-approach-to-diagnosing-lameness-in-equine-patients
• Feuser et al. 2022, AI for lameness detection, Animals: https://pmc.ncbi.nlm.nih.gov/articles/PMC9597839/
• Agreement between veterinarians and three objective systems (2026): https://pmc.ncbi.nlm.nih.gov/articles/PMC13447877/
• Comparing IMUs to markerless video analysis in reining Quarter Horses: https://pmc.ncbi.nlm.nih.gov/articles/PMC10610735/
• Barrett et al. 2023, ultrasound vs MRI in the equine digit: https://ouci.dntb.gov.ua/en/works/98QNxDx7/
• Byrne et al. 2020, MRI image quality by system (abstract): https://madbarn.com/research-topics/imaging-techniques/page/11/
• Lin, University of Cambridge thesis on CBCT, CT and MRI of the fetlock: https://www.repository.cam.ac.uk/items/6dc258a9-74e8-4ec6-a8b2-21f62bbd5e9e
• University of Liège, CBCT vs low-field MRI: https://orbi.uliege.be/handle/2268/337876
• UC Davis, First year of equine PET scans at Santa Anita: https://www.vetmed.ucdavis.edu/news/first-year-equine-pet-scans-santa-anita-success

---