In December 2019 California’s Santa Anita Park became the first racetrack to install a positron emission tomography (sodium fluoride PET, 18F-NA PET) unit. This PET unit allows veterinarians to image standing, sedated horses, producing easy-to-interpret three-dimensional (3D) images. By directly comparing PET scans to traditional nuclear scintigraphy (bone scans), veterinarians found that PET scans could provide more information about a horse’s fetlock than bone scanning in certain situations.

“Sodium fluoride PET is like a 3D bone scan: The radiotracer accumulates in the bone and veterinarians look for areas of increased uptake,” said Mathieu Spriet, DVM, MS, Dipl. ACVR, ECVDI, ACVR-EDI, associate professor of diagnostic imaging at the University of California, Davis, and the radiologist leading equine PET development.

Other imaging modalities such as bone scans are already available to racehorses, but those techniques have limitations and can miss certain lesions, putting horses at risk for injury and potentially catastrophic breakdown. Spriet and colleagues therefore studied the ease of use and clinical value of PET scan in Thoroughbred racehorses with lameness or performance issues related to the fetlock. They compared those results to bone scans to help clarify when PET might be a better tool than nuclear scintigraphy in the track setting. Spriet presented their findings at the 2021 American Association of Equine Practitioners (AAEP) Convention, held Dec. 4-8 in Nashville, Tennessee.

In the study Spriet’s team imaged 72 fetlocks in 33 racehorses using both 18F-NA PET and nuclear scintigraphy within two weeks of one another. Seven observers (racetrack practitioners, surgical and imaging residents, and a board-certified radiologist) independently reviewed the scans, grading 10 distinct regions of interest on a scale of 1 to 3. Regions of interest included the medial and lateral condyles of the cannon bone, the proximal sesamoids, and the long pastern bone. They then calculated interobserver agreement, which is a measure of how closely the seven individuals shared the same observations.

“If perfect interobserver agreement was observed, the kappa weighted value would be 1. Both modalities had a kappa weighted greater than 0.6, which is good! It indicates substantial agreement,” Spriet explained.

Agreement among the seven observers for reading the PET scans was higher at 0.73 than for nuclear scintigraphy at 0.61.

For the condyles, which are often a primary area of interest in Thoroughbred racehorses (condylar fractures occur relatively commonly in this population), the kappa weighted was 0.75 for PET and only 0.53 for scintigraphy. For the proximal sesamoid bones, kappa weighted was a little lower for both, 0.62 for PET and 0.53 for scintigraphy.

“The independent observers also recognized overall more areas of uptake on the PET scans than scintigraphy,” said Spriet. “For the proximal sesamoid bones, PET identified abnormalities in 16% of cases, whereas scintigraphy only detected abnormalities in 2.8%.”

Spriet suggested the main reasons for these differences were the 3D nature of 18F-NA PET providing more information than the 2D scintigraphy scans.

“18F-NA PET seems to be a particularly helpful tool for sesamoid assessment, which is important because the majority of catastrophic breakdowns in racehorses involve the sesamoid bones,” he said.