Overcoming veterinary diagnostic challenges with AI-driven POC technologies

By leveraging new methods, diagnostics can become more efficient, accurate, and cost-effective

As the emphasis of veterinary diagnostics shifts away from curative medicine toward prevention, as well as early detection and disease management, it has become clear the standard diagnostic tools used in veterinary care have yet to embrace today’s technological advancementsAs the emphasis of veterinary diagnostics shifts away from curative medicine toward prevention, as well as early detection and disease management, it has become clear the standard diagnostic tools used in veterinary care have yet to embrace today’s technological advancements. The typical methods of diagnosis lack the innovation that exists in human medicine. It is time to bring these advancements to the veterinary sphere.

Veterinarians today fall into two primary classes: those treating farm animals in the field and those treating companion animals in a clinic. Each of these requires its own set of expertise and tools, and each faces unique diagnostic challenges.

Methods of diagnosis differ greatly in these two groups. Veterinarians treating livestock and other farm animals generally draw blood samples in the field and send them to a centralized laboratory for analysis and await results. They often perform an additional on-site blood smear, which is also sent to the lab, although these tests are often of low quality because of the conditions under which they are performed.

Smaller veterinary clinics treating companion animals usually utilize in-house hematology analyzers, which are smaller versions of their lab-grade counterparts, to detect pathogens and cell abnormalities. This equipment requires maintenance, calibration, and may be financially draining for low-volume clinics.

The use of new technologies, such as point-of-care devices powered by artificial intelligence (AI) and machine vision algorithms, offer solutions to the most pressing problems in veterinary diagnostics, whether in the field or in a clinic. By leveraging these new methods, diagnostics can become more efficient, accurate, and cost-effective. 

Lack of efficiency

Technological advancements and the search for greater efficiency led diagnosis and detection of pathogens to take place in large centralized labs, which have large hematology analyzers that can work through high volumes of samples each day. Veterinarians practicing in the field generally take samples from animals and send them to these labs, which take time to process, possibly delaying diagnosis. We all know the earlier a pathogen is identified and diagnosed, the better the outcome.

For clinics, the process of sending a sample to a centralized lab and waiting for results can delay diagnosis and treatment. Thus, most clinics have in-house analyzers they can use to perform tests. However, this equipment requires routine maintenance, calibration, and expertise to operate. Further, it is possible they do not deliver the most accurate results, as they employ a relatively basic technology that is limited in its ability to discern different pathologies.

Consequently, many veterinarians perform a manual blood smear and inspect the sample under a microscope. Moreover, the cost of these devices can add up for clinics, as the reagents needed to perform assays can be expensive, and they often expire faster than they can be used.

For low-volume clinics that run a handful of samples a day, hematology analyzers are a costly luxury that can be seen as impractical.

Lack of accuracy

Another major challenge when it comes to veterinary diagnostics is that of accuracy and reliability. Most of the technology used today was developed and designed for human samples, and eventually converted for veterinary use. In contrast to the human medical sphere, regulation and performance-validation of veterinary hematology analyzers are very low, and often these tools are not as precise as they ought to be.

Because standard analyzer technology and reagents were developed initially for human samples, they are not optimized for testing animal samples, let alone multiple species. Given these limitations, in-house analyzers can extract a handful of features from animal blood samples, and often miss valuable and telling information.

As many veterinarians don’t feel comfortable relying on these analyzers; they often prefer blood smears, which provide more accurate information. A trained professional performs this test by inspecting a sample of blood under a microscope, looking at the morphological characteristics of cells and their staining properties.

These blood smears are still considered the gold standard and are routinely used to detect pathologies, such as blood cancers and blood parasites. This testing method is more accurate than its automatic counterparts, yet blood smear analysis is laborious, time-consuming, and requires the expertise of a trained professional.

New developments in microfluidics, engineering, artificial intelligence, and medicine are paving the way for diagnostic devices that address each of these issues, both in the field and in veterinary clinics.

Addressing efficiency

Point-of-care (POC) devices are designed to perform diagnostic tests near the patient, in or out of a clinical setting, and deliver virtually immediate results. Having on-demand diagnostic information can allow veterinarians to quickly prescribe medication or map out a treatment plan, beginning the process to recovery as quickly as possible.

The way a veterinary team treats and diagnoses farm animals differs significantly from how they work with house pets.
The way a veterinary team treats and diagnoses farm animals differs significantly from how they work with house pets.

In comparison to centralized laboratory testing, POC devices offer the substantial advantage of timeliness. Instead of requiring a follow-up visit or a phone call a week after the initial test, the animal can be diagnosed on the spot and treatment can begin immediately. In addition, due to the miniaturization of new technologies, as well as their robust design, these devices may be taken into the field for farm animal diagnostics, or implemented in individual clinics with a very small footprint.

Utilizing “lab on a cartridge” technology, POC devices do not require the use of wholesale reagents and are simple to use because each test is run on a factory-calibrated, self-contained reagent cartridge that is easily inserted into the POC instrument for analysis. This saves not only the cost of materials, but also time, as these devices require no calibration or maintenance. Unlike tabletop analyzers, no expertise is needed to run these tests, so they can be done quickly, efficiently, and cost effectively.

Addressing accuracy

Much of the technology used today to analyze blood samples was developed in the 1960s. And although there have been some improvements since, the basic underlying technology has not matured, leaving much to be desired in the way of both efficacy and precision.

POC alone is not enough to overcome the diagnostic challenges outlined above. By utilizing artificial intelligence and machine vision algorithms, new diagnostic instruments can be taught to recognize hundreds of parameters in each sample, wringing out the most information from a minimal amount of blood.

Crucial to accurate performance in a veterinary setting, in which a veterinarian may treat many different species of animals, AI algorithms can be taught and trained to analyze and differentiate between different animal species’ blood samples.

The accuracy and specificity AI enables will eventually surpass the accuracy of a manual blood smear analysis. Instead of performing a blood smear for each sample, the algorithms can learn to identify dog, cat, or cow cells, and spot abnormalities within each group.

These tests also only require very small amounts of blood, as AI-driven analysis is much more sensitive than the human eye to identifying anomalies, and can inspect significantly more cells. The sample is added to the cartridge, which flows the cells in a single-cell plane for analysis (a.k.a viscoelastic focusing), allowing machine vision algorithms to identify each and every cell, making the most of the resources available and using a minimal amount of reagents and materials.

Avishay Bransky, PhD, CEO and cofounder of PixCell Medical, is skilled in microfluidics and POC testing and has extensive industrial experience in applied physics, software, and systems engineering. He is one of the inventors of the Viscoelastic Focusing technique, cell analysis methods, and the microfluidic-based lab-on-a-cartridge. Dr. Bransky holds a BA in physics, BSc in materials engineering, and a PhD in biomedical engineering, all from the Technion Israel Institute of Technology.

Post a Comment