Veterinary medicine is on the verge of losing one of its most tedious rituals. For decades, the microscopic hunt for heartworm larvae in a drop of canine blood has demanded a technician’s patience, a well-trained eye and a willingness to stare down a microscope while other patients wait. A new study published in Parasites & Vectors suggests that this chore can now be handed to a machine that watches the blood move, thinks about what it sees and delivers a verdict in about six minutes. The result, its developers report, is a fully automated test that is more than ten times more sensitive than the manual method it is designed to replace, and it costs less than ten dollars to run.
The work, led by Britt Ripley, Mackenzie Smith, Alex Chen, Kevin Calcote and Paul Slusarewicz of Parasight System Inc., together with collaborators at North Carolina State University and the University of Georgia, tackles a stubborn gap between what veterinary guidelines recommend and what busy clinics actually do. The American Heartworm Society and the Companion Animal Parasite Council advise that dogs over seven months of age be tested annually for heartworm using both an antigen test and a microscopic examination for microfilariae, the threadlike offspring of the adult worm Dirofilaria immitis that circulate in the bloodstream. Antigen testing is easy: lateral-flow assays require minimal equipment and can be run in the clinic. Microscopy is not. The problem, the researchers argue, is that the microscopic half of this dual-testing strategy has become such an inconvenience that some veterinarians skip it entirely, potentially missing patent infections that antigen tests alone can fail to catch.
The anatomy of the problem is worth understanding, because it explains why the new approach is technically interesting. The simplest microscopic option, the wet mount, involves placing roughly 20 microliters of blood on a slide, covering it with a coverslip and scanning the entire area at 100-fold magnification, looking for the sinuous, wriggling larvae that betray an active infection. Its virtue is speed and simplicity; its weakness is arithmetic. Twenty microliters is a vanishingly small fraction of a dog’s total blood volume, so a lightly infected animal may simply have no larvae in the sample being watched. More sensitive techniques exist, including the modified Knott test, in which blood is hemolyzed and centrifuged to concentrate any larvae into a pellet, and membrane filtration, in which hemolyzed blood is drawn through a fine mesh that traps the parasites. Both examine far more blood, but both kill or pacify the larvae in the process, robbing the analyst of the telltale motion that makes them easy to spot. They also demand additional time, multiple slides and the kind of motivated, monotony-tolerant expertise that is in short supply in a clinic juggling appointments. Fluorescent staining with acridine orange or fluoresceinated lectins can enhance visibility, but most veterinary practices do not own a fluorescence microscope.
The team’s solution repurposes an existing piece of hardware. The Parasight All-in-One, or AIO, is a veterinary parasitology analyzer originally built to count the eggs and oocysts of gastrointestinal parasites in fecal samples. It works by labeling parasitic objects with a fluorescent dye, imaging them with a fluorescence-based optical column and using computational image analysis to identify and count objects of clinical interest. Adapting that machinery to moving larvae in blood required a clever rethinking of what the instrument should look for. Instead of photographing static eggs, the AIO now records short videos of larvae crawling across a filter and infers their presence from the one property that distinguishes them from everything else in a blood sample: motion.
The workflow is straightforward enough to describe in a sentence. One milliliter of whole blood is mixed with five milliliters of deionized water, which bursts the red blood cells by osmotic shock while leaving the tough nematode larvae intact. The hemolyzed suspension is poured onto a single-use sample chamber containing an integrated stainless-steel mesh with five-micrometer pores, and gentle vacuum draws the liquid through, leaving microfilariae stranded on the mesh surface. After three washes with phosphate-buffered saline, 0.1 milliliters of a dilute sodium fluorescein solution is added. When illuminated by the instrument’s array of blue LEDs, the fluorophore emits green light that passes through a long-pass emission filter built into the instrument’s macro lens and lands on an 18-megapixel monochrome camera sensor. A fifteen-frame video is then recorded at three frames per second.
Here is where the physics becomes subtle. The microfilariae are only about six micrometers wide, and the AIO’s optical column operates at just one-fold magnification, which would seem to make them invisible. But the instrument’s high-resolution lens and sensor achieve an effective resolving power of about 2.5 micrometers, sufficient to detect a larva indirectly: as light refracts through the worm’s body, it distorts the fine pattern of the mesh underneath, and that distortion betrays the larva’s position and shape. Raw video is nonetheless low in contrast, so the researchers turned to computational enhancement. For human review, the video is processed with FFmpeg software to highlight only pixels that change significantly between adjacent frames, a difference-processing technique that makes the wriggling larvae leap out of the frame. For automated detection, the video is fed to OpenCV, an open-source computer vision library. A nearest-neighbor-based background/foreground segmentation algorithm identifies moving objects in each frame by monitoring differences in the moving averages of each pixel’s channels. Moving objects are masked, frames whose average intensity exceeds a threshold are discarded to eliminate artifacts from sudden lighting changes or camera shake, and the remaining frames are summed into a composite image in which white pixel patches that pass a size filter are counted as microfilariae.
The performance numbers are striking. When the team diluted infected blood serially and counted each dilution five times with both methods, the manual wet mount showed a clean linear relationship between count and parasite concentration, as expected. The automated method tracked the same samples closely but with roughly thirteen times the magnitude, reflecting the fifty-fold larger blood volume it examines, partly offset by losses during processing. Within the linear range, automated and manual counts correlated superbly, with a coefficient of determination of 0.9851 and a regression slope of 13.2, meaning the machine counted more than ten times as many larvae in the same blood. At low concentrations the gap became decisive. At a dilution corresponding to roughly 50 microfilariae per milliliter, the wet mount detected larvae in only six of ten replicate samples, a sensitivity of 60 percent, while the automated test was positive in all ten. After a further twofold dilution, the manual method’s sensitivity collapsed to 10 percent while the automated test remained at 100 percent. Neither method flagged any of ten uninfected blood samples, giving both a specificity of 100 percent with respect to filarioid detection.
The team also verified that refrigerated storage did not distort the comparison. Counts performed on the same blood stock on day 0 and again on days 1, 2, 3, 4 and 6 showed no statistically significant drift by either method, though by day six the larvae were visibly more lethargic in their movements. A one-way analysis of variance confirmed no significant differences across the time course for either technique, allowing the multi-day experiments to proceed on firm footing.
The study is candid about limitations. The automated method’s counts plateau at high larval concentrations, above roughly 1,000 microfilariae per milliliter, because the worms begin to overlap and abut, forming clumps that the size-filter in the detection algorithm ignores. That compressed dynamic range could matter in the future if anthelmintic resistance, already documented in D. immitis and in canine hookworms, drives guidelines toward quantitative pre- and post-treatment counts to monitor drug efficacy. The authors are developing algorithms to flag suppressed counts and are exploring deep-learning classifiers that may better separate juxtaposed larvae. The system also cannot distinguish D. immitis from the closely related and more zoonotic D. repens, or from the largely harmless Acanthocheilonema reconditum, though pairing the test with antigen testing should resolve species identity. And because the instrument’s depth of field captures only about half the mesh surface in full focus at a single focal length, the true recovery rate is around 20 percent of the larvae loaded; shooting multiple videos per test should at least double that figure.
Even in its current form, the economics are persuasive. In-clinic antigen assays cost American practices roughly 20 to 30 dollars each, and manual microscopic methods run about 10 to 15 dollars once technician time is counted. The automated test retails for under 10 dollars, and because it is essentially hands-off after the sample is poured, it frees staff to do other work during the six minutes the machine runs. The researchers plan next to benchmark the optimized system against the modified Knott test, the acknowledged gold standard, and to validate it on samples from naturally infected animals. If those efforts succeed, the annual heartworm checkup may finally become what guidelines have long assumed it to be: a test veterinarians actually perform.
Subject of Research: Automated in-clinic detection and enumeration of Dirofilaria immitis microfilariae in canine blood using fluorescence videography and computer vision
Subject of Research: Biology
Article Title: Fluorescence videography for the rapid and automated in-clinic detection of microfilariae
Article References: Ripley, B., Smith, M., Chen, A., Calcote, K., Moorhead, A. R., Campbell, E., & Slusarewicz, P. (2026). Fluorescence videography for the rapid and automated in-clinic detection of microfilariae. Parasites & Vectors, 19(1), Article 355. https://doi.org/10.1186/s13071-026-07486-y
Image Credits: AI Generated
DOI: 10.1186/s13071-026-07486-y
Keywords: Dirofilaria immitis, Heartworm, Microfilaria, Veterinary diagnostics, Fluorescence videography, Computer vision, Parasight All-in-One, Automated parasite detection, Wet mount comparison, Canine blood testing
Cite Scienmag News
APA MLA Chicago
Ophelia Keating. (September 10, 2026). Fluorescence videography enables rapid automated in-clinic microfilariae detection. Scienmag. https://scienmag.com/fluorescence-videography-enables-rapid-automated-in-clinic-microfilariae-detection/
Ophelia Keating. “Fluorescence videography enables rapid automated in-clinic microfilariae detection.” Scienmag, 10 September 2026, https://scienmag.com/fluorescence-videography-enables-rapid-automated-in-clinic-microfilariae-detection/. Accessed 10 September 2026.
Ophelia Keating. “Fluorescence videography enables rapid automated in-clinic microfilariae detection.” Scienmag. September 10, 2026. https://scienmag.com/fluorescence-videography-enables-rapid-automated-in-clinic-microfilariae-detection/
Copy citation Download RIS
Tags: automated heartworm testingautomated in-clinic parasite testingcost-effective veterinary testsfluorescence videography in parasitologyfluorescence videography in veterinary diagnosticsheartworm diagnosis advancementsheartworm disease diagnosis in dogsheartworm microfilariae detectionin-clinic veterinary diagnosticsin-office veterinary diagnostic toolslow-cost heartworm testing methodsmicrofilariae detection technologymicrofilariae identification using fluorescence imagingparasite detection automationparasitology imaging techniquesrapid heartworm detection technologyrapid in-clinic parasite screeningsensitive automated blood analysis for dogssensitive blood parasite testingveterinary diagnostic innovationsveterinary disease screening innovationsVeterinary microfilariae detectionveterinary parasitology advancements
