AI & CT Scan Advance Dog Heart Diagnosis in 2026

Introduction

A seven-year-old Afghan Hound’s case of fever and cough led to the first-ever canine CT scan description of a rare heart condition. Simultaneously, French researchers developed an artificial intelligence system that can measure a dog’s heart size from an X-ray in seconds. Together, these 2026 studies, published in Veterinary Radiology & Ultrasound and The Veterinary Journal, highlight the advancing tools for diagnosing complex cardiac and pericardial diseases in dogs.

CT Scan Reveals a Rare Lung-to-Heart Sac Connection

Veterinarians at VCA West Coast Specialty and Emergency Animal Hospital in California diagnosed a condition so rare its CT features had never been documented in dogs. The patient, a female Afghan Hound, presented with classic signs of infection: fever, lethargy, and cough. Standard X-rays showed suspicious soft tissue in the lung and a small amount of gas around the heart, a condition called pneumopericardium.

To find the source, the team performed a computed tomography (CT) scan. It revealed the definitive cause: multiple bronchopericardial fistulas. These are abnormal, tunnel-like connections between the small airways (bronchi) of the lung and the pericardium, the tough, protective sac surrounding the heart. In this case, the fistulas, measuring up to 6.0 mm wide, originated from the left cranial lung lobe. Air and fluid from the inflamed lung were leaking directly into the pericardial sac, causing it to dilate. This put pressure on the heart and led to the dog’s clinical signs. Following precise surgical removal of the affected lung lobes and most of the pericardium, the dog recovered completely, with a follow-up ultrasound two weeks later showing no remaining fluid.

AI Automates a Cornerstone Measurement for Heart Enlargement

While the California case involved advanced imaging for a rare problem, a separate study addresses a very common need in general practice: consistently measuring heart size on X-rays. Cardiomegaly, or heart enlargement, is a major indicator of disease, and the Vertebral Heart Size (VHS) is a standard screening tool. Veterinarians measure the heart’s length and width on an X-ray and compare it to the length of specific vertebrae behind the heart. A high score suggests enlargement.

Manual measurement takes time and can vary between observers. Researchers from the aivancity School of AI & Data for Business & Society in France built an integrated AI framework to solve this. Their system uses a deep learning computer vision model trained on annotated X-rays to automatically find the necessary anatomical landmarks and compute the VHS. Once calculated, a Large Language Model (LLM) processes the measurement along with the image’s metadata to produce a structured veterinary summary for the clinician’s review. The entire process from image upload to report generation takes seconds. Quantitative tests confirmed the AI’s measurements are accurate, while qualitative assessment found the generated summaries to be coherent and context-aware.

Precision Diagnosis Informs Targeted Treatment

These studies demonstrate a clear progression in veterinary cardiology: from detecting the problem to defining its exact nature. The AI-assisted VHS tool acts as a sensitive and rapid screening filter. It can flag potential cardiomegaly during routine check-ups or when a dog presents with vague symptoms like a cough or reduced exercise tolerance, prompting further investigation. This is particularly valuable for breeds predisposed to heart conditions, where early detection is essential for managing diseases like mitral valve disease or dilated cardiomyopathy.

If screening or symptoms suggest a deeper issue, advanced imaging like CT provides a definitive, three-dimensional map. As shown in the Afghan Hound’s case, CT can identify complex structural abnormalities—like fistulas, masses, or congenital defects—that are difficult to see on standard X-rays or ultrasound. This precise anatomical information is critical for planning effective treatment, whether it’s targeted surgery, guided medication, or a specific therapeutic procedure. Accurate diagnosis prevents unnecessary treatments and directs resources toward the actual cause, improving the likelihood of a successful outcome like the dog’s full recovery.

Integrating New Tools into Veterinary Practice

For pet owners, these developments translate to more efficient and informed care. The AI VHS tool can be integrated into a general practice workflow, providing a consistent, second-opinion analysis on heart size during annual wellness exams. This objective data helps veterinarians decide if a more detailed cardiac workup is warranted. For dogs showing persistent respiratory or systemic signs that don’t resolve with initial treatment, as in the case study, veterinarians may recommend advanced imaging like CT to rule out rare but serious conditions.

It is important to acknowledge that these are supportive tools, not replacements for clinical judgment. The AI summary requires veterinarian review, and CT scanning requires general anesthesia. However, by automating tedious measurement tasks and revealing hidden anatomy, they empower veterinarians to diagnose with greater confidence and speed. This allows for earlier intervention, which is a fundamental principle for managing chronic conditions and curing acute ones, ultimately supporting longer, healthier lives for canine companions.

Conclusion

From an AI that streamlines heart screening to advanced CT diagnosing a singular case, veterinary medicine is applying new technology to old problems. These tools enhance the diagnostic process, allowing veterinarians to move efficiently from suspicion to precise understanding. For dogs with heart and related thoracic diseases, this means earlier detection and more accurately targeted treatments, improving the standard of care one patient—and one algorithm—at a time.

Sources:
https://pubmed.ncbi.nlm.nih.gov/42007624/
https://pubmed.ncbi.nlm.nih.gov/42002024/