By Dave DeFusco
Dr. Youshan Zhang, an assistant professor of artificial intelligence and computer science at the Katz School of Science and Health, has been awarded a $175,000 grant from the National Science Foundation (NSF) for research aimed at improving cardiomegaly diagnosis in animals by the development of an innovative, AI-driven diagnostic tool.
Cardiomegaly, the enlargement of the heart, is a critical early indicator of heart disease, particularly in dogs, and is one of the leading causes of death in humans and animals. Traditionally, the detection of this condition relies on the manual analysis of thoracic radiographs, also known as a chest X-ray, using the Vertebral Heart Scale (VHS). However, this method can be time-consuming, prone to human error and requires specialized expertise, making it less efficient and accessible for widespread use.
Dr. Zhang’s project, "Cardiac Disease Detection with AI for Veterinary Medicine," tackles these challenges by developing deep learning models capable of automating the VHS process with greater accuracy and speed.
“The primary goal of the project is to bridge the gap between traditional clinical methods and advanced AI models,” said Dr. Zhang. “Many clinicians, especially those without deep learning backgrounds, struggle to trust AI-generated results due to a lack of transparency and explainability in current models.”
To build trust and usability, the project aims to integrate traditional VHS metrics into the deep learning framework. This will allow clinicians to better understand how AI-derived predictions align with established medical standards. The research will focus on improving the transparency and accuracy of these models, making the diagnosis process more intuitive for veterinary professionals and reducing the reliance on manual calculations.
The project is based on Dr. Zhang’s earlier work published in Scientific Reports, in which he introduced the Regressive Vision Transformer (RVT) for cardiomegaly assessment in dogs. Building on this foundation, his new project outlines three primary goals:
Development of New Detection Models: Dr. Zhang will create a new tool called a perpendicular fully connected layer (PFCL), which will make sure that the way the heart is measured in the X-ray image is more precise by ensuring that certain lines used in the measurements are at perfect right angles to each other. This will help the computer model better detect important features of the heart and calculate its size more accurately, making the diagnosis of heart problems more reliable.
Automatic Report Generation: Using deep semantic mapping and few-shot generation techniques, Dr. Zhang will develop tools capable of generating cardiomegaly reports with minimal training data. This will streamline the diagnostic process, particularly for initial evaluations by general practitioners.
User-Friendly Software Interface: A major outcome of the project will be the creation of an accessible software interface for clinicians and the general public. This tool will combine data labeling, result prediction, report generation and modification into one platform, making it easy to use without requiring prior domain knowledge.
“By developing a more precise and accessible diagnostic tool, the project aims to lower the cost of cardiomegaly detection while improving diagnostic accuracy and reducing stress for pet owners,” said Dr. Zhang. “The project’s deep learning models could also pave the way for similar AI applications in human medicine, particularly in improving the early detection of heart disease.”
