Abstract: Cardiovascular disease is one of the primary global health issues since it leads to the death of millions of people each year worldwide. To advance the treatment outcomes and alleviate the resulting health care pressure, an early diagnosis plays a vital role. We review in this paper whether the VGG-16 model, specifically one of CNN architectures, may be used to detect CVD automatically with the help of the analysis of medical images. VGG-16 exploits a deep, sequential arrangement by employing small 3x3 convolutional filters to obtain a fully connected configuration capturing fine spatial detail in echocardiogram images, MRI images, and CT scan images for identifying patterns in cardiovascular illnesses that traditional methods cannot match. This investigation further points to the dataset pre-processing technique that has the capacity of enhancing generalization, with regard to model explanations for its prediction and crucially significant for its adoption in a clinical setting. These results ultimately prove that the VGG-16 model is a potentially sound early CVD detector tool and a promising addition to diagnostic practices, especially in contexts with limited access to healthcare professional expertise. The current review contributes to the growing body of literature on the role of deep learning in medical imaging and advocates for the incorporation of AI technologies into routine clinical workflows for enhanced patient care.

Keywords: Cardiovascular Disease, ECG, VGG-16, Convolutional Neural Network, AI-driven diagnostics

| DOI: 10.17148/IARJSET.2025.12647