A group of Grade 11 students from Quezon City Science High School has developed a cutting-edge artificial intelligence (AI) tool that could transform how heart disease is diagnosed. Their software, called Pintig, is designed to detect plaque buildup in the heart from CT scan images with remarkable accuracy and speed.
Tackling the Heart Disease Detection Problem
Heart disease, caused in part by atherosclerosis (plaque buildup in the arteries), remains one of the world’s leading causes of death. Traditionally, detecting plaque buildup can be difficult, often only spotted at an advanced stage when it’s harder to treat. This is where Pintig steps in. The AI system can analyze CT scans at a microscopic level, detecting plaques as small as 50 micrometers—far beyond what the human eye can see.
According to one of the developers, Kenzo Miguel Tayko, the main challenge today is that plaque is usually detected too late, requiring costly and invasive treatments. Pintig aims to change that by offering a non-invasive, fast, and more accurate way of diagnosing plaque buildup in its early stages.
How It Works
The software works by analyzing thousands of CT scan images, with a dataset of 14,000 images from over 200 patients. Using machine learning, Pintig examines the scan in seconds, highlighting any potential plaque and classifying it as positive or negative for coronary artery disease (CAD). The results are then displayed instantly, making it easier for doctors to make informed decisions.
With an accuracy rate of up to 98%, Pintig offers a promising tool for early heart disease detection, potentially reducing the need for expensive surgeries and allowing for earlier, more effective treatment.
Next Steps: Clinical Trials and Future Potential
Having already won top honors at a regional science fair, the team is now preparing Pintig for clinical trials in partnership with the Philippine Heart Center. With further testing and development, Pintig could soon be a key tool in preventing heart disease before it becomes life-threatening.
