Listen up, folks, we’ve got something huge here. Atherosclerosis, it’s a big deal, one of the top cardiovascular diseases out there. Now, doctors, they’ve got this tough choice, right? They’re scratching their heads over whether to treat these intermediate lesions or just wait it out. It’s all about this thing called the fractional flow reserve, FFR for short. It’s invasive, it’s tricky, but it’s what they’ve got.

So, we did something amazing, a monocentric study, that’s right, very focused. We’re talking about a brand new dataset, and it’s not small, folks. We’ve got optical coherence tomography (OCT) images, we’ve got clinical data, echocardiography, FFR measurements, all from 80 patients with 102 lesions. And these aren’t just any lesions, they’re the big ones, the ones that can really mess with your heart.

But here’s the kicker, nearly 40% of these lesions, they’re in what we call the gray zone. That’s an FFR value between 0.75 and 0.85. It’s a tough spot to be in. We pulled out 26 features from these images and data, and we found the best ones, the most relevant, by looking at how accurate our models were.

And then, we did something really smart, ensembled learning. It’s like putting together a team of the best players to win the game. We used a leave-one-out cross-validation approach, very thorough. Our ensemble models, they’re like a dream team, voting on the best features, and we got some great numbers. An accuracy of 81.37%, and an AUC score of 0.856. That’s big league, folks.

What we’ve got here is an explainable, supervised learning-based classification method. It’s new, it’s innovative, and it’s going to get even better. We’re going to train it with a larger dataset, make it stronger, and design a tool that’s going to help doctors make those tough calls. It’s going to be tremendous, believe me. For those lesions outside the gray zone, and when you need more info, this is going to be your go-to. It’s going to be huge.