04.10-11.24-HST-Startups-General-Prognostics

SEAN MATSUOKA: Hi, everybody. My name is Sean Matsuoka, one of the co-founders of GPx. My co-founder and CEO, Javier Echenique, is a proud MIT graduate.

So at GPx, we are developing bloodless blood tests. We believe that we cracked the nut to something that people have tried for decades, which is to convert a simple smartwatch into a medical device. We're starting off with heart failure, but we have bigger aspirations to expand into other therapeutic areas as well as global scale. Excuse me.

So today, nobody-- no doctor will use a smartwatch to monitor their heart failure patients. This is actual data you see from our patients across the span of five months of heart rate. It's just too noisy.

Nobody can interpret it. Nobody wants more data to mine. But when you pass this data through our algorithm, you can tell exactly when the exacerbation started. It also correlates with vital blood biomarker. So doctors would know exactly what to do.

In the world of heart failure, patients need more proactive heart failure monitoring solutions. The only FDA-approved clinically validated solution is an implantable device. It requires surgery. It's also indicated for a very small segment of the patient population. So 90%-- over 90% of the patient lack proactive monitoring solutions, and racking up huge, huge hospital bills. So we're providing a solution to keep patients out of the hospital.

An interesting approach to monitoring heart failure patients is also at the core of a company is doing serial blood testing of NT-proBNP. NT-proBNP, it's a terrible name. It's a hormone released by the heart when the heart experiences stretch. It's recently been risen to a class I recommendation in 2017 for diagnostic as well as prognostication of heart failure patients.

Several clinical studies have shown the efficacy of using this vital blood biomarker in keeping patients out of the hospital, including a landmark study by Roche called STRONG-HF. However, at the standard of care today, it's only done twice a year. It's costly, it requires a patient to go to the clinic to get their blood drawn. And when we ask a doctor, How often would you like access to the data? They want it on a weekly basis. There's a huge unmet need, and we're not trying to fulfill that gap by doing more blood tests. We're trying to do that algorithmically.

Now since the inception of the company, we've enrolled over 250 patients. And we collected a very unique pairing of digital and blood biomarker for about 6 to 12 months. And from that, we are able to derive a very unique predictive algorithm. This database is very unique and it's bigger than that of Roche or Mayo Clinic.

Let me just show you a case illustration and then talk about the performance of the algorithm. So in our current observational clinical study, we're doing blood tests every other week. This is to train the algorithm as well as to validate. Once we have an FDA approval, this will be a completely digital system.

As you can see, between the third and fourth, there was a steep rise in NT-proBNP, meaning that the patient is experiencing exacerbation. And then two weeks later, the patient was readmitted to the hospital. Our algorithm runs concurrent to this and predicted three weeks ahead of the NT-proBNP rise. And also, you can notice that the patient was even asymptomatic five weeks ahead of the hospitalization. So once we get an FDA approval, the doctors can use this to keep patients out of the hospital.

Now in our latest observational study of 235 patients across 10 hospitals in the US and India, we have shown a very promising result in the algorithm. We are performing at 87% sensitivity and 99.5% specificity. The takeaway message here is that, one, it exceeds the threshold set forth by the FDA for clearance. And then, two, it's better than implantable device that requires $30,000 and requires the patient to go through a surgery.

So our ask today is we're looking for partners. We're looking for partners who are looking to partner with startups in the chronic disease state. Some of the use cases are the following. We can help you in your clinical study. We can help identify exacerbation ahead so you can reduce the risk of adverse events. We can be used for companion diagnostic. We can be used to update the patient on GDMT so that-- so that they can-- their outcome will be better.

Thirdly, the patient candidate identification. We can help you bring patients on second-line or third-line therapy. And then lastly, remote patient monitoring. So we can help you keep patients away from the hospital. We also have a booth next door. So if you're interested, please come by and visit us. Thank you so much.

[APPLAUSE]