Back to search

BIA-Brukerstyrt innovasjonsarena

Echocardiography Automated diagnostic Interpretation system

Alternative title: Ekkokardiografi Automatisert Diagnostisk Interpreterings system

Awarded: NOK 9.3 mill.

Project Number:

313756

Project Period:

2020 - 2025

Funding received from:

The Echo AI project addresses the need to improve efficiency and confidence in diagnosing Cardiovascular Disease using medical ultrasound imaging. This will be accomplished by using artificial intelligence (AI) to help cardiologists analyze ultrasound images of the heart. Cardiovascular disease is the number one cause of death globally, resulting in grave consequences for patients and their families. Medical ultrasound scanners are commonly used by cardiologists to diagnose patients with cardiovascular disease because they are relatively inexpensive ,portable, and they are easier and safer to use than for instance CT scanners. Cardiologists analyze ultrasound images of the heart using many different techniques both based on visual inspection and measurements as part of clinical guidelines. It is very common to measure distances between structures as well as the size of different structures in the heart. These measurements are typically performed by using manual drawing tools. One example is that the cardiologist use a manual line drawing tool to measure the distance between two structures in the image. One of the main goals of the Echo AI project is to train AI methods to perform many of these measurements automatically. This will yield more consistent results and it will help the cardiologist to become more efficient. In the initial phase of the project we have developed an algorithm for automated region-of-interest determination in connection to automated analysis of wall motion. The result is that this analysis can be done automatically Individual measurements are usually not enough to give a diagnosis. Often several measurements need to be analyzed together and this can be complicated. Another goal of Echo AI is to perform this kind of analysis automatically based on established clinical formulas and guidelines. We have focused on one specific analysis in the initial phase of the project. AI methods can be very good and efficient but they are never perfect. In the Echo AI project we will develop an application that can present the measurement and analysis results in an intuitive way. This will help the cardiologists to easily verify that they agree with the results. We have a first prototype of such an application in the early phase of the project.

Cardiovascular disease (CVD) is the number one cause of death globally, representing almost one third of all deaths (World Health Organization 2019). There are devastating consequences for both patients and their families, as well as overwhelming healthcare expenditures. CVD costs the Norwegian economy more than 120 NOK billion per year. Long-term management of chronic CVD is an especially high fraction of these costs. Echocardiography (echo) is the primary imaging tool for cardiologists to assess cardiac function and structures. It is essential in evaluating and monitoring chronic and acute cardiac disease as it provides a low cost, mobile and real-time view into the heart. Echo can be used to visualize and quantify overall structure and function of the heart, cardiac muscle, valves, pathologies and blood flow. The vision for this project is to create an innovative new product offering: ECHOcardiography diAgnostic Interpration system(ECHO AI) Echo AI will utilize artificial intelligence to provide workflow improvements by seamlessly integrating image acquisition, analysis and reporting. It will maximize consistency, accuracy, and completeness in accordance with established clinical guidelines. This will be done by developing fully automatic measurements and interpretation of cardiac images originating from an ultrasound scanner. The automated analysis will, in combination with established clinical guidelines, lead to a set of clinical findings. The new product offering will allow for a new and more efficient workflow, where the physician starts with a pre-populated report and can drill down from the high-level findings to individual measurements and images that the findings are based on.

Funding scheme:

BIA-Brukerstyrt innovasjonsarena