Periodic Reporting for period 2 - AI-driven cardiac ultrasound analysis (AI-driven cardiac ultrasound analysis)
Berichtszeitraum: 2023-08-01 bis 2024-04-30
Cardiovascular diseases (CVD) are the leading cause of death globally claiming 18.6M lives a year (32% of global deaths). There are over 6M new cases of CVD in the EU annually and the estimated cost of CVD to the EU economy is €210Bn a year.
Heart ultrasound - the two-dimensional transthoracic echocardiography (2D TTE) – is the most versatile, most widely used, and cost-effective heart imaging method. Accessibility to ultrasound imaging is growing rapidly as the devices are getting cheaper and smaller. Access to quality heart ultrasound examination and interpretation of the acquired images remains full of obstacles: examination requires a high level of expertise, is user dependent, time consuming, prone to error and variation.
Studies showed that inaccuracies can be found in up to 30% of transthoracic echocardiography reports. Moreover, an experienced cardiologist spends 50-85% of the average total test time (30-90 min) on performing only manual measurements. This results in the average waiting time of 4-5 weeks across most EU countries. Considering the global healthcare staff shortage, which will reach over 12.9M workers in 2035 and ageing population, it is clear that there is a need to introduce a faster and less human-reliant solution.
Ligence is remodelling the quality, difficulty, and length of echocardiography with an AI-driven tool to automate the whole analysis of heart ultrasound images and standardise echocardiography reporting. 2TTE has a highly standardised protocol – a list of specific measurements (approximately 50 qualitative and quantitative heart parameters ) that need to be evaluated for each patient and it is curated by cardiologist associations, e.g. the European Society of Cardiology (ESC) or American Heart Association (AHA).
Ligence Heart receives heart ultrasound images and performs a sequence of steps which are normally performed by the user during a regular heart ultrasound image analysis – performs measurements, classify heart image views, detect heart cycle phases, produced standartised report. The report with analysed images is accessible moments after the images are uploaded to the server through any connected device in the hospital's network. The performed measurements and final report are then reviewed and approved by the user. The report includes most important measurements, summarizes them in a fluent synopsis, and attaches the images of the performed measurements to allow the users to easily follow up with patients and track changes over time.
The achieved performance is non-inferior to a cardiologist and increases reproducibility of the reporting ensuring higher quality monitoring. Standardized reporting will greatly increase the readability of the reports and contain all the information that otherwise would be left out due to time constraints.
It seamlessly integrates with existing infrastructure in hospitals, meaning that moments after images are loaded onto the hospital's network the results are accessible on any workstation.
This results in dramatically increased accessibility and analysis quality, earlier diagnosis, and better patient risk stratification, monitoring, and patient management.
Heart ultrasound - the two-dimensional transthoracic echocardiography (2D TTE) – is the most versatile, most widely used, and cost-effective heart imaging method. Accessibility to ultrasound imaging is growing rapidly as the devices are getting cheaper and smaller. Access to quality heart ultrasound examination and interpretation of the acquired images remains full of obstacles: examination requires a high level of expertise, is user dependent, time consuming, prone to error and variation.
Studies showed that inaccuracies can be found in up to 30% of transthoracic echocardiography reports. Moreover, an experienced cardiologist spends 50-85% of the average total test time (30-90 min) on performing only manual measurements. This results in the average waiting time of 4-5 weeks across most EU countries. Considering the global healthcare staff shortage, which will reach over 12.9M workers in 2035 and ageing population, it is clear that there is a need to introduce a faster and less human-reliant solution.
Ligence is remodelling the quality, difficulty, and length of echocardiography with an AI-driven tool to automate the whole analysis of heart ultrasound images and standardise echocardiography reporting. 2TTE has a highly standardised protocol – a list of specific measurements (approximately 50 qualitative and quantitative heart parameters ) that need to be evaluated for each patient and it is curated by cardiologist associations, e.g. the European Society of Cardiology (ESC) or American Heart Association (AHA).
Ligence Heart receives heart ultrasound images and performs a sequence of steps which are normally performed by the user during a regular heart ultrasound image analysis – performs measurements, classify heart image views, detect heart cycle phases, produced standartised report. The report with analysed images is accessible moments after the images are uploaded to the server through any connected device in the hospital's network. The performed measurements and final report are then reviewed and approved by the user. The report includes most important measurements, summarizes them in a fluent synopsis, and attaches the images of the performed measurements to allow the users to easily follow up with patients and track changes over time.
The achieved performance is non-inferior to a cardiologist and increases reproducibility of the reporting ensuring higher quality monitoring. Standardized reporting will greatly increase the readability of the reports and contain all the information that otherwise would be left out due to time constraints.
It seamlessly integrates with existing infrastructure in hospitals, meaning that moments after images are loaded onto the hospital's network the results are accessible on any workstation.
This results in dramatically increased accessibility and analysis quality, earlier diagnosis, and better patient risk stratification, monitoring, and patient management.
Ligence Heart web app software operation optimised and new functionalities developed, AI functionalities enhanced.
Software optimisation improvements allowed company to distribute software updates easier, to get faster feedback to developers, to reduce time from detection to bug fixing, to increase interoperability with other software. Optimisation also increased and eased access to Ligence Heart product testing, and flexibility of the system with more functionalities meeting users’ needs, responding to the users’ feedback received. System was improved to make the usage of the application as intuitive, easy, and quick as possible without any loss in the quality of the exam.
AI functionalities enhancement included the automated strain, Spectral and Tissue Doppler, automated B mode and M mode measurement optimisation.
Software optimisation improvements allowed company to distribute software updates easier, to get faster feedback to developers, to reduce time from detection to bug fixing, to increase interoperability with other software. Optimisation also increased and eased access to Ligence Heart product testing, and flexibility of the system with more functionalities meeting users’ needs, responding to the users’ feedback received. System was improved to make the usage of the application as intuitive, easy, and quick as possible without any loss in the quality of the exam.
AI functionalities enhancement included the automated strain, Spectral and Tissue Doppler, automated B mode and M mode measurement optimisation.
Project activities implemented allowed to reach following results and impacts:
1. United States Patent Application “Echocardiogram classified with machine learning” No 18/081886 detailing automated analysis of colour flow Doppler modality images submitted and accepted by the United States Patent Office, and that grant of a patent can therefore be expected.
2. Improved access to markets. Commercialisation perspectives strengthening cooperation initiatives started in US and UK markets with the leading hospitals.
1. United States Patent Application “Echocardiogram classified with machine learning” No 18/081886 detailing automated analysis of colour flow Doppler modality images submitted and accepted by the United States Patent Office, and that grant of a patent can therefore be expected.
2. Improved access to markets. Commercialisation perspectives strengthening cooperation initiatives started in US and UK markets with the leading hospitals.