LUMI has granted us Extreme Scale access to their supercomputer for one year. And we will begin our research this autumn. LUMI is one of the fastest supercomputers in Europe and the 9th fastest in the world.
What we are researching:
We are trying to find prediction methods for sudden cardiac death (SCD), one of the leading causes of mortality, especially in cardiomyopathy patients. Current SCD prediction methods lack accuracy and predicting SCD remains an unmet clinical challenge. If SCD could be accurately predicted, it would be preventable with an implantable defibrillator.
Traditional SCD risk markers, such as left ventricular ejection fraction have low diagnostic value. Raw electrocardiogram (ECG) data may contain predictive features not easily identified using traditional methods, requiring advanced computational approaches. We are using LUMI’s extreme-scale high-performance computing (HPC) capabilities to develop a deep learning model capable of identifying SCD risk from non-invasive biosignals (signals that can be measured externally f.i. by using an ECG. This research will facilitate earlier detection and intervention, potentially reducing mortality from cardiac events.
Unlike typical methods, we thoroughly test different settings for turning ECG signals into detailed frequency-based images. This helps our deep learning model find important spectral biomarkers, or patterns, that indicate SCD, which conventional techniques fail to detect. To do this, we use a method called short-time Fourier transform (STFT) to turn the ECG signals into visual representations that show how the signal’s energy is spread across different frequencies over time. The research uses the publicly available MUSIC database to develop and validate the predictive models.
This research is part of the SMASH-HCM project and will be used for our digital twin-based decision support solution for hypertrophic cardiomyopathy (HCM) management to improve prognosis and treatment of HCM and related conditions.
More about LUMI’s Extreme Scale call.
To learn more about HCM and Digital Twins take a look at our Educational Series.