A better future? A New Study Claims Machine learning could predict death or heart attack with over 90% accuracy
Artificial Intelligence is the future and with constant efforts of researchers to train machine learning algorithms in the field of healthcare made us capable of predicting death or heart attack with more than 90 percent accuracy.
This is not mere speculations these statistics are the finding of the study which was presented at The International Conference on Nuclear Cardiology and Cardiac CT (ICNC) 2019.
Scientists constantly trained machine learning algorithms feeding them data and By repeatedly analyzing 85 variables in 950 patients with known six-year outcomes, the machine learning algorithm finally ‘learned’ the procedure of interaction of imaging data. The algorithm then identified patterns correlating the variables to death and heart attack with more than 90 percent accuracy which is astounding and well above the human predictions.
This Study was conducted under the authorship of Dr. Luis Eduardo Juarez-Orozco of the Turku PET Centre in Finland. According to Dr.Luis “These advances are far beyond what has been done in the field of medicine, where we(Doctors) need to be cautious about how we evaluate risk and outcomes of patients. Doctors have the data but we as doctors are not using it to its full potential yet.”
Even after years of experience Doctors use risk scores to make treatment decisions. Doctors are doing their best But these scores are based on just a handful of variables and the success and accuracy rate depends on individual patients.
Doctors are specially trained and even with years of expertise, they cannot predict life and death outcomes with 90 percent accuracy. The times are changing and in the field of machine learning Through repetition and adjustment, it can exploit large amounts of data and identify complex patterns that may not be evident to humans.
Dr.Juarez-Orozco explained the foundations of a human mind and the capabilities of machine learning he said “Humans have a very hard time thinking further than three dimensions or four dimensions. The moment humans jump into the fifth dimension we are lost. Our study of machine learning shows that very high dimensional patterns are more useful than single dimensional patterns to predict outcomes in individuals and for that, we need machine learning.”
In their study, Dr.Juarez-Orozco and his team enrolled 950 patients with chest pain who underwent the center’s usual protocol to look for coronary artery disease.
A coronary computed tomography angiography (CCTA) scan produced 58 pieces of data on the presence of coronary plaque, vessel narrowing, and calcification. The patients who had these scans suggestive of disease underwent a positron emission tomography (PET) scan which further produced 17 variables on blood flow. They obtained Ten clinical variables from this study which were gathered from medical records including sex, age of patients and their habits like smoking, and their health conditions like diabetes.
They used the data from this study to help make a better prediction for the future. Doctors saw that During an average six-year follow-up there were 24 heart attacks and 49 deaths from any cause. The more the data the more refined the algorithm can perform so, 85 variables were then entered into a machine learning algorithm called ‘LogitBoost’, which analyzed them over and over again until it found the best structure to predict who had a heart attack or death.
Dr. Juarez-Orozco was excited and motivated about this study, he said that “The machine learning algorithm progressively learns from the data and after numerous rounds of analyses, the algorithm figures out the high dimensional patterns that should be used to help us in efficiently identifying patients who have the event. The result is a score of individual risk for individual patients”.
Dr. Juarez-Orozco further added that “Doctors already collects a lot of information about patients. Our team of researchers found that machine learning can integrate these data and accurately predict individual risk to individual patients. This should allow doctors to recommend personalized treatment and ultimately lead to better outcomes for patients”.