Researchers from Brigham and Japan’s Keio University developed an artificial intelligence and deep learning model to scan an electrocardiogram (ECG) for signs of atrial septal defects (ASD) of the heart. This condition can cause heart failure and is not reported due to a lack of symptoms before irreversible complications develop.
“If we can deploy our model of ECG screening at the population level, we will be able to catch more of these patients before they become irreversibly damaged,” said lead author of the new study, whose findings were published in the journal eClinicalMedicine, Dr. ». This is according to what was mentioned by the specialized medical website, “Medical Express”.
And “atrial septal defects” or (ASD) is a common congenital heart disease in adults; It is caused by a hole in the septum of the heart that allows blood to flow between the right and left atria.
Goto added that the disease has been diagnosed in about 0.1% to 0.2% of the population, but it is likely that it is not reported. Symptoms of ASD are usually very mild or in many cases not present until later in life.
Symptoms include:
Inability to do strenuous exercise
Affecting the heart rate or rhythm
Heart palpitations and increased risk of pneumonia.
Even if ASD does not cause symptoms, it can put pressure on the heart and increase the risk of atrial fibrillation, stroke, heart failure, and pulmonary hypertension. At this stage, complications of ASD are irreversible even if the defect is repaired later. If caught early, ASD can be corrected with minimally invasive surgery to improve life expectancy and reduce complications.
There are several ways to detect ASDP; First, the largest defects can be found by listening to the heart with a stethoscope. But only about 30% of patients can be detected this way. the other by echocardiogram; It is a time intensive test and not a good screening option. Another test, the ECG, takes only about one minute, which makes it possible to use it as a screening tool.
However, when humans analyze ECG readings for known abnormalities associated with autism, there is limited sensitivity to picking up an autism spectrum disorder.
To see if the AI model could better detect ASD from ECG readings, the study team fed the ECG deep learning model data from 80,947 patients over the age of 18 who underwent both an electrocardiogram and an echocardiogram to detect ASD.
A total of 857 patients have also been diagnosed. Data was collected from three hospitals; two large educational institutions, one BWH in the United States and the other Keio University in Japan, Dokyo Medical University and Saitama Medical Center in Japan; It is a community hospital.
The model was then tested using scans from Doku; which has a larger population and does not specifically screen ASD patients. The model was more sensitive than using known ECG abnormalities to screen for ASD. The model correctly detected ASD 93.7% of the time, while using known abnormalities it found ASD 80.6% of the time.
“I’ve picked up a lot more than what an expert does using known abnormalities to identify cases of ASD,” Goto said.
A limitation of the study is that the model was trained on user samples from academic institutions; which deals more with rare diseases such as ASD. All patients who used to train the model were also screened for ASD and received echocardiograms; So it is not clear how well the model works on the general population. That’s why they tested it at Dokyo; The performance of the model was maintained even in the general population of the community hospital, indicating that the model is well generalizable.
Notably, the authors also note that even using an echocardiogram to detect ASD will not find every defect; Some can bypass both the regular scan and the AI model, although these smaller defects are less likely to require surgery.
Goto concludes, “The problem with machine learning is that it is a black box; We don’t really know what features were captured. This means that we can’t tell from the model what features to look for in the ECG either. While the results indicate that this technique can be used in population-level screening to detect ASD before it leads to irreversible heart damage.”
Ultimately, ECGs are relatively inexpensive and are now performed in many contexts, says Goto, who notes, “Perhaps this test could be incorporated into an annual appointment for PCP or used to screen for EKGs taken for other reasons.”
2023-08-17 12:15:22
#vegetarian #diet #reduce #hot #flashes #menopause
