One of the biggest research areas in the medical industry involves understanding mortality and identifying methods to expand the human lifespan. Researchers and health experts have used several markers over the years. Body Mass Index remained a common choice, but it had too many limitations that affected the accuracy of the predictions made. As DEXA scans continue to gain traction, researchers are turning to AI to predict all-cause mortality based on body composition instead of simply relying on BMI.
In this article, we explore recent findings from studies that focused on building neural networks and using deep learning to predict mortality through this DEXA imaging.
Artificial Intelligence, Neural Networks, and Deep Learning in Medical Research
Mortality studies and research papers generally involve multiple databases, which can sometimes include thousands of documents, overviews, reports, and other content. For a long time, medical research focused on a manual approach. Researchers would review all the data they could obtain, manually count up the results, and eventually provide a report based on their research findings.
Unfortunately, this takes a long time to complete and creates the potential for human error.
With the recent advancements in artificial intelligence, we’re seeing more of this technology introduced in the medical industry. This includes medical research, as AI has the potential to process data significantly faster than human researchers and offers a greater level of accuracy.
A research paper published in the Future Healthcare Journal explains the impact that AI could have in the healthcare sector. The paper focuses on different areas of healthcare and medicine and covers the fact that there are already several instances where different types of AI are used in clinical settings.
Currently, machine learning is an important AI technology for this industry. It is the type of AI that allows us to create neural networks and implement deep learning systems.
While machine learning is the focus of much attention, other types of AI have also entered medicine and healthcare. Natural language processing is a good example and works closely with the deep learning models we see in clinical settings. AI is also used in other areas, such as to create automatic processes for surgery where physical robots are used.
How AI Can Impact Mortality Studies Using DEXA Scans
When considering the use of AI in medical research, several areas of interest arise. Many scientists have also started to turn to this type of tech to help with their mortality studies.
One exciting finding came from scientists who considered AI’s ability, with neural networks and deep learning, to predict all-cause mortality using data from DEXA scans.
DEXA scans offer a more detailed overview of the body’s composition since it provides imagery that can be used to determine lean, fat, and bone mass. The study was published in the Journal of Communications Medicine and presented data after the researchers developed multiple neural networks. All these networks utilized deep learning models.
The researchers had access to more than 15,000 scans from more than 3,000 patients. Both male and female patients were included in the study, which was part of a Health, Aging, and Body Composition Study. Data was collected for a period of 16 years.
The AI models were able to accurately predict all-cause mortality based on the body composition data provided. The performance of the deep learning system was compared to traditional methods used to predict mortality. It performed better, making it a viable choice for mortality studies in the future.
The study also shows the potential that DEXA scans offer. By simply considering the data that comes with a full-body DEXA scan, the AI system was able to use the information to make accurate predictions.
This also shows that AI should be explored further in mortality studies. By implementing neural networks, researchers can process vast amounts of data in a short time.
Conclusion
Artificial intelligence is starting to play an essential role in the medical industry, primarily because it can process large amounts of data in a short period. Studies have now shown that deep learning AI holds the potential to provide a more accurate prediction of all-cause mortality when it gains access to full-body DEXA scans. While more research is still needed, the evidence that these studies have provided offers a good foundation on which to focus to determine future directions.
References
https://www.ibm.com/topics/artificial-intelligence-medicine