Detecting and preventing diseases with AI.
One field exposed to innovation every day is healthcare as there are so many opportunities for artificial intelligence to help medical experts and researchers provide better attention and treatment for people around the world.
Among all the possibilities for computer-aided systems, some of the most impressive medical developments assist doctors in the detection and prevention of diseases.
CADe
Within this field, there are different solutions that optimize these kinds of clinical procedures, for example: Computer-aided detection (CADe). This tool consists of computer-output that helps doctors make a diagnosis with greater accuracy and faster results.. A CADe system would help by detecting pathologies (the behavior of a disease) among medical images, having that information classified and ready for a doctor.
A highly-accurate computer could point out all the inconsistencies in multiple images. Furthermore, the system could be enhanced with machine learning over time as it would have to go through many algorithms and heavy datasets and would also receive inputs from medical professionals in order to guarantee accuracy (VITech, 2022).
This type of technology isn’t designed to replace a clinician, but instead to relieve them from time-consuming tasks by using dependable AI that would allow this professional to have accurate and straight-forward data, letting them invest more of their time on making diagnoses and working towards treatments for more patients.
A CADe that has been developed recently is an AI software that aids doctors in the detection of lung cancer. Mozziyar Etemadi, a biomedical engineer at Northwestern University’s Feinberg School of Medicine, created a program that looks at patient’s computerized topographies (CTs) and is able to detect distinct patterns and irregularities that could indicate lung cancer in patients.
According to the Nature Journal’s article on Etemadi and his AI system: “In 2019, he and his team reported that their system correctly identified the early stages of lung cancer 94% of the time, outperforming a panel of 6 veteran radiologists .
By training the computer with over 40,000 CT scans, it was able to learn what factors signal cancer. Researchers would tell the computer which “early-stage scans” would turn out to be lung cancer diagnoses later on. Therefore, the AI program was able to find subtle patterns of this disease that doctors wouldn’t recognize at such early phases and would actually develop into cancer. These miniscule details would be 1–3 millimeters in size, making it difficult for doctors to identify but a computer was able to pick up on them.
CAP
Not steering too far from CADes, another form of computer-aided systems are CAPs, or computer-aided prognoses. CAP, being a sub-field of CAD, goes even further in medical diagnosis accuracy by performing image detection but also integrating an analysis of the patient’s data including medical history, a profile of the patient, among others. This takes the capabilities of computer-aided systems in medicine further because the computer has more data to work with, resulting in more accurate and personalized data (VITech, 2022).
Different forms of CAPs are already being used in the healthcare world. An AI software developed by researchers at Houston Methodist Research Institute in Texas is able to scan “patient charts, collected diagnostic features and correlated mammogram findings with breast cancer subtype” in order to accurately diagnose women with breast cancer. In this case, this software is combining many different inputs, generating an output that’s a result of many different factors pertaining to one patient.
This program has been immensely useful to doctors because it takes exponentially less time to interpret mammograms and review a patient’s data. While it would take two professionals 50 to 70 hours to review the information of 50 patients, this AI system would scan 500 of them in a matter of hours.
Monitoring Health
One form of this technology that many people are familiar with is health monitorization, usually done through a device such as a smart watch. These gadgets that millions of individuals have integrated in their daily lives can use AI in certain functions linked to the user’s health and wellbeing. These watches can track your heart rate or even stress levels and alert you if there are any irregularities that put your health at risk.
In fact, there are some developments in technology that would allow a smart watch, with help of AI, detect low blood sugar based on irregular heart behaviors that the watch would pick up. This would be a great tool for somebody with diabetes as it would be able to track low blood sugar in a painless and discrete way. Of course, the irregularities in a patient’s heart that would indicate said issue would vary depending on the patient but, through deep-learning, the AI would recognize what an inconsistency would look like in each individual and have its alerts be personalized.
Reducing Medical Errors
Finally, this use of AI is related to the jobs professionals in healthcare have to do on a day-to-day basis whether it is in a hospital, a clinic, or any health institution. An issue in this field has been the use and management of electronic health records (EHRs) and electronic medical records (EMRs). EHR and EMR systems are currently not very efficient as they are considered “typically complex, cumbersome, expensive to maintain and customize, inflexible, and not very user-friendly”.
The use of AI could help tackle all of these massive amounts of data and help healthcare workers navigate through their patient’s information quickly and efficiently. This is highly relevant because a doctor would rather spend their time helping their patient instead of trying to find what they need among heaps of information.
With and AI model with natural language processing (NLP) capacities, a doctor could take clinical notes and a computer could interpret this data and categorize it in a way that could be easily managed or found later on. In fact, the AI model could receive inputs such as voice-recordings, video or images and process them into the system. This would save time in tasks such as reporting and result notifications.
What’s next?
Even with all the remarkable investigations and developments with AI in healthcare, there is still so much space for new ideas and inventions to come along and improve the system for both healthcare workers and patients. By facilitating administrative tasks, detecting medical issues early on and helping researchers discover new solutions, AI could contribute in making these processes more efficient.
All of these discoveries and all the others out there should inspire the next generation of biomedical engineers, data scientists, researchers, software developers or anyone who is interested in this area to not limit themselves and work towards making quality health accessible every human being.
Sources
Deng, F. (2019,). Computer aided diagnosis: Radiology reference article. Radiopaedia Blog RSS. Retrieved October 5, 2022, from https://radiopaedia.org/articles/computer-aided-diagnosis-1
Griffiths, S. (2016). This AI software can tell if you’re at risk from cancer before symptoms appear. WIRED UK. Retrieved October 5, 2022, from https://www.wired.co.uk/article/cancer-risk-ai-mammograms
Grimes , K. (2020). How smart watches and AI can help people with diabetes. Babylon Health. Retrieved October 5, 2022, from https://www.babylonhealth.com/en-gb/blog/tech/how-smart-watches-and-ai-can-help-people-with-diabetes
Mehta, N. (2022,). 10 application of artificial intelligence in Modern Healthcare — Techtic Solutions, Inc.. Techtic Solutions. Retrieved October 5, 2022, from https://www.techtic.com/blog/applications-of-ai-in-healthcare/
SEDGE, The Solutions Edge. (2021). How deploying AI in hospital systems helps reduce medical errors. Medium. Retrieved October 5, 2022, from https://medium.com/@solutionsedge/how-deploying-ai-in-hospital-systems-helps-reduce-medical-errors-c6f20eab3a90
Svoboda, E. (2020). Artificial intelligence is improving the detection of lung cancer. Nature News. Retrieved October 5, 2022, from https://www.nature.com/articles/d41586-020-03157-9
VITech. (2022). Why we use computer aided systems in healthcare. VITech. Retrieved October 5, 2022, from https://vitechteam.com/computer-aided-systems-in-healthcare/
Zewe, A. (2022). A smarter way to develop new drugs. MIT News | Massachusetts Institute of Technology. Retrieved October 5, 2022, from https://news.mit.edu/2022/ai-molecules-new-drugs-0426
Our other news about AI here.