The field of medicine is one of the most positively affected by developments in Artificial Intelligence (AI). AI (Artificial Intelligence) In Medicine.
According to a report by Frost & Sullivan, the use of IA solutions for hospital workflows will significantly improve patient care. Overall, IA has the potential to improve outcomes by 30 to 40 percent while reducing treatment costs by up to 50 percent.
“By 2025, artificial intelligence systems may be involved in everything from population-level health management to digital avatars capable of answering specific patient questions,” says Harpreet Singh Buttar, an analyst at Frost & Sullivan.
There are several excellent examples of IA implementation that make us optimistic about the near future, here are a few.
- SkinVision.
- Biofourmis.
- ContinUse Biometrics.
- BioMedIA
- Artificial Intelligence in Mental Health.
- RadIO
1. SkinVision
Melanoma is a type of skin cancer and one of the most common causes of death, more than 60% of cases are caused by ultraviolet radiation from the sun.
To use this app, the user has to photograph the marks of his skin that he wants to analyze with the camera of the cell phone, then the system studies different parameters of the mole or stain to assign a level of danger: low, medium or high.
“The algorithm currently analyzes seven different criteria and will be further improved by the continuous growth in our database (currently more than 1 million images),” explains Dick Uyttewaal, CEO of SkinVision.
2. Biofourmis
Biofourmis, a startup founded in September 2015 in Singapore, harnesses the power of AI to detect personalized patterns that predict the deterioration of the patient’s health.
They use medical devices such as wearable and biological sensors to capture physiological signals and detect anomalies. This IA-enabled continuous monitoring platform and alerts medical professionals to intervene even days before a critical event.
“At Biofourmis, we are reinventing remote patient monitoring that will improve clinical outcomes and reduce re-admissions after discharge. Our IA technology identifies deteriorating health weeks before readmission to the hospital,” explains founder and CEO Kuldeep Singh Rajput.
They are currently working with the Mayo Clinic to access anonymous medical data from clinical trials and other medical data sources. This collaboration allows them to optimize their “Biovitals” product.
3. ContinUse Biometrics
ContinUse Biometrics develops a remote, contactless patient sensing technology that simultaneously identifies and authenticates users and monitors their physiological status.
It has developed a laser-based optical sensor that combines hardware and automatic image processing technologies to measure the physical health of patients at the molecular level.
The sensor can communicate with common electronic devices (e.g., cell phones, assistants, among others), allowing people to have continuous medical monitoring, wherever they are: at the clinic, at home, at work or while driving.
The sensor allows remote detection of heart and respiratory rates, including auscultation of heart and lung sounds, blood pressure, myography, peripheral hemodynamics and even some biochemical data, all from a distance, without any physical contact with the user.
4. BioMedIA
BioMedIA has developed algorithms for the acquisition, analysis and interpretation of images.
They use automatic learning to extract clinically useful information from medical images, in particular for detection and diagnosis.
They are currently developing and evaluating SonoNet for real-time fetal monitoring using conventional ultrasound. Automatic image processing can provide tools to assist experienced or inexperienced operators in clinical evaluation and thus detect malformations or other alterations.
Technical details on the use of convolutional neural networks in this new technique have been presented in an open-access scientific paper.
5. Artificial Intelligence in Mental Health
Language is one of the main tools with which psychiatrists evaluate their patients for certain mental disorders, such as depression or schizophrenia.
That is why two teams of IBM researchers (from the Computational Psychiatry and Neuroimaging Groups) led by Argentine researcher Guillermo A. Cecchi decided to develop an artificial intelligence capable of predicting the production of a psychotic episode in a patient, even several months in advance.
In a study already published in the journal World Psychiatry, transcripts of interviews with several individuals were used, divided and graded according to the coherence of the sentences.
An automatic learning model then determined, based on those speech patterns, who was at risk of developing psychosis and who was not with an accuracy of more than 80 percent.
“We believe this is an important step toward the goal of developing a tool for mental health professionals, caregivers and patients. A tool that can expand and multiply the scope of neuropsychiatric evaluation outside the clinic,” Dr. Cecchi explained in an interview.
6. Corti
Emergency dispatchers in Copenhagen, Denmark, are using a new voice assistance technology called Corti.
Corti works by listening to all emergency phone calls. The system analyzes the conversation and detects clues that it then links to its database. It then shows dispatchers the related information.
The combination of humans and machines has proven successful, explains the CEO of EMS in Copenhagen.
Corti has saved many people’s lives, because in emergency situations, people are often under a lot of pressure. Both people in trouble and dispatchers can be nervous. And that’s when information about symptoms and medical conditions is vital.
Corti seeks to be an aid to human work, never a replacement, explain its creators.
7. RadIO
Artificial intelligence is one of the most powerful allies we have in the fight against cancer. One of the examples mentioned by the media is RadIO technology.
The IT Department of the Moscow Government launched this new system. It is an open source cancer detection code, which is available at Github, and which uses ‘deep learning’ to find signs of lung cancer on x-rays.
RadIO is freely available to anyone who wants to use it, and there are tutorials on how to run it.
It is not a magic cancer detection button, but it is likely to save many lives. According to DIT, RadIO allows you to create ‘deep learning’ algorithms in a short, easy-to-read piece of Python code.
That system is so fast that, according to the researchers, it is capable of processing the X-rays of the entire population of Moscow (12 million people) in 30 seconds.
