Artificial Intelligence (AI), through machine learning technique, uses huge amounts of data to find disease patterns and provide more accurate diagnoses without human intervention.
However, reaching clinical-level accuracy demands analyzing a great volume of information from several institutions. It may come up against issues such as data privacy. Currently, this is one of the major obstacles to the use of AI in medicine. It grows stronger with the General Data Protection Regulation (GDPR).
A new tool called Federated Learning may solve this problem by allowing training with real data, not exposing the information owner.
Learn about federated learning: what it is and how it can boost the use of artificial intelligence in medicine.
Federated learning
In this model, algorithm training occurs in multiple devices or decentralized systems using local data, without sharing it. Once trained, the improved algorithm can be sent to a central server and shared or improved by users from other centers.
Undoubtedly, federated learning is one of the key approaches to ensuring data privacy. This is because it allows the running algorithms collaboratively, without transferring private data to a cloud server.
Moreover, the characteristics of patients, devices used or clinical specializations may influence information from a single medical institution. Providing assertive diagnostics demands a diversified dataset as well, with data regarding different genders, ages, characteristics, and environmental exposures. A large database is essential for that.
Federated learning in brain tumor detection
The Perelman School of Medicine at the University of Pennsylvania (Penn Medicine) in the United States, in partnership with Intel, is leading 29 international health and research institutions in training AI models to identify brain tumors through federated learning.
The study began with a model previously trained with more than 2.6 thousand images from BraTS (Brazilian health technology assessment bulletin), of 660 patients, obtained by MRI. Ten hospitals also collaborated by training an AI model with data from their own patients. After that, federated learning gathered these ten algorithms into a consensus.
The researchers compared federated learning with algorithms trained by only one institution to other collaborative learning approaches. The effectiveness of each method was compared to tests manually analyzed by neurologists.
When compared to a trained centralized-data model that does not protect a patient’s privacy, federated learning had a 99% identical result. Results also indicate that access to more data, from several institutions that respect patient privacy, can benefit model’s performance.
Conclusion
Federated learning assures the preservation of privacy, so it may boost the use of Artificial Intelligence in medicine. In this sense, several researches and companies already offer solutions to health institutions, such as Acuratio.
Reviewed by Paulo Schor, ophthalmologist, free professor and director of innovation of the Federal University of São Paulo (Unifesp) and collaborator of the Faculty of Medicine of the Albert Einstein Hospital.
Follow Phelcom’s blog and stay on top of the main news about artificial intelligence in medicine.
Since the beginning of the pandemic, several researches have investigated how the new coronavirus (SARS-CoV-2) can affect the eyes. A recent study has described two cases of patients who have suffered corneal transplant rejection after being immunized against covid-19. The work was published in the British Journal of Ophthalmology.
Learn about the study and results. In fact, it is still too early to correlate rejection to the vaccine, but it works as a warning signal to doctors.
The research
The work reported two cases of corneal endothelial transplant rejection after patients received the covid-19 vaccine. Those procedures took place before immunization.
The first evaluated case is of a 66-year-old woman who took the first vaccine dose 14 days after transplantation (unilateral DMEK combined with phacoemulsification and IOL implant).
The patient had an acute onset of blurred vision, redness and photophobia seven days after vaccination. Doctors increased the drug dosage to an hourly basis to treat acute rejection of the endothelial graft. Inflammation symptoms and signs began decreasing three days after follow-up. Four weeks later, there were no more signs of inflammation in visual acuity.
Another patient, 83-year-old, developed rejection of corneal transplantation in both eyes at once, three weeks after receiving the second dose of the vaccine. She had had surgery six years earlier on her right eye and three years earlier on her left.
Presented symptoms included sudden onset of bilateral blurred vision, pain, photophobia and redness. The treatment consisted of putting in eye drops every hour. Inflammation signs decreased on the seventh day and the drug dosage reduced.
Results
This is the first work describing evidence of a possible association between rejection of corneal transplantation and covid-19 vaccine. Moreover, these are the first reports of DMEK procedure failing after any immunization.
The hypothetical occurrence is that the host antibody response may have triggered the allogeneic response.
Researchers have considered that doctors carry out non-urgent surgeries only after vaccination. In case of done procedures, assess the need to increase drug intake frequency or avoid reducing treatment at the time of vaccination.
Conclusion
There is no hard evidence that immunization against covid-19 led to rejection of corneal transplantation. This is a low-level case report, with regards to scientific evidence. Researches with more case studies, followed for a long time, can confirm the authors’ hypothesis.
Therefore, one must be attentive to any signs of rejection to start treatment at the very beginning.
Reviewed by Paulo Schor, ophthalmologist, free professor and director of innovation of the Federal University of São Paulo (Unifesp) and collaborator of the Faculty of Medicine of the Albert Einstein Hospital.
Follow Phelcom’s blog and stay on top of the main news about coronavirus and the eyes.
Several countries are investigating if covid-19 potentiates infections by other pathogens, such as fungi. In Brazil, a study by the Federal University of São Paulo (Unifesp) analyzed the first two cases of Candida auris in a hospital in Salvador (BA), in December. Currently, nine more occurrences have been confirmed in the same hospital.
The “super-fungus” quickly grows resistant to the main drugs used against it. The report was recently released on Journal of Fungi.
In India, patients have contracted mucormycosis during or after recovery from the new coronavirus. Their numbers are growing more and more. The “black fungus” provokes that a rare infection.
According to Maharashtra State Health Minister Rajesh Tope, at least 90 people have died and 800 are hospitalized in the state alone due to the problem. In addition, local authorities have also confirmed approximately two thousand cases to date. The information was provided to CNN.
There are also reports of patients who needed to have one of the eyes removed due to worsening of the disease. Only in April, ophthalmologist Akshay Nair said, in BBC interview, he treated about 40 patients with the problem. Eleven of them lost one of their eyes.
Read the reports of Indian doctors on the chances of patients developping mucormycosis after covid-19, especially in diabetics, and how it affects the eyes.
Eyes and covid-19: mucormycosis after infection
Mucormycosis is a rare disease that affects the nose, face, eyes, brain and lung. It is provoked by the” black fungus”, very common in damp places such as soil, plants, manure, decaying fruits and vegetables and even in the nose and mucus of healthy people. However, when it attacks immunosuppressed people, such as patients with diabetes, HIV and cancer, it can have sequelae and even be fatal.
From 58 cases of the fungal infection investigated in India, most of the subjects caught it from 12 to 15 days after recovering from covid-19. Such figures are according to a BBC report.
In addition to these occurrences and those reported by Nair, there were more infections reported.
Ophthalmologist Raghuraj Hegde also told the BBC he had treated 19 cases in two weeks, most of them young. Mumbai’s Sion Hospital reported 24 occurrences in two months, four times more compared to the same period last year.
Most had diabetes and had been infected two weeks after recovering from covid-19. From these patients, eleven lost one eye and six died.
Doctors believe that the problem occurs due to falling immunity during treatment with steroids to fight the new coronavirus in severe patients. Another aggravating factor that the drug can cause is increased blood glucose, which further worsens the picture of covid-19 positive diabetics.
Eye treatment
Removing one of the eyes – or even both, in some cases – is an option to prevent the fungus from reaching the brain. In most cases, patients only sought help when they were already losing vision.
Presented symptoms included: stuffy and bleeding nose, swelling and pain in the eyes, drooping eyelids, blurred vision and loss of vision. Black skin spots around the nose may also occur.
The only effective treatment option is an antifungal intravenous injection (amphotericin B) every day, for up to eight weeks.
However, Nair states a way to prevent it: apply the correct dose of steroids, following the designated time span. Then it is necessary to monitor possible rises in the glucose levels of recovered patients.
It is now essential to follow the evolution of mucormycosis cases in India and study its correlation with covid-19 and the eyes.
Reviewed by Paulo Schor, ophthalmologist, free professor and director of innovation of the Federal University of São Paulo (Unifesp) and collaborator of the Faculty of Medicine of the Albert Einstein Hospital.
Follow Phelcom’s blog and stay on top of the main news about coronavirus and the eyes.
Undoubtedly, one of the trends in health technology in 2021 is the consolidation of healthtechs. These startups have spread widely in recent years. Since 2014, they sum up US$ 430 million in investments, according to the survey Healthtech Report Brazil 2020.
Basically, these businesses are aimed at solving problems in healthcare services. They offer innovative solutions that improve both access to healthcare and the quality of life of patients. Innovations range from Artificial Intelligence (AI) diagnostic systems to digital medicine.
Moreover, they operate in the optimized management of public health entities, in smart clinics and offices, in self-service and self-care services and offer cutting-edge technologies for clinical and laboratory examinations.
Learn about how healthtechs work the importance of such segment and some of their solutions that have been revolutionizing health.
Healthtechs: what are they?
First, let’s better understand what a startup is: an innovative company with a repeatable and scalable business model in a scenario of uncertainties. That is: a business model that generates value, but without the certainty that the project will actually work or be sustainable; it can deliver the same product or service on a large scale without many adaptations; and grow more and more in revenue at the same time as costs slowly rise.
To do this, they use cutting-edge technologies such as Artificial Intelligence (AI), machine learning, big data and robotics, inter alia.
Undoubtedly, one of the main advantages of startups is the speed with which they create disruptive products and services. That is, they transform the market and deliver more value to the consumer.
By the way, these are the characteristics that differentiate them from conventional companies, in which any change needs to go through a slower internal process.
So, briefly, healthtechs are the name given to health startups. In English, health means healthcare services and tech, technology. In recent years, the suffix tech has been increasingly used to designate startups.
Photo: Freepik
Importance of healthtechs for health
In Brazil, the lack of access to quality healthcare affects millions of people. Many live in places without basic infrastructure, such as specialists, exams, treatments and medicines, for example.
In this scenario, healthtechs help democratize health by seeking solutions to these bottlenecks. According to Global Health Care Outlook Survey , conducted by Deloitte, these startups are helping solve some healthcare services challenges, such as faster digital adherence; reduction of treatment costs due to the agility and accuracy of examinations and diagnoses; and regulation and compliance, through more effective cost and resource management.
Nowadays, we can say that healthtechs serve the main areas below:
Preventive medicine: prevent or slow the progression of diseases;
Predictive medicine: detect the predisposition of certain diseases;
Proactive medicine: close doctor-patient relationship, supported by technology;
Personalized medicine: using data to personalize treatments;
Medical management: automation and data analysis to optimize processes in hospitals, clinics, offices and laboratories.
Solutions developed by healthtechs
In fact, healthtechs offer services that benefit patients and help doctors and health institutions on a daily basis. Learn about some solutions that made positive changes in the industry:
Accessibility and quick fundus and refraction exams and diagnoses.
In-office, the fundus camera provides a quicker acess to patients’ fundus images. The specialist can diagnose early, start treatment and monitor various diseases, such as glaucoma, diabetic retinopathy and age-related macular degeneration (AMD).
However, the high cost of the equipment – with models of up to BRL100,000 – can make a purchase inviable. The startup Phelcom Technologies created the smart device Eyer to solve this problem.
Coupled to a smartphone, the technology captures high-quality images from the eye fundus and sends them to an online platform, Eyer Cloud. It enables making exam reports, filing a patient’s history and other features.
Price adds an advantage: the equipment is up to 4 times cheaper than a conventional retinograph.
On its turn, startup EyeNetra offers the Mobile Refraction Kit. It can run a vision test through virtual reality. In the end, the technology measures degrees of myopia, astigmatism or hyperopia and prescribes the glasses, which will need the final approval of an ophthalmologist.
More than this, there are products to check the accuracy of the measured degrees and a set of physical lens to evaluate and validate the first test results. Finally, the company also offers a set of software to collect and store all this data for the experts to review.
Machine learning detects disease patterns
More than remote diagnosis, some remote reporting software work with the principle of machine learning. Basically, it collects data, learns from it and improves automatically. It does not necessarily require programming.
In diagnostic medicine, the tool evaluates an extensive database of patients’ symptoms to find patterns for each disease. This way it cheks for a disease in an individual, based on traits presented.
Some systems are available in Brazil. For example, the system Portal Telemedicina analytically compares face-to-face scans to similar cases from a database of 30 million images and scans. The platform provides medical recommendations based on reliable and accurate criteria.
If the medical exam and the algorithm recommendation do not match, it forwards the exam to three other specialists for further evaluation. In addition, the program learns from each report issued, collecting clinical repertoire to its database.
Another innovative aspect is the automatic exam screening, which prioritizes emergency cases in the doctor’s queue.
It is worth noting that data privacy, one of the current challenges of artificial intelligence, is a field of study of the scientific community. In this sense goes the concept of federated learning, which circulates only algorithms and not patient information. This allows multiple locations to share results and improve algorithms.
Accurate diagnosis of breast cancer
Imagine accurately identifying the area of the breast with a suspected change and, on top of that, easing the biopsy? Some artificial intelligence softwares are able to detect breast cancer more precisely.
One of them can even predict unusual image patterns in a mammography and point out the area that requires further investigation. The algorithm was created by Dasa in partnership with CureMetrix, an American startup.
Another software, developed by Google, may be a “second medical opinion” about mammography. The algorithm presented 11.5% more accuracy compared to human analysis.
However, humans show the same result as the machine when two doctors evaluate the exam. Since it is common that two experts analyze the image, Google’s tool can be the “second expert”.
Predicting breast cancer growth and metastasis
EndoPredict genetical test measures the high or low chance of breast cancer spreading to other parts of the body for the next ten years. It avoided chemotherapy in 70% of patients with negative nodules in clinical studies.
However, testing is recommended only for recent disease diagnoses, at early stages , with positive for estrogen receptors and negative for HER2 protein.
With more than 90,000 exam evaluations in the database, the Artificial Intelligence (AI) has learned the subtle patterns in breast tissue that are harbingers of malignant tumors.
The team’s model was significantly better at predicting risk than the existing approaches. It accurately positioned 31% of all cancer patients in their highest risk category, compared to only 18% of traditional models.
Technology helps monitoring diabetes
Imagine checking the glucose rate only by reading a small sensor installed on the back of the arm? With the help of a mobile-like device, the patient tracks the blood glucose levels on the screen without having to stick his fingers several times a day.
Instead of analyzing a drop of blood, as usual, the technology uses the interstitial fluid under the arm to capture the amount of glucose without having to pierce the tissue collect fluid around the vessels.
Moreover, the device checks levels 16 times a day and predicts rises or falls of glucose in the following hours. Thus, it helps to avoid hyper and hypoglycemia.
The technology also allows the user to scan the captured information every eight hours. It remains stored on the device itself or is accessible by the app. Data may be available to the doctor via cloud storage. This way, the expert doctor follows the patient conditions real time and decides for eventual changes in treatment, if necessary.
Another advantage is the possibility to share data with others, such as parents and caregivers. Thus, it is even safer to control diabetes.
Other innovations are the insulin pumps with continuous glucose monitoring, which allow the doctor to determine the exact amounts that should be released to the patient’s blood. Also the inhalable insulin, sold in powder cartridges with three sizes of dosages. The patient inserts the cartridge into an inhaler and, while aspirating, takes the substance to the lung, which is absorbed into the bloodstream. Therefore, it reduces blood glucose levels.
Telemedicine systems: from online consultations to process management
The expansion of telemedicine in Brazil, since last year, has made many clinics and offices look for technologies. The market has several healthtechs that offer complete and innovative telemedicine systems.
We are approximately 7.8 billion people worldwide. Many live in countries with poor access to healthcare. In this sense, it is essential to invest in affordable solutions, in terms of space and cost, which help prevention, treatment and healthcare support.
The healthtechs have come and are committed to such proposal. Their main goal is to improve the people’s quality of life.
Reviewed by Paulo Schor, ophthalmologist, free professor and director of innovation of the Federal University of São Paulo (Unifesp) and collaborator of the Faculty of Medicine of the Albert Einstein Hospital.
Learn more about health technology on the Phelcom blog.
Imagine losing sight in childhood and regaining it partially after many years? This happened to a patient who received an experimental genetic treatment aimed to recover the vision of patients with Leber Congenital Amaurosis (LCA) and severe visual impairment. The rare disease can cause blindness.
The research is led by the Scheie Eye Institute of the Perelman School of Medicine at the University of Pennsylvania in the United States (USA). The report case was recently published in Nature journal.
Learn more about the study and results. Also learn how the patient recovered his vision for longer than a year, after receiving the drug.
The research
The researchers applied intraocular injections of sepofarsen, a small molecule (oligonucleotide), in patients with Leber Congenital Amaurosis (LCA), which caused severe visual impairment.
The oligonucleotide is a small piece of RNA (genetic code) that acts on increasing levels of the CEP290 protein in the photoreceptors of the eye. This results in improved retinal function in daytime vision conditions.
The research used genetic engineering, seeking to improve the function of non-responsive cells. For this, the patient needs to present a relatively good amount of cells so that the surrounding ones could achieve vision recovery.
In 2019, scientists noted that reapplying the injections every three months increasingly improved the vision of the ten patients who took part in tests.
However, recently, the study showed surprising results by injecting only one dose.
Recovering vision after a single dose
After the first dose, the 11th study volunteer did not want to receive the subsequent injections, since he feared developing cataract. Up to then, the patient had reduced visual acuity, narrow visual fields and absent night vision.
However, what happened next surprised the researchers: after a month, the patient had his vision improved. In the following month, however, the patient presented the best results of visual acuity. During 15 months, the effects of the injection were still noticeable.
The results showed that the treatment promoted remarkable changes in the fovea, the core of human vision. “This sets a new standard in which biological improvements are possible with oligonucleotide therapy in Leber Amaurosis caused by mutations,” believes professor and ophthalmology researcher Artur Cideciyan.
Next steps
Now, researchers are investigating the reasons why only one dose generated such satisfactory results, as well as vision recovery. One hypothesis is that, even though they are tiny, RNA molecules are too large to be eliminated early, which allows for continuous visual improvement.
For the next studies, the researchers plan to create specific gene therapies for other hereditary disorders that affect the retina and that are incurable.
Conclusion
In fact, the result of this single patient demonstrates that genetic treatments can increasingly bring more advantages to patients, with longer lasting results.
Several researches are still necessary to confirm such theory. However, this discovery already encourages new studies based on gene therapy.
Reviewed by Paulo Schor, ophthalmologist, free professor and director of innovation of the Federal University of São Paulo (Unifesp) and collaborator of the Faculty of Medicine of the Albert Einstein Hospital.
Follow the main news on eye health on Phelcom blog.
Virtual reality (VR) is a computer-simulated environment that provides visual, sound and even tactile effects to the user. It allows complete immersion in the virtual scenario, as if the person was actually present there. It requires using technologies with stereoscopic displays, such as the popular headsets (special glasses that transmit the simulation).
By such features, it has been increasingly used in several areas. Includind in healthcare.
Check out how virtual reality, apllied to medicine, helps treating various diseases, training professionals and even improving the quality of life of the elderly and physically disabled.
1. Virtual reality in medicine – treatments
Pain
Imagine yourself in a beautiful, relaxing and fun place while undergoing, at the same time, an extremely painful treatment? Experts have found that virtual reality therapy fills the brain with so much information that it leaves no room to process pain sensations at the same time.
Far beyond a distraction, the technique offers a multisensory experience that engages the patient on a much deeper level than watching TV or reading, for example.
One of the first VR programs aims to treat burns. SnowWorld is a game that patients play while having wounds cleaned – a highly painful procedure. It’s simple: users throw snowballs at Penguins. However, it has shown significant results. It was developed by the University of Washington in the United States.
Virtual reality exposure therapy (VRET) has assisted Psychiatric treatment of phobias. The technique provides the most direct approach to the patient’s fears and anxieties in a controlled environment. Instead of visualizing or mentalizing fears, patients can experience the situation in the doctor’s room, a safe and welcoming space.
During therapy, doctors and psychologists can evaluate the patient’s responses and indicate the best treatments.
Anxiety, depression and obsessive-compulsive disorder (OCD)
Virtual reality exposure therapy (VRET) is useful to treat other disorders, such as anxiety, depression, panic syndrome and obsessive-compulsive disorder (OCD).
The virtual environment enables patients to experience changes in body and mind, similar to real-life symptoms. All in a controlled scenario followed by specialist psychiatrists and psychologists.
A study, conducted at St George’s Hospital in London, evaluated the use of VR in patients suffering anxiety during hospitalization for surgical procedures. They viewed soft landscapes and listened to relaxing music through a headset.
The results: 94% were more relaxed, 80% felt less pain and 73% less anxiety.
Phantom Pain
Virtual reality in medicine has presented results even for treating phantom pain in paraplegic and amputee patients. For instance, games stimulate the user to move the missing member.
They help the brain reconnect the stimulus areas, temporarily alleviating or even eliminating pain.
2. Physiotherapy
Virtual reality assists physiotherapy in the evaluation and rehabilitation of patients with movement, postural balance and pace disorders. For example, non-immersive technologies create machine-learning exercises to customize each activity to the patients’ therapeutic needs.
Several studies have shown that VR reduces dizziness and other otoneurological symptoms and improves body balance, functional capacity and quality of life.
3. Cognitive rehabilitation
Several researches study virtual reality applied to rehabilitation and cognitive development of people suffering from neuropsychiatric disorders and autism. In the latter case, scientists from the University of Texas, in the United States, simulated virtual realities where patients interacted with other people, such as shopping in the supermarket.
After training, eight young autistic adults – without great impairment of understanding – showed significant increases in social cognitive measures of theory of mind and recognition of emotions and in the social and occupational functioning of real life.
Currently, the university offers a training program that helps young autistic people exercise their social skills.
4. Training doctors
Virtual reality applied to medicine enables both seeing details of the human body and training procedures.
Student, have you ever thought of practicing surgery in a virtual environment? Stanford University has a simulator which allows the most complete training for the future surgeon.
According to researchers, a three-dimensional virtual surgical environment can enable enhanced pre-operative planning and testing. It thus improves results of patients, decreases complication rates and improves technical skills.
Stanford Virtual Surgical Environment (VSE) is developed for rhinological procedures. The technology allows the surgeon to interact with patient-specific three-dimensional reconstructions of nasal sinus CT data sets using a modified haptic interface device, triggering a virtual endoscope.
There are also options of simulators in Brazil. For example, Eyesi aims to train intraocular cataract surgery and vitreous retinal procedures. It presents performance score and evaluates tremor, movement accuracy and repetitiveness, inter alia. Therefore, it is a good tool either for interns to practice or to introduce new techniques in ophthalmic surgery.
There are also the Da Vinci surgical systems for robotic surgery, already in their fourth generation. The equipment carries out minimally invasive surgeries in different procedures. One of its tools is a console – inspired by flight simulators – in which doctors view the high-definition 3D images and make the operative movements with their own hands, which are transmitted to the robot.
It also applies to train and requalify students, residents and specialists in attendance or within the surgical center. Thus, it is possible to analyze ethnographically the performance of the professional – alone during the procedure – by capturing images from a local camera.
5. Quality of life for the elderly and physically disabled
+
Elderly and physically disabled may have difficulty moving. So, imagine them able to experience cycling and diving in the sea? Some projects develop these scenarios so realistically that they even change the temperature, light and wind direction.
There are also technologies that help reduce depression in elderly people who live bedridden or in nursing homes. The startup Rendever uses virtual reality devices, personalized software and a tablet to recreate memories chosen by the elderly, such as their home during childhood.
Reviewed by Paulo Schor, ophthalmologist, free professor and director of innovation of the Federal University of São Paulo (Unifesp) and collaborator of the Faculty of Medicine of the Albert Einstein Hospital.
Are you interested in virtual reality applied to medicine? Follow Phelcom’s blog and stay on top of the main news.
English
Português
Español
