In fact, artificial intelligence in the diagnostic medicine has been showing positive and promising results. More than that, the development of more and more innovative software has been revolutionizing the sector by delivering more speed, precision, quality, accessibility and cost reduction to patients, professionals and institutions involved.

For example, Google carried out a research that showed the use of AI can improve the diagnostic accuracy of ophthalmologists . Another study by the company also proved that ophthalmologists and algorithms are more effective when working together .

Undoubtedly, the use of computerized systems and computational symbols in the prediction and imitation of human behavior directly impacts the patient health and the cost reduction in treatments, among other numerous benefits.

Learn about 4 technologies that apply artificial intelligence in the diagnostic medicine: Eyer handheld fundus camera, remote reporting software, breast cancer genetic test and mammography software.

1.   Eyer handheld fundus camera

The Eyer handheld fundus camera was one of the main equipments used in the project “Unidos pelo Diabetes em Ação, in Itabuna (BA), Brazil.

Connected to a smartphone, Eyer handheld fundus camera carries out high-quality retinal exams, in only a few minutes, with no need to dilate the pupil. Integrated with an online platform, the data is automatically sent and can be analyzed by a specialist anywhere in the world. That is, it allows remote diagnosis.

Learn about the equipment functionalities:

• High-quality retinal exams by smartphone;
• Accurate and fast diagnostics;
• Lower cost compared to traditional fundus cameras;
• Portability, which allows exams in several locations;
• Democratization of Retina exams, mainly in places with low infrastructure of quality services in the area, such as doctors, health professionals, equipment, medicines, etc;
• Faster service through computerized systems integrated to an online platform accessible through computers, smartphones and tablets;
• Exams are easier to carry out in clinics and health centers;
• Diagnosis made by specialists and leading professionals from anywhere in the world;
• Reduction of service time and operational costs;
• Decrease in the displacement of patients to hospitals and large urban centers;
• Improvement in the quality of the issued reports;
• No use of eye drops for pupil dilation;
• Increased prevention and early diagnosis of diseases such as diabetic retinopathy, glaucoma, cataract, macular degeneration, retinoblastoma, retinal detachment, retinopathy of prematurity and blindness, inter alia.

In Brazil, the startup Phelcom Technologies offers the Phelcom Eyer handheld fundus camera . In addition, it also offers the online platform Eyer Cloud , which allows storing and managing patient exams. It ensures data backup on a secure server and the physician has all data organized in a friendly, functional and intuitive interface.

Remote reporting software

Some remote reporting software enable remote diagnosis and work with machine learning, an AI application. In Portuguese, the term means “machine learning”.

How does it work? Basically, it collects data, learns from it and improves automatically – without necessarily being programmed.

In the diagnostic medicine, the tool evaluates an extensive database of patients’ symptoms to find patterns for each illness. The software can check which disease the individual has, according to the evidence it presents.

There are some systems available in Brazil. Portal Telemedicina, for example, analytically compares face-to-face exams to similar cases in a database with 30 million images and exams.

The platform develops medical recommendations with reliable and accurate criteria when using Deep Learning. This method is based on complex algorithms that copies our brain neural net, giving the system an ability to detect medical findings at the superhuman level.

If the medical exam and the algorithm recommendation do not match, it sends the exam to three other specialists for a more detailed evaluation. The program even incorporates learning into each report issued, accumulating clinical repertoire in its database.

Another innovative aspect is its capacity to perform automatic exam screening allowing emergency cases to take priority in the doctor’s queue.

Software for mammography reports

Imagine accurately identifying the location of the breast with a suspicious change and, on top of that, easing the biopsy. Some artificial intelligence software are able to detect breast cancer more accurately.

One of them can predict unusual patterns of the mammography image and point out the region that demands further inquiry. American startup Dasa, in partnership with CureMetrix, created the algorithm.

Another software developed by Google may be the “second medical opinion” on the mammography. This is because the algorithm showed 11.5% more hits in relation to human analysis.

However, when evaluated by two doctors, humans show the same result as the machine. And, as it is common for two experts to analyze the image, Google tool can be the “second expert”.

Genetic testing – EndoPredict

Artificial intelligence in diagnostic medicine is also managing to predict the development of breast cancer and possible metastasis for the years to come.

The  EndoPredict genetic test, for example, evaluates 12 tumor tissue genes related to the probability of recurrence. The result is a score that indicates whether the chance of the cancer to spread elsewhere in the body is high or low over the next ten years. Thus, it avoided chemotherapy in 70% of patients with negative nodules in clinical studies.

However, the test is indicated only for recent diagnoses of the disease and at an early stage. Moreover, positive for estrogen receptors and negative for HER2/neu protein.

The test is offered exclusively by GeneOne, Dasa genomics laboratory.

Machine learning

The Massachusetts General Hospital (MGH) and MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) , in the United States created a new model of machine learning. It allows predicting, from a mammography, if a patient will develop breast cancer within five years.

With more than 90,000 exam evaluations in the database, the AI has learned the subtle patterns in the breast tissue that are malignant tumor precursors.

The team model was significantly better at predicting risk than existing approaches. It accurately positioned 31% of all cancer patients in its highest risk category, compared to only 18% of traditional models.

Conclusion

Artificial intelligence in diagnostic medicine will surely revolutionize the area and help doctors and health services to increase productivity and attend more people.

Undoubtedly, it is also essential to democratize access to healthcare. Mainly in places with low infrastructure of quality services in the area, such as doctors, health professionals, equipment, medicines etc;

In fact, intelligent algorithms will not replace doctors who specialize in this type of diagnosis. However, it can serve as support, offering safer and more accurate reports.
Is artificial intelligence in diagnostic medicine a subject that interests you? Follow Phelcom’s blog.