Using the example of Steve Jobs, we’ll discuss the challenges of early diagnosis and treatment of rare diseases and explore the potential for AI to drive future medical advances. The future of healthcare will be centered on AI’s diagnostic and therapeutic capabilities and the use of big data.
On October 5, 2011, Steve Jobs, an icon of innovation and convergence, passed away. His life was cut short by the failure to treat his pancreatic cancer early. His pancreatic neuroendocrine tumor is extremely rare, with fewer than 1 in 100,000 people diagnosed each year. It is very difficult to see the tumor visually using imaging modalities, so it is diagnosed by the hormones it releases and the symptoms it causes. However, in clinical practice, the disease is so rare that it”s difficult to recognize it, and in some cases, there are no symptoms. In addition, because it is impossible to diagnose with blood tests, treatment is often delayed. However, in the future, it will be easier to detect tumors that are difficult to diagnose early. Not only will we be able to diagnose them early, but we will also be able to treat them immediately. The doctor of the future who will give us early diagnosis and treatment is artificial intelligence. In the future, various forms of medical AI will be developed. They will be able to cooperate with each other and form a competent team that will be in charge of the diagnosis and treatment process. And these teams can be very small and can be located anywhere in the world.
In Korea, there are no regular conferences on AI in healthcare. In Europe, on the other hand, forums and conferences on AI in healthcare have been held regularly for more than 30 years since 1985, and specialized journals have been published since 1998. The main focus of these conferences has been on technologies to diagnose diseases and find their causes based on data. AI is better at analyzing data than humans, so data is the key ingredient for AI to excel. AI utilizes meaningfully collected “big data”. When we say “big data,” we don’t just mean large amounts of information. It means information that has been carefully refined to bring new knowledge. Big data fulfills three conditions: it’s quickly available, it’s massive, and it’s in many different forms. An AI medical team will have three main tasks. First, the team will need to collect data to diagnose the disease. Then there is a team that analyzes the data to diagnose the disease. The data analysis and diagnosis team will use deep learning from electronic medical records to provide doctors with useful treatment methods and diagnoses. Once the diagnosis is made, tiny robots can be used for early treatment.
Let’s take a look at the data-gathering part of the AI healthcare team first. To see the role of the medical team in action, let’s consider a situation where an image is taken to find cancer cells. The team consists of two main parts One is responsible for improving the quality of the images, and the other is responsible for obtaining objective data. The AI that improves the image quality creates a statistical model based on the data. Based on the statistical model, it determines whether the information in the image provides useful data for diagnosis. Patch Group Prior Based Denoising (PGPD) technology can be used here. PGPD removes noise, which is information that is not useful for image interpretation, and replaces it with information that is organized by the statistical model by comparing it to the surrounding information. Data is key in the process of finding and correcting unhelpful parts of the image. Using big data, AI can build a statistical model and use it to correct the video.
Once the footage has been corrected, data needs to be extracted from it. Before the development of artificial intelligence, there was no processing involved once the images were taken. Doctors would take the images as they were and use them for diagnosis. In recent years, technology has developed to allow computers to extract more objective information from the images. The computer converts the image into digital information and then analyzes the brightness of the pixels and their context with the surroundings to give them a quantifiable value. It reconstructs blurry and hard-to-see areas using the surrounding context. The AI obtains information about the size, shape, and texture of the cells in the image. This information about size, shape, and texture can then be used to detect cancer cells.
Once the data extraction process is complete, it needs to be analyzed to make a diagnosis. The diagnosis made by A.I. is not just a simple knowledge based on the patient’s symptoms. It collects information based on accumulated big data to make a diagnosis. Big data is created through the process of extracting data, converting it into a form, and storing it under an analyzable system. Artificial neural network models are used as a way to utilize accumulated big data. An artificial neural network consists of an input layer, multiple hidden layers, and an output layer. The input layer receives the data first and passes it to the first hidden layer of the neural network with the greatest degree of association. Once the data reaches the hidden layer, it is mathematically processed in a complex way and passed to other hidden layers for correlation before finally reaching the output layer. As the data moves through the hidden layers, it is processed and the degree of relevance to a particular hidden layer is iteratively modified to ensure high accuracy. In a medical context, the inputs are symptoms and patient information, and the output is a diagnosis.
Once we have a diagnosis of the patient’s disease, we want to be able to treat the disease immediately. If the disease does not require surgery, it can be treated in a simple way using AI. This is done by injecting tiny robots into the blood. These tiny robots are able to recognize the location of cancer cells and can travel through the blood to locate and eliminate them. This method of injecting robots into the blood can also help treat chronic diseases. Diabetes requires constant measurement and control of blood sugar. Therefore, artificial intelligence can monitor the condition of the blood at all times and recognize dangerous conditions. If there is a robot in the blood, the hormones that regulate blood sugar can be used appropriately. AI also uses artificial neural networks to prevent medical errors during surgeries for treatment. It can prevent mistakes in coronary artery bypass grafting procedures, which require quick judgment and can be prone to mistakes. By building neural network models, AI can analyze the relevance of a decision to the situation in a short period of time to catch suspected mistakes.
Now that you know what AI medical teams are capable of, you can imagine them in action in the future. Going to the doctor is not easy and difficult. In the future, patients won’t have to go to the hospital, but will be treated at home by an AI medical team using a medical kit. Medical kits can be distributed to every home. The kits can take pictures of a person’s body or analyze blood. Through the photos and blood, the kit uses the AI system and obtains data. The data is managed by a centralized control center. The center stores the individual’s family history and medical history so that the photo and blood data can be analyzed precisely. The data is further analyzed through an artificial neural network model and a final diagnosis is made. The kit includes a device to insert a miniature robot. The robot receives the diagnosis and the individual’s biometric information and performs the appropriate treatment. If these medical kits are mass-produced and distributed in large numbers, our lifespan will be greatly improved.
Artificial intelligence technology is advancing at a furious pace. It’s not only making our lives easier, but it’s also extending our lives. In order to extend human life, AI fits into a very small kit and creates a medical team of doctors. The medical team fulfills three main roles quickly and easily. It can take pictures, and it can enhance the images to make them better than they are. They enhance the images and then derive objective numbers from the images that are difficult for humans to judge. The central control center stores big data about an individual’s family history and biometric information. The data obtained from the centralized control center is sent to the centralized control center, and analysis begins by synthesizing big data and extracted data. An artificial neural network model is used for analysis. The final diagnosis is made through data analysis. Even after the diagnosis is made, a small robot that enters the human blood can be used for immediate treatment. In conjunction with artificial intelligence, tiny robots can quickly eliminate the cause of the disease and treat chronic diseases.
In the future, nanotechnology will allow us to include many devices in a small space, which will lead to the development of small-sized universal medical kits. These kits will help humans quickly and continuously treat diseases with the help of artificial intelligence, as described earlier. AI will disrupt the medical field by putting doctors in medical kits. In the future, hospitals will be unnecessary in the long run and may even disappear. A society where you don’t have to travel far or spend a lot of time in the hospital to treat your illness is on the horizon. In an extreme case, as Ray Kurzweil suggests, advances in artificial intelligence could make it possible for humans to live forever.
