The Emergence of Internet of Things and Digital Biomarkers in Medical Artificial Intelligence
Medical Artificial Intelligence is the process of combining AI into healthcare and or preventive medicine. Currently, mankind is undergoing the fourth industrial revolution- “blurring of boundaries between the physical, digital, and biological worlds” (Mcginns, 2018). It is the advancement of technologies such as artificial intelligence (AI), the Internet of Things (IoT) and quantum computing. Medical AI can be used in three different ways: (1) to diagnose a disease or predict the treatment outcome based on clinical data or genomic data; (2) read medical images in order to diagnose diseases, replacing doctors in this aspect and (3) to prevent or predict a disease by monitoring biomedical parameters in real time.
There has been an increase in interest regarding AI in the medical field and how it is closely related to IoT. IoT is “the network of physical objects that contain embedded technology to communicate and sense or interact with their internal states or the external environment” (Hoy, 2015). The “thing” in IoT can be anything, not limited to mobile or wearable devices. IoT is usually composed of devices with embedded sensors, gateway, cloud, analytics and a user interface (Lo, 2019). The device with the embedded sensor collects information continuously from the environment and it is received in the gateway. After which, this information received is delivered to a cloud system in order to collect, analyze and store large amounts of data produced from devices in real time. Here, AI plays a role in analyzing this data received, then feedbacks its analysis to the user interface for patients or medical staff. In this system, the data collected from patients in real time is the digital biomarker, a biomarker that can be measured using digital devices and be used to explain or predict health-related outcomes (Wang, 2016). This includes all human data that can be measured via a digital tool.
This is especially prominent in our human nervous system, specifically our peripheral nervous system, the main component being the spinal cord. AI can be used to predict the outcome after surgery or patients with the condition spinal metastasis. Here are some examples.
Spinal Posture. This can be monitored via a wearable device. By monitoring spinal posture in real time, feedback can be constantly provided to the user if he has bad posture persistently. This feedback can be used by doctors to take measures in order to prevent spinal disease caused by bad posture.
The Electromyography (EMG) test is most commonly used to evaluate spine patients. The signals read by the EMG is also a digital biomarker, since it detects electrical signals that muscles receive from nerve cells (specifically, motor neurons). This is conducted only within a hospital. Based on the feedback the system gets, doctors can deduce the electrical activity in response to a nerve stimulation of the muscle and conclude whether the patient has any neuromuscular diseases, motor problems, nerve injuries or degenerative conditions.
Human gait is another digital biomarker in the spine. In Parkinson’s disease, some common symptoms are walking difficulties, balance and coordination issues. In these cases, a patient's gait pattern can be monitored via sensor build-in shoes or other wearable devices. Furthermore, gait disturbance is also an early sign of compressive myelopathy (injury to spinal cord due to severe compression). If gait is monitored in real time, it is possible to diagnose the patient quicker and more accurately, pinpointing exactly the progression of the patient’s myelopathy.
Figure 1: Illustration of the human gate, taken from kodsi engineering
Thumbnail credit: IE University