Chapter 1

Why AI for Medical Applications?

Digitization has transformed healthcare. From electronic health records and telemedicine to digital imagery for diagnostics and other smart medical devices, digital tools abound. The result? An abundance of data. Engineers, scientists, and researchers in the healthcare space see the potential to use artificial intelligence (AI) to tap into these datastores and create diagnostics, interventions, and services that improve healthcare outcomes.

One example is Kaiser Permanente’s SureNet System, which uses AI to scan electronic health records to identify patients with undetected conditions. In one example, the system searched 3.86 million patient records for patients who had not been screened for abdominal aortic aneurysms (recommended for most men ages 65 to 75). Screening resulted in 2,062 new diagnoses of aneurisms, 87 of which required surgery. The system slashed the percentage of unscreened patients from 51.7 to 20 percent [1].

An AI-enabled record scanning system reduces the proportion of unscreened patients (in yellow) in the database from 51% to 20%.

AI simulates intelligent behavior, but as seen in the SureNet example, it can rapidly process stores of data too large and too complex for a human to interpret. AI-driven systems can be used in many ways to integrate AI algorithms, such as machine learning and deep learning, into complex environments that enable automation.

There are many spaces in healthcare where AI is playing a significant role in improving patient experiences and generally improving the health of populations:

  • Improving medical decisions in large populations.
  • Improving diagnoses directly from images or physiological signals.
  • Enabling new kinds of treatments and interventions.
  • Providing scalable screening and therapeutic options that improve access to healthcare.

Delivering on the Potential of AI

AI has the potential to be the future in medical devices. To deliver on its promise, you—engineers and scientists in the healthcare space—need to be able to:

  • Implement AI solutions even if you aren’t a data science expert.
  • Design, simulate, and validate AI-based models quickly and cost-effectively.
  • Generate, collect, and prepare good quality, labeled data.
  • Integrate AI into existing algorithms and systems cost-effectively.
  • Certify your AI-based models so you can launch them in a regulated market.

MATLAB® and Simulink® make it easy to start working with AI models for medical industry applications. They provide a complete workflow for AI, which involves preparing the data, creating a model, designing the system on which the model will run, and deploying to hardware or enterprise systems. MATLAB and Simulink also provide workflows to help certify your AI-based software as a medical device.

Six boxes connected by arrows from left to right in a flow from data preparation to AI modeling to system design to verification and validation to deployment to certification.

MATLAB and Simulink workflow for successful AI implementation.

The following chapters describe case studies showing how MATLAB and Simulink were used by organizations to implement AI for improving patient outcomes in various areas of diagnosis and therapeutics and improving health of populations.


[1] Rochman, Sue. “Researchers conduct largest-ever study of abdominal aortic aneurysms.” Kaiser Permanente. Updated September 20, 2021. https://divisionofresearch.kaiserpermanente.org/abdominal-aortic-aneurysms/