It often feels like there’s no corner of healthcare that artificial intelligence (AI) hasn’t touched. Medical devices are no exception. The use of AI in medical devices originated back in the 1990s with imaging applications that were dependent on locked algorithms operating statically. We are now dealing with adaptive and evolutionary AI algorithms. These capabilities provide revolutionary unlocks in technology from personalised care to faster diagnostics and better decision-making in health services.

Then, in the early 2000s, AI started to live up to its earlier promise. Healthcare workers were using AI to screen for diseases such as diabetic retinopathy and skin cancer with great precision. The US Food and Drug Administration (FDA) appreciates that AI and devices enabled by machine learning (ML) will revolutionise healthcare. Since 2020, the FDA has approved nearly 730 "new" AI-enabled devices; now we are observing how the field has evolved. As of June 25, 2024, there are 950 FDA-approved AI devices, indicating the speed at which AI technologies are being adopted by the medical device industry.

AI and ML in medical devices have made significant advancements and possibilities for diagnostics, treatment, patient care, and data analytics, while advancing the possibilities for medical advances. There are administrative uses, as well as clinical, but both in clinical and primary care. In either case, it comes down to medical devices to support, enable, and inform healthcare and pharmaceutical professionals. For patients, this has shown benefits in timely, accurate diagnoses, treatment efficacy, and even more personalized patient care.

Radiology accounts for the bulk of AI medical device approvals

AI-enabled medical devices are making significant inroads across various specialties, with radiology leading the way. Companies such as GE Healthcare, Siemens Healthineers, and Philips are known as major manufacturers of medical devices and lead the AI devices market with the most FDA-cleared products. Approximately three-quarters of all FDA-approved AI-enabled devices are in the radiology space, and one-quarter have the specific designation of radiological imaging processing systems or software. AI technology is ideally suited for the analysis of medical images (such as X-rays, CT scans, and MRIs), which are at the core of radiology. The abundance of imaging data, rapid and accurate assessments, and workflow automation all position AI as particularly useful in radiology.

‍

Companies are developing more AI-enabled devices in other areas. For example, cardiovascular was the second most prevalent specialty, which includes devices like e-stethoscopes and other software that can identify heart arrhythmias and signs of heart failure. Neurology AI devices utilize EEG technology to track brain activity in brain-injured patients, as well as software that detects abnormal levels of electrical activity in the brain. Overall, the numbers are promising for AI-enabled medical devices, particularly in radiology, and indicate significant room for growth in other specialties. 

How AI is helping in device development and innovation 

The use of AI as a revolutionary development and tool for supporting researchers and discovering innovative treatments in the medical devices space is huge. The areas where AI is driving change in medical devices are as follows:

Diagnostic reliability

AI-enabled medical devices can analyze complex medical images (X-rays, MRI, and CT scans) for abnormality analysis, sometimes with precision that rivals or even exceeds the best human experts. AI models are also enabling early and accurate detection of diseases, such as cancer, cardiovascular diseases, and neurological diseases. This is a hopeful sign for the future of AI in medical imaging.

Personalized patient care

One of the most impactful ways that AI is being used in healthcare is in its capacity to personalise treatments for patients through medical devices. This is because AI can analyse data in real time as medical devices are being used by clinicians. The medical device can then apply the treatment outcomes, guided by the real-time data recorded by AI, to provide a personalized service to patients.

Automation and workflow optimization

The use of AI in medical devices has resulted in increased amounts of streamlined automation. It can reduce the risk of human error in complex procedures. For example, AI algorithms are used to help map brain tumour surgeries. In addition to this, AI has the potential to spot opportunities that may be missed by humans by combining a large amount of real-time data and patient records.

Enhancement of medical devices

AI applications in the healthcare and life sciences are used to improve medical devices to improve the experience for both health professionals and patients. Machine learning systems use existing techniques to analyze large amounts of data, identify patterns, assist in decision making, and improve performance with little help from humans.

Real-world applications of AI in medical devices

Imaging and diagnostics

AI is already improving medical imaging machines in real-life healthcare systems to reconstruct images with a higher, clearer resolution. Hospitals and clinics are employing such systems in order to expedite diagnosis and identify complications more quickly, enhancing the evaluation and decision-making.

‍

Reduced reporting time in radiology

Hospitals that have deployed AI-driven radiology platforms are reporting a 30% reduction in reporting time for radiologists. For example, deep learning systems that are trained on large datasets automate the detection and labeling of features in scans, streamlining workflows and lessening the burden on specialists. This is allowing hospitals to save money on radiology, in addition to improving the accuracy of care and quantifiable outcomes by speeding diagnosis and enabling intervention earlier.

Surgical innovation

AI is being applied in actual operating theatre scenarios to analyze surgical video footage and automatically generate "highlight reels" of key surgical moments. Surgeons utilize these videos to evaluate their performance in the operating room, which permits learning again from complex circumstances and refining their surgical techniques. This approach allows surgeons to enhance their skills without having to review hours of video recordings.

How Segmed supports the AI evolution in medical devices

We have repositories of medical images and data, with access to over 100 million medical images and data points from every region of the world, providing a world of real, clinical information and access to broad and diverse datasets to be used to build and improve AI models.

By providing curated, regulatory-grade datasets with images from different demographics and different medical conditions, Segmed reduces bias while simultaneously increasing model accuracy and generalizability. This addresses challenges with AI-powered medical devices achieving precise anomaly detection across different patient populations. 

Segmed's datasets are also used to enhance AI algorithms for diagnostics and decision support for improved and earlier detection of a medical condition. Better-performing AI models on medical images result in better patient outcomes primarily through earlier, safer, effective, and personalized interventions in the healthcare experience.

‍Connect with us to find out how our high-quality, diverse, tokenized imaging datasets can support a myriad of research and development for medical devices R&D.