Computer Vision in Healthcare
In the life sciences and healthcare sector, computer vision can improve diagnostics, medical treatments, and procedures while also accelerating healthcare research. The technology can give patients an improved overall experience and provide medical professionals with the real-time visual data they need to make better decisions while ensuring safety and cost efficiency.
Current Trends
Examining samples of body tissue in a laboratory is a cornerstone of modern medicine, ensuring correct diagnosis and treatment. But a shortage of pathologists – exacerbated in the post-pandemic era – and rising biopsy volumes are challenging. Meanwhile, advancement in cell and gene therapy has created opportunities for highly personalized patient medication. Also on the increase is remote monitoring of patient condition and compliance with treatment regimes, enabled by digital healthcare and the new healthcare delivery paradigm. Patients as consumers are more proactively involved in their own healthcare decision making and they use more digital and wearable devices to collect information about their health status.
Here we explore three important computer vision healthcare applications.
Medical Image Reading and Reporting
Accuracy, speed, and fewer errors. These are the key benefits of computer vision in recognizing patterns to inform diagnoses and improve patient outcomes. Many healthcare facilities now use this AI-based technology to extract and analyze vital visual information from MRI and CT scans, X-rays, ultrasounds, and other medical images.
Computer vision is used mostly for image interpretation and reporting, while enabling health system interoperability, efficient data management, and cloud access to shared visual data.
The overall number of intelligent screening solutions to support diagnosis is growing fast. One example is the AI-powered smart robotic microscope from Indian startup SigTuple. By analyzing visual medical data, this device removes drudgery from routine microscopy, allowing pathologists to spend time where it matters most – with critical patients.
Lithuanian startup Oxipit offers an autonomous AI computer vision tool identifying healthy chest X-rays (those without any abnormality) and producing finalized patient reports without requiring any human intervention at all. This frees up the radiologist’s time to focus on more pressing cases.
Early Detection of Diseases
Doctors use their experience and intuition to identify disease, often using tests to verify their diagnoses. But traditional tests don’t always reveal the problem, particularly early indicators of disease. Computer vision provides the support that medical professionals require. Systems can go beyond identifying anomalies to also detect disease; they are trained, by learning from images of healthy and diseased tissue types, and algorithms can recognize patterns and notify doctors.
A study from researchers at Google Health in collaboration with Imperial College of London designed an algorithm that shows this technology is more efficient and accurate in diagnosing breast cancer than human radiologists.
Clinicians can send their medical images to the lab of New York startup PreciseDX to receive a detailed report with AI-powered insights within 2-3 days, using this additional information to inform and enhance their usual treatment plans. The AI-based diagnostic software startup Paige has developed a solution to detect the spread of breast cancer to the lymphatic system – a spread that is most at risk of being missed and a task that’s tedious and time-consuming for the pathologist but critical to patient health.
Accurately and swiftly analyzing prostate biopsies is challenging due to increasing case volumes, subjectivity in grades, the small size of certain tumors, the large number of samples per case, and more. The Israel-based medical technology manufacturer Ibex has developed an AI-powered, computer vision solution to help pathologists in improving the detection and grading of prostate cancer.
Enhanced Medical Procedure Efficiency
When performing complex surgeries, medics need to make critical decisions that can profoundly affect the patient. Computer vision can be used to boost surgical success rates. For example, a computer vision system can automate the recording of a surgical procedure and then superimpose this virtual video footage and still images on the surgeon’s view in real time as this person undertakes the same type of operation. This can guide and train surgeons as they work and can also be used to support them with a variety of repetitive and typically error-prone processes.
Promising “surgery at your fingertips,” health technology startup Touch Surgery, part of Medtronic, has developed an AI-powered surgical video and analytics platform for the operating room. Viewing procedures on a monitor and even on a smartphone, this solution makes it easy to prepare, practice, and teach more than 200 surgical procedures across 17 different specialties and explore new techniques.
