Clinical Fellowship Chronicles: 4 Medical Specialties Using Computer Vision Right Now
This blog post is the third in a 12-part series called the Clinical Fellowship Chronicles. In this series, we’ll be following The DICE Group’s latest clinical fellow Tiffany D’souza as she explores new ways to bring the medical and tech worlds together. Follow us on Medium to learn more!
Computer vision, a form of artificial intelligence where computers learn to interpret digital images, could have a tremendous impact on imaging-based medical specialties. As I learn more about computer vision, I see potential for this technology to expand into the following four areas of medicine.
Diagnostic Radiology and Nuclear Medicine
Computer vision is quickly expanding into Diagnostic Radiology, a field that uses medical images like X-rays, CT scans and MRIs to reach clinical diagnoses. Today, systems are being developed to correct out-of-focus radiographs, read “noisy” scans and even clear and interpret those scans. This could help physicians pick up abnormalities most humans might overlook while decreasing their exposure to radiation.
Machine learning models are also being used to estimate patient wait times for scans and predict disease onset much earlier. This allows Radiologists to focus on treating complicated cases and dispensing quality patient care.
Pathology
Pathology is an image-driven field where pathologists spend hours reviewing blood smear slides and biopsy samples to locate and diagnose diseases. Pathology characteristics that can be viewed under a microscope are being converted to digital images, which can then be put into comprehensive databases programmers can develop models on.
In addition to slide imaging, research is being done in tissue genomics, intraoperative diagnostic processes and technology for clinical research. This growing trend towards digitization makes pathology a great place to explore use cases for computer vision.
Ophthalmology
Since most eye diseases like cataracts, glaucoma and macular degeneration are diagnosed through imaging tests, computer vision is starting to take off in Ophthalmology.
Some systems can even predict the onset of diabetic retinopathy and deliver a prognosis, helping with early detection and intervention. Once we fully understand how to leverage this technology, computer vision could have a huge impact on how we diagnose and treat eye diseases.
Dermatology
Dermatology examines skin abnormalities and pathologies, making it a highly visual field. In fact, computer vision is already being used in early detection for skin cancer and completing risk assessments for lesions.
While it’s still important to have a doctor complete a physical exam, machine learning systems can help with primary diagnosis and triage by processing a range of skin disorder images. This has huge implications for telehealth and remote consults, in addition to diagnostic processes.
Impact on the Industry
Image-driven medical specialties can benefit from pattern recognition and computer vision.
Just by looking at a radiograph or skin image, these systems can quickly and accurately identify early stage abnormalities, which could have an amazing impact on teleconsultations, diagnostics, clinical decision making, research and even medical training.
However, these machines do have limitations, since they can’t provide the empathy and nuance needed when dealing with patients. Plus, machines need close human supervision to ensure they’re making accurate diagnoses, so it will take some time before certain parts of medicine can be completely automated. Nevertheless, these systems show tremendous potential as diagnostic tools that allow providers to focus on complicated cases and create meaningful patient interactions.
Tiffany D’souza
Tiffany is completing a Healthcare Innovation Fellowship at The DICE Group before applying to her residency. She is also an online fitness coach and social media marketing consultant who loves baking cookies, solving crossword puzzles and exploring the potential of healthcare technology.
