Diabetics Have New Super Power in Fight Against Retinopathy - Big Data

Ruth Smith
April 23, 2019

WPC Healthcare implements a “better together” approach by using retinal images with data science to provide early warning of disease progression for diabetic retinopathy. EyePACS, a provider of picture archive communication systems, offered the original data set for analysis in order to improve accuracy of the stage of diabetic retinopathy.  Using the industry standard, a five-part scoring system, each individual patient was evaluated for degree of risk, and the results compared with the physicians original diagnosis. Most ophthalmologist see about 5,000 patients per year (or 10,000 eyes). In this case, the analysis included a half a million images to train a computer to recognize the disease stages by “looking at” a more expansive view of the image, and discounting errors. Essentially, 50 years of clinical experience compressed into a single 24-hour period. This provides an enhanced ability to analyze a vast number of images using software greatly improving the quality of diagnosis. It is an example where data science combined with the physician’s expertise can confidently classify various stages of diabetic retinopathy. Are you doing something innovative to combine physician expertise with data science? Please share your story or question here.  

Several potential business uses and conclusions came from this project:

  • In rural settings, physicians can read images with software and obtain “second opinions” without requiring patients to go anywhere else.
  • In high-volume radiology, use cases such as mammography typically require a dedicated radiologic resource. Most radiologist read everything (broken bones to scans) with a computer backup providing expertise that a single radiologist can’t gather in a lifetime. Imagine having data science expertise to quickly and accurately offer a second opinion or confirm a diagnosis.  
  • Images form a significant part of a patient’s medical history. Relying on a computer to process large volumes of images and approve the results is a game changer in precision medicine.  

A well-known fact is that diabetes is a dangerous and potentially life threatening disease. What some may not realize is that one of the conditions associated with diabetes is a type of vision loss called diabetic retinopathy. This condition is treatable and best outcomes are usually due to an early diagnosis by a physician or optometrist. This solution was aimed at the more than 10 million Americans (and 93 million people globally) who suffer with diabetic retinopathy, according to the Centers for Disease Control and Prevention (CDC). Diabetic retinopathy is a condition suffered by diabetics that causes progressive damage to the retina resulting in damage to the tiny blood vessels that nourish the retina. They leak blood and other fluids that cause swelling of retinal tissue and clouding of vision.  In other diabetics, abnormal new blood vessels grow on the surface of the retina. The condition typically affects both eyes. Usually, there are no symptoms of early diabetic retinopathy, and the person’s sight is often not affected until the condition is severe. The ability to detect this condition in it's early stages can prevent blindness and loss of vision.

The challenge with this project was the sheer size of the data. The total dataset required a Graphics Processing Unit (GPU) infrastructure enabling the processing of images in 24 hours in order to train the model. On a regular CPU, it would have taken 2 weeks or more. The goal was to have enough data to minimize the outliers and train the computer to read the image correctly. Many data scientists don’t touch imaging because it’s layered and has a spatial aspect, it’s not flat and not labeled.  A computer reads this kind of data differently and can retain the memory of what is being viewed.  

Images are an example of unstructured data, and the use of them is a new frontier in healthcare data analytics. Precision medicine is being advanced today by a combination of advanced analytics and physician expertise. Ultimately, the benefits will be in the form of early detection and treatment, which will result in improved quality and lower costs of care for patient populations.

Where is your organization using unstructured data? What impact could data science have on healthcare that would have better outcomes for patients? Is your organization considering using unstructured data to solve a business challenge? We would love to hear your stories or questions if you want to contact us here

Looking for your next meaningful IT career opportunity?
We’ve got you covered too. Click here to set up a consult.