Donate to Science & Enterprise

S&E on Mastodon

S&E on LinkedIn

S&E on Flipboard

Please share Science & Enterprise

NIH Grant Funds A.I.-Based Tuberculosis Detector

Human lungs illustration


27 Nov. 2019. A company developing a system to quickly detect tuberculosis by analyzing sputum images received a federal grant to advance and validate its technology. Diascopic LLC in Cleveland, Ohio is the recipient of a six-month, $225,000 award from National Institute of Biomedical Imaging and Bioengineering, part of National Institutes of Health.

Tuberculosis is caused by the Mycobacterium tuberculosis microbe and affects mainly the lungs. The disease is marked by a bad cough lasting for weeks at a time, chest pains, and coughing up blood and sputum — a combination of saliva and mucous — from deep in the lungs. Many people may have a latent form of tuberculosis that does not cause symptoms, nor is spread to others, but in people with a weakened immune system can become the full-fledged disease.

According to World Health Organization, 10 million people worldwide became ill from tuberculosis in 2018, leading to 1.5 million deaths. While tuberculosis is preventable and treatable, the disease is often a complication of people with HIV, responsible for 251,000 deaths in 2018. Tuberculosis, says WHO, is also often overlooked and difficult to diagnose in children and adolescents. The agency says eight countries accounted for two-thirds of tuberculosis cases in 2018: India, China, Indonesia, the Philippines, Pakistan, Nigeria, Bangladesh and South Africa.

Diascopic is a spin-off enterprise from Case Western Reserve University in Cleveland developing the iON system for detecting tuberculosis from a patient’s sputum sample spread on a microscope slide. The iON software analyzes the sputum sample microscope image to detect a Mycobacterium tuberculosis infection. Diascopic says its system today can detect tuberculosis infections in 60 seconds with 95 percent accuracy of visual microscope inspections.

With the NIH grant, Diascopic aims to boost iON’s analytical engine. The software now uses an algorithm developed by the company, but in the new project will add in more deep-learning algorithms with a convolutional neural network. A convolutional neural network combines image analysis and machine learning to dissect an image by layers for understanding features in the image. Different aspects of each layer discovered and analyzed by the algorithm are translated into data that the algorithm then uses to train its understanding of the problem being solved, with that understanding enhanced and refined as more images and data are encountered.

The added analytical power is expected to raise iON’s accuracy rate to 99 percent compared to visual inspection. Diascopic also aims to meet WHO’s target product profile for a point-of-care sputum-based test, issued in 2014.

“By digitizing the process, we’re able to reuse the image in perpetuity, which makes the test highly repeatable,” says Diascopic founder and CEO Cary Serif in a company statement released through EurekAlert. Serif adds that “the digitization allows us to build a massive reference library to which we can apply artificial intelligence and data analytics to continue improving the test’s accuracy.”

The company is collaborating with Anant Madabhushi, professor of biomedical engineering at Case Western Reserve and director of the university’s Center for Computational Imaging and Personalized Diagnostics. Diascopic is also joining the university’s research collaboration with Makerere University and other institutions in Uganda to field test the upgraded iON system, analyzing 400 sputum samples and generating some 60,000 digital images.

Diascopic says the iON technology can test for more diseases than tuberculosis. If the current system upgrade is successful, the company plans to expand the platform to diagnose other infectious diseases in developing regions such as river blindness, malaria, and schistosomaisis.

The early-stage award to Diascopic is made under NIH’s Small Business Innovation Research, or SBIR, program that sets aside a portion of its overall research funding for small U.S.-based companies with science-based products. NIH says it invests more than $1 billion in SBIR and related Small Business Technology Transfer awards.

More from Science & Enterprise:

*     *     *

Comments are closed.