How AI and Smartphones May Help Fight Antibiotic Resistance

Source: quimono/Pixabay

By 2050, the number of human deaths resulting from antimicrobial resistant (AMR) infections is projected to increase to 10 million per annum unless there is a global response to address the problem, according to the “Review on Antimicrobial Resistance” report by the UK Department of Health with Wellcome Trust. Antimicrobial resistance is an existential threat that not only impacts global human health, but also food supplies, consumer packaged goods, livestock, and tourism. The contributing factors to AMR include the use of antibiotics in animal husbandry, the proliferation of antibacterial agents in consumer products, and the over-prescribing and misuse of antibiotics in health care. Recently on May 7, 2019, the Doctors Without Borders / Médecins Sans Frontières (MSF) Foundation was awarded a $1.3 million Google Artificial Intelligence Impact Challenge grant to create a smartphone app that will use artificial intelligence (AI) to analyze antibiotic resistance tests and recommend treatment in low-resource environments.

The MSF Foundation is an incubator for innovation, research, and development. It is part of Doctors Without Borders / Médecins Sans Frontières—a humanitarian organization that was awarded the Nobel Peace Prize in 1999 for its work spanning several continents to provide emergency medical aid to people impacted by epidemics, natural disasters, armed conflict, and exclusion from health care. It was one of 20 organizations selected from 2,602 applications from 119 countries to share $25 million in grants from the Google Artificial Intelligence Impact Challenge. The grant is part of Google’s “AI for Social Good” program under the leadership of Jacquelline Fuller, President of Google.org, and Jeff Dean, Google Senior Fellow and head of Google AI.

To directly address the issue of medical overuse of antibiotics, the MSF Foundation plans to create a smartphone app called “ASTapp” that uses AI to analyze antibiogram images to guide non-expert microbiologists in interpreting antibiotic resistance tests.

Testing antimicrobial resistance and efficacy of antibiotics is done by antibiograms, where an antibiotic is seeded in a petri dish that has been cultured and populated with a particular microorganism. Bactericidal activity is occurring if a clear zone emerges—the larger the zone, the greater the antibiotic efficacy.

Antibiograms are typically analyzed by microbiologists. However, in low-resource environments, there is often a lack of access to microbiologists to perform the interpretations. To solve this problem, the MSF Foundation team plans to use artificial intelligence to both analyze antibiogram images and recommend treatment using a smartphone app.

Image processing by computer vision is a growing trend that is part of the AI boom in recent years with investments by venture capitalists, corporations, and other organizations in life sciences, biotech, pharmaceuticals, neuroscience, and health care. This is mostly due to vastly improved pattern recognition precision and capabilities due to deep learning, a subset of AI machine learning with architecture inspired by the biological brain consisting of artificial neural networks with more than two layers of processing.

According to a MSF Foundation report, the team’s goal is to have the ASTapp ready for launch in three years, with an estimated four months to develop the app, eight to 12 months planned for quality assurance and testing, and the remaining time allotted for global training and rollout.

The team plans to have ASTapp ready for testing as early as the last quarter of 2019 in MSF field-based laboratories, starting with the MSF lab in Amman, Jordan. Antibacterial resistance have been found “everywhere MSF has tested for it, but we have identified the issue as especially prevalent in the Middle East, where we treat severe wounds and encounter the wide availability of over-the-counter antibiotics,” reported the MSF Foundation in the same report.

Innovative technology is the proverbial double-edged sword—it can be used to help solve some of the greatest challenges facing humankind, or it can become an existential threat—depending on how and where it is applied. Artificial intelligence sits squarely in this familiar adage. Fortunately, forward-thinking organizations recognize the need to ensure that AI is used to help improve the human condition, and are putting resources and funding towards AI projects to benefit humanity.

Copyright © 2019 Cami Rosso All rights reserved.