Artificial Intelligence Can Detect Pneumonia-Causing Bacteria

Artificial intelligence can use information available in the emergency room to predict the kind of bacteria that is causing infection in patients with pneumonia, according to research presented at the annual meeting of the American Society of Microbiology.

Researchers stated that infection caused by antibiotic-resistant bacteria is hard to treat and can be life-threatening. Pneumonia caused by bacteria such as Methicillin-resistant Staphylococcus Aureus (MRSA) or pseudomonas can be fatal because they are resistant to commonly prescribed antibiotics.

Although there are effective antibiotics against these infections, the sputum culture test takes at least 48 hours to incubate and identify those bacteria from the sputum, while these patients may deteriorate within hours.

Researchers analyzed more than 50,000 intensive care unit admissions data from Beth Israel Deaconess Medical Center (BIDMC) in Boston, Massachusetts. The team examined records of patients who were admitted with pneumonia and trained an AI neural network using the dataset. The AI agent showed promising results in predicting bacteria that caused the infection.

The algorithm was able to use information available in the emergency room and predict if the patient has MRSA or pseudomonas so that physicians can immediately prescribe specific antibiotics targeting specific bacteria.

“This research highlights the potential of AI as a supplementary tool for physicians in identifying causal pathogens of pneumonia, even before sputum culture results are available,” said Joowhan Sung, MD, hospitalist at MedStar Southern Maryland Hospital. “We demonstrated that physicians could be assisted by AI to decide appropriate antibiotics.”

Researchers noted that the results also have important implications for patients with COVID-19.

“Similar techniques can be applied to future research on pneumonia amid the current pandemic, such as capturing bacterial co-infection in those with known COVID-19, which could be fatal if undetected,” said Sung.

Artificial intelligence has played a significant role in the COVID-19 pandemic. A recent report from Johns Hopkins Medicine showed that AI has the potential to enhance the role of chest imaging and leverage large-scale data to quickly find solutions for detecting, containing, and treating COVID-19.

The team argued that AI has the power to enhance chest imaging beyond just screening for signs of COVID-19 in a patient’s lungs.

“Although chest imaging is not presently recommended for initial diagnosis of COVID-19 pneumonia, the scale of the COVID-19 pandemic calls for dedicated research on diagnostic and therapeutic approaches that are robust, reliable, and rapid. Data-driven AI applications could address this unmet need to allocate resources in a timely manner,” the team stated.

AI has also helped investigators identify the most pertinent research on COVID-19. A team from Northwestern University has developed an AI platform that can quickly detect research that has the most potential to produce COVID-19 treatments and solutions.

The AI model is just as accurate as the human scoring system at making these predictions, and it can scale up to review a larger number of papers in minutes instead of months.

“The standard process is too expensive, both financially and in terms of opportunity costs,” said Brian Uzzi, Richard L. Thomas Professor of Leadership at Kellogg and co-director of the Northwestern Institute on Complex Systems.

“First, it takes take too long to move on to the second phase of testing and second, when experts are spending their time reviewing other people’s work, it means they are not in the lab conducting their own research.”