Machine learning algorithms can help distinguish between acute cholangitis and alcohol-associated hepatitis

Acute cholangitis is a life-threatening bacterial infection that is often associated with gallstones. Symptoms include fever, jaundice, right upper quadrant pain, and elevated liver enzymes.

While these symptoms may seem distinctive and telling, they are unfortunately similar to those of a very different condition: alcohol-associated hepatitis. This presents a challenge for emergency department personnel and other healthcare professionals who must diagnose and treat patients with liver enzyme abnormalities and systemic inflammatory responses.

New research from the Mayo Clinic reveals that machine learning algorithms can help healthcare workers distinguish between the two conditions. In an article published in Mayo Clinic Proceedingsresearchers show how algorithms can be effective predictive tools using a few simple variables and routinely available structured clinical information.

This study was prompted by the fact that many medical providers in the emergency department or intensive care have difficulty distinguishing between acute cholangitis and alcohol-associated hepatitis, which are very different conditions that can occur. present in a similar way.

Joseph Ahn, MD, third-year fellow in gastroenterology and hepatology, Mayo Clinic, Rochester

Dr. Ahn is the first author of the study.

“We developed and trained machine learning algorithms to distinguish between the two conditions using some of the commonly available lab values ​​that all of these patients should have,” says Dr. Ahn. “The machine learning algorithms demonstrated excellent performance in discriminating the two conditions, with an accuracy of over 93 percent.”

Researchers analyzed the electronic health records of 459 patients over the age of 18 who were admitted to the Mayo Clinic in Rochester between January 1, 2010 and December 31, 2019. The patients were diagnosed with acute cholangitis or alcohol-associated hepatitis.

Ten routinely available laboratory values ​​were collected at the time of admission. After withdrawal of patients with incomplete data, there remained 260 patients with alcoholic hepatitis and 194 with acute cholangitis. This data was used to train eight machine learning algorithms.

The researchers also validated the results externally using a cohort of intensive care patients who were seen at Beth Israel Deaconess Medical Center in Boston between 2001 and 2012. The algorithms also outperformed doctors who participated in an online survey. , which is described in the article.

“The study highlights the potential of machine learning algorithms to aid in clinical decision-making under uncertainty,” says Dr. Ahn. “There are many cases of gastroenterologists being consulted for urgent endoscopic retrograde cholangiopancreatography in patients who initially deny a history of alcohol use but later turn out to have alcohol-associated hepatitis. In some situations , inability to obtain a reliable history of patients with altered mental status or lack of access to imaging modalities in underserved areas may require providers to make the decision based on a limited amount of objective data.”

If machine learning algorithms can be made easily accessible with an online calculator or smartphone app, they can help healthcare staff who are presented in an emergency with a critically ill patient with abnormal liver enzymes, the study finds. .

“For patients, this would lead to better diagnostic accuracy and reduce the number of additional tests or the inappropriate ordering of invasive procedures, which could delay the correct diagnosis or put patients at risk of unnecessary complications,” says Dr. Ahn. .


Journal reference:

Ah, J.C. et al. (2022) Machine learning techniques differentiate alcohol-associated hepatitis from acute cholangitis in patients with systemic inflammation and elevated liver enzymes. Mayo Clinic Proceedings.

Sharon D. Cole