Rare but dangerous cholesterol condition gives it up to an algorithm

A rare and difficult-to-diagnose genetic condition that raises LDL (bad) cholesterol to dangerous levels is now vulnerable to an AI tool.

The disease is called familial hypercholesterolemia (FH). It’s often mistaken for ordinary elevated cholesterol, but it makes heart attack as much as 20 times more likely in the small slice of the population (around 0.4%) that has it.

The tool is an algorithm developed by researchers at Stanford University and tested by colleagues working with FH patients (and controls) at Geisinger Health System in Pennsylvania.

Senior author Joshua Knowles, MD, PhD, of Stanford and co-authors had the research published online April 11 in NPJ Digital Medicine.

The team trained a random forest classifier to distinguish the presence of FH from its lack using data from Stanford Medicine’s EHR. The data comprised records from 197 patients with known FH and more than 6,500 unaffected matches.

On an initial test set, the classifier achieved both high positive predictive value and high sensitivity.

Further, when applied to 100 patients at risk of FH who were not part of the original dataset, the tool correctly flagged 84% of patients at the highest probability threshold.  

When the researchers went to validate their classifier on 466 FH patients and 5,000 matched controls at Geisinger, they found its accuracy clocked in at 85%.

“Our EHR-derived FH classifier is effective in finding candidate patients for further FH screening,” the authors concluded. “Such machine learning guided strategies can lead to effective identification of the highest risk patients for enhanced management strategies.”

Study co-author Nigam Shah, MBBS, PhD, also of Stanford, explained the main problem the project set out to address.

Because FH is so rare, he told the institution’s news division, it doesn’t make sense to screen all heart patients for it on a general scale.

“Theoretically, when someone comes into the clinic with high cholesterol or heart disease, we would run this algorithm,” Shah said. “If they’re flagged, it means there’s an 80% chance that they have FH. Those few individuals could then get sequenced to confirm the diagnosis and could start an LDL-lowering treatment right away.”

Dave Pearson

Dave P. has worked in journalism, marketing and public relations for more than 30 years, frequently concentrating on hospitals, healthcare technology and Catholic communications. He has also specialized in fundraising communications, ghostwriting for CEOs of local, national and global charities, nonprofits and foundations.

Trimed Popup
Trimed Popup