A quick, noninvasive toe-based light measurement combined with artificial intelligence demonstrated efficacy in screening for peripheral artery disease (PAD), according to a recent study published in npj Digital Medicine by a 2024 recipient of an ACC Accreditation Services' QI.PI Project Grant. The novel technique, developed and validated by a multidisciplinary team of researchers at University of California San Diego, shows promise in improving access to diagnosis and helping identify high-risk patients sooner.

Mattheus Ramsis, MD; Ava J. Fascetti, MS; et al., included 2,362 patients in their study exploring whether photoplethysmography (PPG) paired with a machine learning model could detect PAD, noting significant correlations between multiple PPG features and ankle-brachial index (ABI), a commonly used noninvasive test to diagnose PAD.

The investigators report that the machine learning model for PPG-based PAD detection performed well, with an area under the receiver operating characteristic curve (AUC) of 0.831. Adding the smoking status of the patient improved the model's performance, marked by an AUC of 0.845.

"When we built the model using only the PPG data, it demonstrated strong performance in distinguishing patients with PAD (defined by an abnormal ABI) from those without the disease, correctly distinguishing PAD cases approximately 83% of the time, compared with the roughly 60-65% performance typically achieved using traditional clinical risk factor assessment alone," said Ramsis. "Importantly, the signal reflects physiologic blood flow changes in the toes, providing information beyond conventional clinical evaluation."

In addition, the authors note no statistically significant differences in model performance across racial or ethnic groups, and its consistent performance among key clinical subgroups such as patients with end-stage renal disease and diabetes.

"The present findings are not intended to represent the final stage of validation, but rather to establish the physiologic and methodological foundation required for prospective evaluation," write the authors. "Based on the robustness of short-duration signal acquisition, feature interpretability and performance consistency observed in this study, this framework is now the foundation for prospective studies designed to assess performance across clinical workflows and complementary reference standards."