Key Points
– Prompt identification of patients suffering from ACS requires measurements of the blood level of the cardiac troponin biomarker, which is traditionally dependent on laboratory turn-around times.
– A novel wrist-worn, point-of-care test for estimating high-sensitivity cardiac troponin-I (hs-cTnI) based on transdermal infra-red spectrophotometric sensors, and coupled with a machine-learning algorithm, was trained and externally validated on a cohort of 238 patients presenting with ACS across 5 hospitals in India.
– The externally validated model demonstrated an excellent performance with AUC of 0.92 (95% CI, 0.80-0.98; sensitivity, 0.94; specificity, 0.64), and also predicted obstructive coronary disease and regional wall motion abnormalities. As such, this innovative wearable technology may have important implications in real-world settings for diagnosing patients presenting with ACS inside and outside of a hospital.
Promptly diagnosing acute coronary syndrome (ACS) and differentiating it from unstable angina or unrelated non-cardiac mimickers is critical to implement time-sensitive interventions that can save lives, such as percutaneous coronary intervention. One of the fundamental steps for diagnosing ACS is measurement of blood concentration of the troponin biomarker, which is traditionally dependent on laboratory turn-around times. As such, the implementation of a hand-held, point-of-care test for estimating high-sensitivity cardiac troponin-I (hs-cTnI) can revolutionize diagnosis and treatment options for patients with ACS.
During the annual American College of Cardiology (ACC) scientific session, held this year in partnership with the World College of Cardiology (WCC) on March 3-6th in New Orleans, LA, Dr. Partho Sengupta (Rutgers Robert Wood Johnson Medical School, New Jersey) presented a late-breaking study on the clinical feasibility of a wrist-worn transdermal infra-red spectrophotometric sensor (Transdermal-ISS), coupled with a remotely-analyzable machine-learning algorithm for identifying elevated hs-cTnI levels among patients hospitalized with ACS.
The study enrolled 238 ACS patients in five hospitals in India. From March to November 2022, the study team performed a head-to-head comparison of traditional blood-drawn hs-cTnI levels with the Transdermal-ISS recordings at a single random time point during their hospital stay, where the blood draw occurred within 15 minutes of using the optical device. Of these patients, the development cohort comprised data from three hospitals, while the external validation cohort comprised data from 2 separate hospitals. Patients with tattoos, scars or other skin lesions who may interfere with transdermal devices were excluded from the study. The primary endpoint of training the machine learning model was the presence of elevated 7 troponin, as defined as a binary endpoint based upon central lab analysis (elevated versus 8 normal).
The mean age of the study cohort was 55 years, 23% female sex, of which 57% presenting with STEMI, 22% with NSTEMI and 21% with unstable angina or other causes. The trained model (n=134, after excluding patients with unmeasurable hs-cTnI levels by blood draw) showed excellent performance with an area under the curve (AUC) of 0.90 (95% confidence interval [CI] 0.84-0.94; sensitivity, 0.86; specificity, 0.82), and the externally validated model (n=45) demonstrated a similar performance with AUC of 0.92 (95% CI, 0.80-0.98; sensitivity, 0.94; specificity, 0.64). No clinical covariates were found to influence the predictive performance of the Transdermal-ISS. In secondary analyses, adjusted for sex and smoking, abnormal Transdermal-ISS was significantly associated with the presence of obstructive CAD (OR, 4.69 [1.27- 17.26], p = 0.019) as well as the presence of regional wall motion abnormalities on echocardiography (OR 3.37, 95% CI [1.02 to 12 11.15], p =0.046).
In discussing next steps for this promising technology, Dr. Sengupta explained that future studies with larger sample size should address the diagnostic performance of the algorithm among symptomatic patients with a larger prevalence of normal hs-cTnI. Furthermore, as the diagnosis of ACS requires a dynamics rise and/or fall of troponin, therefore future studies should test the performance of Transdermal-ISS with serial measurements.
In conclusion, the use of a wrist-worn, point-of-care transdermal-ISS , coupled with a remotely-analyzable machine-learning algorithm, demonstrated the feasibility of bloodless estimation of hs-cTnI in a rapid and reliable fashion, and may have important implications in real-world settings for diagnosing patients presenting with ACS inside and outside of a hospital.
