Wrist-worn sensor predicts elevated hs-cTnI in ACS patients

A novel transdermal infrared spectrophotometric sensor (transdermal-ISS) worn on the wrist may be used for quick and noninvasive prediction of elevated high-sensitivity cardiac Troponin-I (hs-cTnI) in real world settings, according to a study presented at ACC 2023.

“The use of a transdermal-ISS for bloodless estimation of hs-cTnI shows the feasibility and may have a role in real-world settings for diagnosing acute myocardial infarction (AMI) in patients presenting with acute coronary syndrome (ACS),” said the researchers led by Dr Shantanu Sengupta from Sengupta Hospital and Research Institute in Nagpur, India.

Sengupta and colleagues tested the clinical feasibility of a wrist-worn transdermal-ISS in clinical practice and evaluated the performance of a machine-learning algorithm for identifying elevated hs-cTnI levels in hospitalized patients with ACS.

The proposed algorithm consisted of two components: one for signal processing, transmission, and feature extraction from a wrist-worn wearable device and a deep-learning model to classify samples as clinically normal or abnormal based on a clinically recommended cut-off value of troponin-I.

The research team enrolled a total of 238 ACS patients at five sites and verified the final diagnosis of MI (with or without ST-elevation) and unstable angina using electrocardiogram (ECG), cardiac troponin (cTn), echocardiography (regional wall motion abnormality), or coronary angiography. They then trained the transdermal-ISS-derived deep-learning model at three sites before externally validating it with hs-cTnI (one site) and echocardiography and angiography (two sites), respectively.

Prediction of elevated hs-cTnI levels was successfully made by the transdermal-ISS model, with the area under the receiver operator characteristics of 0.90 (95 percent confidence interval [CI], 0.84‒0.94; sensitivity, 0.86; specificity, 0.82) for the internal validation cohort and 0.92 (95 percent CI, 0.80‒0.98; sensitivity, 0.94; specificity, 0.64) for the external validation cohort. [ACC 2023, abstract 409-12]

“The recording of transdermal-ISS waveforms and the transmission of data over the web occurred accurately and confirms the feasibility of obtaining the optical-sensor readings in routine clinical settings,” the researchers said.

“The machine-learning models showed stability and high accuracy for detecting elevated troponin-I which differentiated patients with MI from unstable angina and other causes,” they added.

Model prediction

Additionally, the predictions of the model correlated with the presence of regional wall motion abnormalities (odds ratio [OR], 3.37, 95 percent CI, 1.02‒11.15; p=0.046) and significant coronary stenosis (OR, 4.69, 95 percent CI, 1.27‒17.26; p=0.019).

“These data should inform the design of future clinical trials where optical sensor-based machine-learning models can be further optimized for detecting myocardial injury and understanding potential underlying pathophysiology for each phenotype of troponin elevation,” the researchers said.

CTn has been used as a biomarker to assess myocardial injury since the 1990s. Although accurate, immunoassays require time-sensitive logistics coordination between the lab and the ordering providers for blood draws and sample transport, according to the researchers.

“While new strategies in using high-sensitivity immunoassays and microneedle patches are worthy of continued consideration, the alternate strategy evaluated in this study using mid-infrared has been used previously for analysis of serum, urine, breath, skin, and other bodily fluids,” they noted. [Analyst 2009;134:1092-1098; Analyst 2018;143:3156-3163; Molecules (Basel, Switzerland) 2020;25:2227]

“Future larger studies and pragmatic clinical trials would need to investigate the impact of transdermal-ISS on early, prehospital infarct diagnosis, triage, and therapy in emergency settings, including its utilization in emergency rooms, chest pain clinics, and implementation in ambulances and its utilization by trained paramedics,” the researchers said.