Connect with us

ECG Equipment

Influencing human over readers with automated ECG interpretation

The role of automatic ECG analysis in clinical practice is limited by the accuracy of existing models. This technology has recently achieved striking success in a variety of tasks, and there are great expectations on how it might improve clinical practice.

According to the World Health Organization, cardiovascular diseases (CVDs) account for an estimated 31 percent of all deaths worldwide. Since their invention over a century ago, ECGs have been instrumental in the diagnosis and monitoring of various CVDs, and innovations have produced diagnostic ECGs that deliver more accurate results in less time, helping doctors make timely treatment decisions that ultimately benefit more patients.

Innovation in healthcare technology, particularly around diagnostic ECG, continue to change how CVD is diagnosed, monitored, and treated. Global demand is also driving growth in the ECG market, which is expected to maintain a compound annual growth rate of 5.6 percent and reach USD 6637 million by 2023, according to Allied Market Research.

Patients are also becoming more empowered to monitor and track their own health, and as products and services are increasingly integrated with cloud interfaces and electronic health record (EHR) systems, diagnostic ECG technology is evolving to help them do so.

Innovations have increased diagnosis accuracy and speed and have made devices smaller, easier to use, and more comfortable to wear – large, bulky electrocardiogram devices have become pen-sized; cumbersome Holter devices with electrodes and wires have transformed into simple, disposable patches.

The Indian ECG equipment market in 2020 is estimated at ₹150.5 crore, at 27,060 units. The market had slowed down in 2Q2020, with the hospitals barely functioning as a result of the nation-wide lockdown, but demand started picking up from 3Q onward, and in 2H2020, the market grew in double digits.

In value terms, the high-end machines account for a 72-percent share, and a 49-percent share by volume. The demand for single-channel machine is declining ever year, as are the margins. So is the 6-channel segment, though for a different reason. Its application is niche, and the reducing price differential between 6-channel and 12-channel is tilting the buyer toward the superior product.

The 12-channel segment is seeing steady growth and this trend shall continue. By volume, the segment has an 89-percent share within the high-end systems and a 44-percent in the overall ECG market. By value, the shares are 71 percent and 52 percent respectively.

In the high-end segment, the Holter machines are gaining popularity and strengthening their position. The ratio between stress/Holter/ambulatory and resting systems in 2020 is estimated at 85:15.

Major vendors* in Indian ECG equipment market – 2020

Tier I BPL and GE
Tier II Schiller, Philips, and Contec
Tier III Mindray and Edan
Tier IV Nidek, Bionet, Mortara, Allengers, and RMS
Others Skanray, Nasan, Medikit, Forest, Silverline Meditech, Nihon Kohden, and local brands
*Vendors are placed in different tiers on the basis of their sales contribution to the overall revenues of the Indian ECG equipment market.
ADI Media Research

2020 saw BPL gain ground. The brand came neck-to-neck with GE, and now GE shares the Tier-I position with BPL. It has increased its national distribution network to 110 dealers, and has the advantage of dominating the single-channel and 3-channel segments.

The unorganized and local brands continue to hold sway in this segment, and have increased their share from about 20 percent in 2019 to 34 percent in 2020.

Like many facets of healthcare, artificial intelligence (AI) is being developed to aid ECG interpretation. One of the top priorities for the application of AI to the interpretation of ECGs is the creation of comprehensive, human-like interpretation capability. Since the advent of the digital ECG more than 60 years ago, ongoing effort has been toward rapid, high-quality, and comprehensive computer-generated interpretation of the ECG. The problem seems tractable; after all, ECG interpretation is a fairly circumscribed application of pattern recognition to a finite dataset. Early programs for the interpretation of digital ECGs could easily recognize fiducial points, make discrete measurements, and define common quantifiable abnormalities. Modern technologies have moved beyond these rule-based approaches to recognize patterns in massive quantities of labelled ECG data.

Several groups have worked to create AI-driven algorithms, and some of these algorithms are already in limited clinical use. Some studies have developed convolutional neural networks (CNNs) from large datasets of single-lead ECGs and then applied them to the 12-lead ECG. For instance, using 2 million labelled single-lead ECG traces collected in the Clinical Outcomes in Digital Electrocardiology study, one group used a CNN to identify six types of abnormalities on the 12-lead ECG9. This study demonstrated the feasibility of this approach, but widespread implementation or external validation in other 12-lead ECG datasets is forthcoming. Another group conducted a similar study of the application of CNNs to single-lead ECGs and demonstrated that the CNN could outperform practicing cardiologists for some diagnoses. However, whether this approach will translate to clinically useful software for 12-lead ECG interpretation remains to be seen.

In an evaluation published in 2020, a CNN was developed for the multilabel diagnosis of 21 distinct heart rhythms based on the 12-lead ECG using a training and validation dataset of >80,000 ECGs from >70,000 patients. The experts demonstrated that a CNN can identify 66 discrete codes or diagnosis labels, with favorable diagnostic performance. Lately, experts have developed a novel method that uses a CNN to extract ECG features, and a transformer network to translate ECG features into ECG codes and text strings. This process creates a model output that more closely resembles that of a human ECG reader – presenting information in a similar order, with similar language – and also makes sense of associated codes, thereby avoiding the presentation of contradictory or mutually exclusive interpretations that would not be presented by a human reader.

This technology will be particularly important as physicians increasingly rely on ECG data, obtained through novel, consumer-facing applications, which are massively scalable. For instance, AI–ECG algorithms have been applied to single-lead ECG traces, obtained through mobile, smartwatch-enabled recordings for the detection of AF13. This democratization of ECG technology will exponentially increase the volume of signals that demand interpretation, which might quickly outstrip the capacity of human ECG readers. It has been anticipated that these models will be essential in facilitating telehealth technologies (that is, automatic, patient-facing or consumer-facing technologies) and could allow the creation of core laboratory facilities, capable of ingesting and processing massive quantities of data. However, the signal quality obtained with these devices can be inconsistent, and AI–ECG might be less able than human expert over-readers to classify the heart rhythm, using poor-quality tracings.

Nevertheless, although great progress has been made toward a comprehensive, human-like ECG-interpretation package, the realization remains on the horizon. Even in its most modern incarnation, the package lacks the accuracy needed for implementation without human oversight. Additionally, computer-derived ECG interpretation has the potential to influence human over-readers and, if inaccurate, can serve as a source of bias or systematic error. This concern is particularly relevant if the algorithms are derived in populations that are distinct from those in which the algorithms are applied. This limitation underscores the need for a diverse derivation sample (ideally including a diverse patient population and varied means of data collection that reflect real-world practices), rigorous external validation studies, and phased implementation with ongoing assessment of model performance and effectiveness. Quality-control systems, based on regular over-reads by expert interpreters of ECGs who are blinded to the output of the CNN, will allow continuous calibration of the model.

error: Content is protected !!