By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Cookie Policy for more information.
The precision of data and results with AI: a faster and more accurate way of detecting PAD
May 23, 2023

Peripheral Arterial Disease (PAD) is the new pandemic, and although it is highly spread among people, it is not given enough attention. The figures are not promising:

What's more, the thing that makes the condition so cruel is the fact that its diagnosis can be challenging. According to the same source, about 20%-50% of people with PAD are asymptomatic. And diagnosing on time is vital for its treatment.

Importance of diagnosing it on time

According to the Compendium of PAD from 2015, many people are dying from PAD because of neglect. Despite being a major atherosclerotic disease with a global prevalence of >200 million, healthcare specialists often fail to diagnose PAD, due in part to conflicting screening recommendations, low patient and provider awareness, and the high prevalence of asymptomatic or atypical symptoms. Therefore, detecting the disease on time can be absolutely treatable, and the condition of the patients could possibly be reversed.

Traditional methods of diagnosing it

Ankle-brachial index (ABI)

The most popular method of diagnosing PAD is an ankle-brachial index (ABI), which measures the blood pressure in the ankles and compares it with the blood pressure of the arm. ABI is a cheap and non-invasive method, but unfortunately, numerous pieces of research show the results from ABI are controversial. For example, due to research, there is little evidence on the value of the ankle-brachial index (ABI) for detection of lower limb peripheral arterial disease (PAD) in patients with exertional leg pain. Also, it is often written that the ABI is not a useful test for detecting PAD in those with diabetes (Bhasin 2007; MacLeod‐Roberts 1995) because the incompressibility of calcified vessels produces false results.

Non-contrast magnetic resonance (NC-MRA)

There are several NC-MRA techniques used for the diagnosis of disease in various vascular districts. All of them have their advantages and disadvantages, but the main problem is the poor visualization of some areas of the body and the longer scanning times. For example, one of the most used techniques for detecting PAD is called QUISS - it has acceptable scanning times (7-8 minutes) but even with this method one can experience a suppression of in-plane oriented vessel segments that reside within the saturation region.

Duplex ultrasound (DUS)

Duplex ultrasonography is also very often used to screen, diagnose, and monitor patients with PAD. It accurately determines the location and degree of stenosis in arteries and helps differentiate stenosis from an occlusion. (Nasra K, Osher M. StatPearls [Internet]. StatPearls Publishing; Treasure Island (FL): Jan 8, 2023. Sonography Vascular Peripheral Arterial Assessment, Protocols, And Interpretation. [PubMed]) However, even though most arterial segments of the lower limb may be visualized adequately using DUS, some segments are difficult to visualize and the results are not fully reliable. Another limitation is the need for highly experienced technicians - especially for the visualization of the intravesicular arteries.

What’s more, ultrasound is often not suitable for people with diabetes as they suffer from Medial Arterial Calcification (MAC) - a non – obstructive condition leading to reduced arterial compliance. For this reason, known testing methods for cardiovascular diseases are often unreliable for diabetics.

X-ray angiography

This is currently the most widespread and reliable method for detecting PAD, and although it is precise, it has some cons. The first and obvious one is repeated exposure to X-rays, which can lead to a number of health complications for patients. Besides that, the procedure is costly and time-consuming - a procedure takes from one to three hours and sometimes that means a hospitalization of the patient is needed.

Infrared thermography

By detecting temperature changes in the lower extremities, infrared thermography offers another non-invasive option in PAD diagnosis. Different medical publications show it does have the potential to provide additional information about circulation, subclinical infections, and the severity of vascular disease. Also, its accuracy is proven to be around 90%. Its main benefit undeniably is the ability to immediately measure the absolute temperature of the skin surface over a large area without direct contact with the skin. Another study also showed portable IRT devices showed high sensitivity and specificity as a screening tool for lower limb arterial disease compared to arterial Doppler ultrasonography. A publication goes further by suggesting the IRT method even has more advantages than ABI; it is said to be a safe, reliable and simple application, that could be a worthy tool for the assessment of the clinical frame and severity of foot blood perfusion in symptomatic PAD patients.

Kelvin Health Solution - a game-changer in healthcare

Kelvin Health tool is based on the thermography procedure as it uses a mobile thermal imaging camera that captures the body’s heat, segments the thermal image, and applies artificial intelligence to detect anomalies related to vascular conditions. The model uses machine learning algorithms such as deep neural networks to detect anomalies which allow robust learning of typical PAD pathology.

Research and data collection so far show that Kelvin Health can also be used for supporting information for every patient without the risk and the blind spots of other approaches. The main advantages of the tool are its rapidity, non-invasiveness, and precision of data. So far current methods are often very slow and ironically, not reliable enough as different researches illustrate, for example in people with diabetes (due to their typical medial arterial calcification).


Subscribe for latest News & Updates from Kelvin Health in your email

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.