Diabetic Foot Automated Assessment

The project aims to develop an automated procedure to assess Diabetes-related foot disease (DF). The underlying motivations are (i) the optimization of human resources in the Healthcare domain, and (ii) simplify the access to an evaluation device to the general population.

The increase of people with Diabetes (currently 10.5% of the population has Diabetes and an estimated 11.3% will develop the condition by 2030, to an estimated 643 million, and 12.2% by 2045, [1], represents a relevant effort to societies, namely national healthcare systems, and the associated economies.

The current guidelines to Diabetic Foot (DF) assessment are composed of 6 independent tests, administered by healthcare professionals to ensure the quality/significance of the results. A routine test application to a single patient may take around 15 minutes and patients are required to take the test regularly (every 3 months). With the increased number of DF patients over the years, it becomes clear that the human resources required represent an increasing economic cost. Therefore, the solution to minimize the use of human resources is to automate the evaluation of the DF tests, this being linked to the first objective of project DFAA.

The development of devices to implement each of the tests and their integration evolves to a multidisciplinarity of scientific/engineering areas, namely three classes (i) image analysis and classification, (ii) signal processing, and (iii) design of robotic systems. In the first, novel metrics to obtain adequate models for the evolution of ulceration are expected to be developed. In the second, a comparison between pedal pulse pressure signal processed to estimate velocity and the comparison with the velocity acquired through a Doppler ultrasound sensor (the standard technology) is expected to lead to a simple yet effective technology. In the third class, low cost robotic devices will be developed to support the sensors and facilitate the use by final end-users, namely the DF patients or healthcare professionals.

The visual analysis of the ulcerations is an area of highly active research in machine learning (ML). Detecting the presence of ulcerations is a relatively simple task that can be accomplished by standard image classifiers in accordance with medical guidelines. For patients with an already installed ulceration, image analysis will be useful to estimate the area and severity of the ulceration. For patients without an explicit ulceration, the analysis of the skin condition may indicate an expected evolution towards an ulceration. The DFAA project will develop new metrics using ML to assess the severity and the probability of a skin condition to evolve to an ulceration.

Nowadays, the acquisition of pedal pulse signal is usually obtained via a Doppler sensor, yielding immediately a signal proportional to the flow of the blood stream, [2]. In the project, this sensor is replaced by a standard blood pressure meter (sphygomanometer). Hydraulics arguments are used to show that the magnitude of the blood pressure is proportional to the blood flow and hence the usual Doppler sensor can be replaced by a cheaper and easily operated blood pressure sensor.

The effectiveness of the automated procedure to administer all the tests with minimal/no intervention of healthcare professionals is strongly dependent on the integration of all the tests into one (or more) physical devices that can be safely operated by patients. This physical integration corresponds to the robotics component of the project and follows guidelines of the literature on the human factors shaping technology acceptance. The physical construction is based on accurate stepper motors and aluminium rails (3D printer technology) that allow a modular construction of prototypes.

The DFAA project will impact scientifically and technologically. At societal level, the project will be useful for different stakeholders, regarding optimization of human resources, and specially to DF patients to get a better diagnostic/prognostic.

References

[1] Mikkel Pape Dysted, Balázs Esztergályos, Sanju Gautam, Bruno Helman, Moritz Pinkepank, Adilson Randi, Agus Salim, Katherine Wallis, Beatriz Yáñez Jiménez, Margaux Ysebaert (2021). “IDF Diabetes Atlas 2021”, 10th edition. Edward J. Boyko, Dianna J. Magliano, Suvi Karuranga, Lorenzo Piemonte, Phil Riley, Pouya Saeedi, Hong Sun (Eds). International Diabetes Federation.

[2] Latifat Tunrayo Oduola-Owoo, Adekunle Ayokunle Adeyomoye, Omodele Abosede Olowoyeye, Ifedayo Adeola Odeniyi, Bukunmi Michael Idowu, Badmus Babatunde Oduola-Owoo, AdeniyiSunday Aderibigbe (2022). “Comparative Doppler Ultrasound Findings of Foot Arteries in Patients with Type 2 Diabetes Mellitus and Normoglycaemic Patients. Journal of West African College of Surgeons, 2(1):55. DOI: 10.4103/jwas.jwas 53 22

• 1 scientific research scholarship, 12 month duration,  to work, mainly, in Tasks 2, 3 and 4 of the project, acquiring the image dataset and implementing ulceration risk assessment methods.

Check the official notice here (BL 61 / 2025 - IST-ID).

João Silva Sequeira - IST-ID

Gonçalo M. Neves - IST-ID

Cláudia Marina de Oliveira Marques Neves - ULS Loures-Odivelas, E.P.E. - Hospital Beatriz Ângelo