A new device improves the quality of ultrasound scans
Cristina G. Pedraz/DICYT An ultrasound scan is a test used for diagnosis and medical control that is based on ultrasounds (US). In other words it uses sound waves to create an image. It is widely used in clinics as it is a non invasive, quick, easy, and economic means of obtaining relevant medical information.
Over the past decade researchers from the Image Processing Laboratory (LPI) of the University of Valladolid have been working to improve the quality of these ultrasound scan images. In order to do so they set up the Selene project, by means of which they have developed a method and a device capable of “refining” the quality of the images provided by ultrasound scanners.
The study focused on the doctoral thesis of the researcher Gonzalo Vegas Sánchez-Ferrero, who is currently an Advanced Fellow of the M+VISION Consortium between the Autonomous Region of Madrid and the Massachusetts Institute of Technology (MIT). He works at the Brigham and Women's Hospital of Boston as part of the Applied Chest Imaging Laboratory (ACIL) Group of the Harvard Medical School and in collaboration with Biomedical Image Technologies (BIT) of the Universidad Politécnica de Madrid.
As he explains, the ultrasound (US) scan “is probably the most widely used diagnostic image in a medical context, mainly because of its low cost and non radiant nature”. However, it is particularly noisy and produces images with a grainy pattern known as speckle that limits its resolution.
“Researchers into ultrasound medical images are fully aware of the undesirable effects of the noise and have been working on its characterisation and its reduction for several decades. Important progress has been made in this field and numerous noise reduction filters have been developed with varying degrees of success, but unfortunately medical protocols do not yet include the application of these filters”, he explains.
The main reason is that the speckle is a pattern produced by the physics itself of the acquisition of ultrasounds. Among other factors it depends on the characteristics of the tissue that is being studied, in such a way that a suitable analysis of the speckle “can provide very valuable information for diagnostic purposes”.
Medical personnel therefore prefer to retain the grainy texture despite the fact that it hinders the anatomical and functional analysis of ultrasound scans, but at the same time they agree that it is highly desirable to improve the image quality in order to make the diagnostic process easier. “This apparent paradox shows the difficulties that exist in communication between the technological and medical fields, as although researchers into images have provided noise reduction filters for ultrasounds doctors are not using them for fear of losing relevant diagnostic information”.
Interaction with the image in real time
The Selene project is trying to solve this problem by proposing a device and a method for filtering and enhancing ultrasound images that preserve diagnostic information. On the one hand the system allows the characterisation of tissues of diagnostic interest that can be selected by the user. “The tissues of interest can be selected automatically (by default) or manually as required, giving the doctor the possibility of interacting with the image in real time. This selection objective gives the method great versatility according to the diagnostic purpose of the image”, the researcher explains. On the other hand it is capable of selective enhancing with the preservation of diagnostic details for 2D, 3D, and 4D images in real time, the aim of which is to facilitate the inclusion of Selene in daily clinical practice.
The device is currently at the stage of having requested an international patent. The team of researchers is developing the first commercial prototype and designing a clinical test to demonstrate the diagnostic improvement that Selene provides.
For the moment the system has been tested in a Valladolid clinic and the preliminary results have been satisfactory. “Quantitative studies have shown that in intravascular ultrasound images Selene provides improved results in terms of contrast and the preservation of diagnostic information”, the researcher points out.
Several prizes and recognition
The Selene project has been awarded the 2014 Prometeo Prize by the Fundación General de la Universidad de Valladolid (FUNGE) for the development of market-orientated prototypes, together with the first 2014 University Business Promoters Vivero Prize, an award promoted by the Regional Ministry of Education through the Castilla y León University and Higher Education Foundation (Fundación Universidades y Enseñanzas Superiores de Castilla y León, Fuescyl) with the sponsorship of the División Global Santander Universidades. Likewise the doctoral thesis of Gonzalo Vegas Sánchez-Ferrero was awarded the Extraordinary Doctorate Prize of the University of Valladolid in 2013 and was also short-listed for the best Doctoral Thesis in Information and Telecommunications Technologies of the Professional Association of Telecommunication Engineers in the same year.
Ramos-Llordén, G., Vegas-Sánchez-Ferrero, G., Martín-Fernández, M., Aja-Fernández, S., Alberola-López, C. (2015) "Anisotropic diffusion filter with memory based on speckle statistics for medical images". IEEE Transactions on Image Processing. Vol. 24, n 1