Diagnosis and Classification of Pleural Diseases Using Ultrasound Channel Data

Background and rationale:

The Pleura is a double-layered membrane surrounding the lungs. It is anatomically divided into Visceral Pleura, which is attached to the outer part of the lungs, and to the Parietal Pleura, which is attached to the inner side of the rib cage. Pathological processes involving the Pleura are incorporated under the name "Pleural diseases" which include, among others, pneumothorax, pleural effusion, and various tumors. This disease group is relatively widespread, and as evidence that the incidence reported in the literature is over 300 to 100,000 people in one year.

Diagnosis and evaluation of pleural disease involve the use of various neuroimaging devices such as chest radiographs, ultrasound, CT, and MRI. Most often, chest radiographs are widely accepted as the first means of identifying pleural diseases, with ultrasound, CT, and MRI supplementing or constituting a diagnostic confirmation if necessary.

Over the past few years, a significant perceptual change has been made in the use of chest ultrasound to diagnose pleural diseases as part of emergent and none emergent situations. Ultrasound has been shown to be an excellent imaging device for the assessment of pleural diseases. It is used as an auxiliary in a variety of clinical situations, such as pneumothorax detection in emergency situations, characterization of pleural fluid as simple or complex, the distinction between pleura thickening and pleural fluid loculations, and more. In addition to the diagnostic aspect, the ultrasound serves as a means of performing invasive operations in the chest. The use of ultrasound imaging for the chest is increasing, resulting in a significant increase in the quality of the diagnosis and a decrease in the incidence of various complications.

The ultrasound test uses sound waves to characterize various organ structure and function in health and disease. The test does not involve the use of ionizing radiation, so its use is safer than other techniques such as chest x-rays or CT examinations. The acquired and viewed images in the ultrasound device are real-time, so that changes can be detected and evaluated, which in many cases affects the final diagnosis. The test is non-invasive and does not require the use of contrast material containing substances that may cause an allergic reaction or impairment of kidney function. In addition, the instrumentation is more readily available for use at bedside examinations, and most tests do not require special preparations except for up to 6 hours of fasting in some cases.

On the other side, ultrasound testing has its drawbacks as well. The method is highly dependent on the operator's skill, so a great deal of experience is required to produce sufficient quality information and to make an accurate diagnosis. The ultrasound does not penetrate well through bones, and the quality of the test is impaired by the presence of air between the transducer and the organ or area being examined. In addition, the quality of the test is highly dependent on the subject's cooperation during the test, such as posture changes and deep breathing.

The standard technique for creating a standard 2D ultrasound image is now called Delay and Sum (DAS). Signals are transmitted and received from a series of elements, and by appropriate spacing of each signal, the signal readout can be focused in the required direction and distance. Due to the need for many sensors, a significant amount of data is required to produce the desired 2D picture. This fact leads to a need of using stronger processors, resulting in higher energy consumption, larger and more expensive operating systems. Therefore, Lowering the number of required samples while effectively processing the signals allows creating a high-quality 2D image, while narrowing down the burden of unnecessary data.

The SAMPL Laboratory at the Weizmann Institute demonstrated this data processing method with ultrasound tests on stationary organs such as the liver and kidneys where the number of elements needed for signal processing could be reduced for the purpose of creating an ultrasound image with a smaller volume of information, while not compromising on picture quality.

Objectives of the medical experiment:

The purpose of this study is to improve the diagnosis and characterization of pleural diseases by ultrasound, using a new algorithm developed in the SAMPL laboratory at the Weizmann Institute. The goal is to enable a fast and reliable diagnostic tool comparable to the ultrasound imaging used today while minimizing the sampling rate and volume of information.

Technical details:

B-Mode, M-Mode, and Doppler operation of an Ultrasonic system can provide spatial-, temporal- and motion-related aspects of Pleural composition and dynamics. The standard technique used by commercial medical ultrasound systems for B-mode imaging is delay and sum (DAS) beamforming. However, DAS often yields images of limited resolution and contrast, which are governed by the center frequency and the aperture size of the ultrasound transducer. A large number of elements leads to improved resolution, but simultaneously increases data size and system cost due to receiver electronics required per element. The reduction of receiving channels while producing high-quality images is thus of great importance. Methods and algorithms developed at SAMPL, such as the Convolutional Beamforming Algorithm (COBA), can be used to achieve significant improvement of lateral resolution and contrast. COBA can also be implemented efficiently using the fast Fourier transform. Based on this concept, sparse beamformers were developed, resulting in the same beam pattern as DAS and COBA while using far fewer array elements. Optimization of the number of elements shows an approximate square-root reduction in required elements, compared to DAS. The performance of the proposed methods was tested and validated using simulated data, phantom scans and in vivo cardiac data. The technique may be applied to Pleural disease detection, allowing for both improved image quality and reduction of the number of elements used to produce an image, and therefore ultimately facilitate cheap, portable, and wireless ultrasound imaging. In addition, the methods developed at SAMPL yield imaging that is fully dedicated to the assessment of pleural pathologies, enabling straightforward interpretation even by untrained clinicians. Jointly, wireless, cheap, and portable lung imaging, along with straightforward interpretation opens up a wealth of opportunities, spanning from fast diagnosis of pneumothorax in ambulatory (trauma) settings as well as in rural clinics and developing countries.

Equipment:

The Verasonics Vantage ultrasound system is an ultrasound system suitable for human research use. This system allows for raw sound wave data to be retrieved when performing ultrasound scanning.

Data security:

Volunteers will be scanned at Haemek Medical Center (As detailed in the "Intervention" section), and the data will be processed at the SAMPL laboratory at the Weizmann Institute. All information transmitted to the laboratory at the Weizmann Institute will be encoded beforehand, i.e. removing identifying information such as name, identification number, address, and assignment of a serial number by the medical center, so that the patient information is will be unavailable to SAMPL laboratory.

Transferring the images from the imaging tests will be done after encoding the tested information in an Excel file. Only the primary investigator will be exposed to pre-coding information, which will be stored on a dedicated computer by the primary investigator, password protected. The encoded information will be transmitted continuously to the Weizmann Institute, in order to ensure reliable data acquisition and the possibility of real-time feedback to the principal investigator in favor of higher quality data enabling data analysis.