Do you really understand four dimensions?

Four-dimensional ultrasound imaging has been around for decades, evolving from mechanical arm acquisition to built-in stepper motor volume probe acquisition, and now to the current array probe acquisition mode. 3D/4D has always been a research hotspot. Obstetrics 4D is undoubtedly the most familiar and easily understood application. It uses various advanced rendering techniques to transform three-dimensional spatial echo data into a two-dimensional image with a three-dimensional effect. Today, let's discuss some technical points in the development process of four-dimensional ultrasound technology.

Question 1: Why are abdominal four-dimensional probes so large?

​

Conventional large convex/mechanical volume probe/array probe

The probe determines the technological advancement of four-dimensional ultrasound, as well as the image quality and refresh rate that users care most about. Currently, most four-dimensional ultrasound probes are divided into two types: array probes and volume probes.

An array probe has thousands of elements, which are not arranged in a traditional linear array but in a matrix array. This arrangement allows the probe to transmit and receive in different directions in three-dimensional space, thus quickly obtaining data. Theoretically, the array probe is the primary component for high frame rates. Array probes also come in various Models. Phased array probes are already quite common in cardiac ultrasound, and major manufacturers are gradually mastering this technology; however, the difficulty of implementing convex array probes for obstetrics and gynecology is much greater. Only in recent years have GE and Philips been able to offer relatively mature products, but they are constantly being optimized and updated. The array probe contains thousands of elements, signal processing chips, and heat dissipation-related components. It can be said that in order to optimize the application of these thousands of elements' signals, the array probe is actually a small ultrasound front-end system. The probe usually has an ASIC chip to solve the front-end signal wave velocity synthesis processing and heat dissipation efficiency problems. The Size differences of probes from different companies are huge. In order to fit all these components into the probe without exceeding the heat limit, the convex array probes are generally larger.

The acoustic head part inside the volume probe is actually a conventional one-dimensional probe. Its large Size is due to the addition of a motor and transmission device inside to drive the probe to swing at a certain angle. A single swing can obtain a volume of spatial data. Currently, the vast majority of obstetrics 4D uses volume probes. This type of probe can be considered a transitional solution for four-dimensional ultrasound compared to array probes (it can also be considered a miniaturization of the mechanical arm solution). However, from an industrial perspective, its process is relatively simple and stable, without the complex process and heat dissipation problems of array probes, and the Price is also much cheaper (in fact, there are no manufacturers selling array probes on the market yet).

The acoustic head part of the volume probe is a conventional one-dimensional acoustic head array.

Question 2: Is the emitted energy of four-dimensional ultrasound greater than that of two-dimensional ultrasound?

Four-dimensional ultrasound adds two more dimensions compared to conventional two-dimensional ultrasound, naturally leading one to wonder whether the amount of sound wave energy received by the human body will be greater during use. In fact, both the US FDA and China's SFDA have strict limits on the emitted sound wave energy of ultrasound. In simple terms, it stipulates that the sound wave energy received at a unit time at an organizational location must meet certain conditions to be considered safe. In fact, relative to B, C, and other two-dimensional imaging or PW, a one-dimensional imaging, when the doctor holds the probe still, the sound power at each organizational location under four-dimensional imaging is generally not greater than that of the others.

1. 1. The number of emissions per second is basically the same under B/C/D modes, constantly emitting. Therefore, it can be considered that the energy emitted per second in four-dimensional mode is not more than in other modes.

   2. The received dose at each location is less in four-dimensional imaging because the motor drives the probe to swing. The emission energy that originally bore on a two-dimensional plane per second is now dispersed into a three-dimensional space.

   3. In practical applications, the four-dimensional continuous scanning time is short. Doctors will not continuously image in three-dimensional and four-dimensional ultrasound as they do in two-dimensional scanning. In four-dimensional ultrasound, doctors generally will not continuously collect four-dimensional data for more than half a minute before freezing, then adjusting the acquisition position and conditions to collect suitable sections. This intermittent operation also ensures the safety of the average unit time energy of sound emission.

Therefore, four-dimensional ultrasound does not emit more energy than two-dimensional ultrasound, and both are energy controllable.

Question 3: Why can't the frame rate be the same as B-mode?

In four-dimensional ultrasound, the probe swings from one direction to another to obtain a volume of data, which is then rendered into a two-dimensional stereoscopic image for display. The number of volumes of data that can be obtained per second determines the number of times that can be displayed or refreshed per second. Relative to the frame rate (fps, frame per second) in B-mode, four-dimensional ultrasound is generally referred to as the volume rate (vps, volume per second). Currently, even for the highest-end ultrasound systems, as long as a volume probe is used, the conventional volume rate is between 2~4 vps, which is somewhat different from the refresh rate of over 20 fps in B-mode.

The volume rate of a four-dimensional ultrasound system is constrained by several factors:

1. The swinging ability of the motor

2. Volume data image quality

3. System platform processing capability

The author has basically come into contact with several volume probes on the market, and even tested the motor swing capability of the highest-end machine's volume probe. It can be considered that the current bottleneck is not the swing capability of the probe motor. At 60 degrees, stable swinging above 5 vps is not a big problem. Higher requirements would require a powerful motor. As mentioned earlier, the volume rate of most obstetrics 4D ultrasound is 2~4 vps, so probe swinging is not the main bottleneck of the volume rate.

The bottleneck mainly lies in the consideration of image quality. There are two aspects to this consideration: one is the reduced data volume due to high volume rate affecting image quality. For acquisition at a specific depth, the distance traveled by each emitted sound wave is known, and the spatial range acquired by the probe (defined by two parameters: swing angle and B-plane scanning range) is also determined by the user.

At this time, the volume rate actually depends on the number of times the space needs to be scanned. The fewer the number of transmissions, the higher the volume rate. For example, if the scanning depth is 12cm and the speed of sound is 1540m/s, then after one transmission, the sound wave travels to this depth and then the echo is received by the probe surface. The time of one transmission and reception is t=12*2/1540 seconds. If a volume of data needs N transmissions, then the theoretical volume rate can be calculated, which is 1/(N*t). Assuming the conventional line density of B (about 160 lines per image, dual beam, i.e., 80 transmissions per frame), a swing angle of 60 degrees, and 60 frames of volume data, then N=80*60=4800, the theoretical volume rate is 1.34vps. The reason why this is called the theoretical volume rate is that the probe swing and system transmission and reception will have various limitations, so the actual volume rate needs to be discounted from this theoretical volume rate. It can be seen that the volume rate is inversely proportional to the amount of data, or more precisely, the volume rate is inversely proportional to the number of transmissions. To improve the volume rate, the number of transmissions should be reduced, which leads to less data, resulting in poor image quality (just like the difference between low line density and high line density images).

On the other hand, high volume rate and fast swing also affect the signal quality of wave speed synthesis. With the same amount of data, the swing of the probe will also lead to inconsistencies in the transmission and reception positions, making the data clutter noise more serious. The problem of poor data quality caused by fast swing is the main factor for the low volume rate in volume probes.

System platform processing capacity. In recent years, due to the strengthening of the general processing capacity of GPUs, more and more ultrasound machines have gradually used GPUs as coprocessors to process ultrasound data and perform high-performance rendering algorithms. General desktop computers use independent high-performance graphics cards as coprocessors. GPUs are very suitable for algorithms with high parallel processing requirements such as three-dimensional/four-dimensional, and can easily meet the performance requirements of volume data rendering processing. It can be said that a 200 yuan PC graphics card can easily defeat the latest i7 processor in terms of 4D rendering speed. However, for portable ultrasound machines, due to the strict requirements for the power consumption of devices, conventional desktop graphics cards cannot be used, so performance is a key issue that cannot be ignored. Currently, there are not many portable machines that support 4D, partly because not many obstetrics and gynecology departments use portable machines, and partly because the processing capacity is weak.

This information is sourced from the internet. Please contact us for removal if there is any infringement.