Trends in ultrasound from RSNA.
I had intended to knock off an article right after the RSNA, featuring commercial equipment first with clinical topics from the scientific sessions to follow. My original assignment was to report on ‘trends’. While I was trying to decide on practice pertinent trendiness, I got distracted by a glimmer of an issue below the radar concerning differences in trendsetting now versus the good old and/or wild west days when ultrasound imaging was still pretty new.
Before I tackle something elusive, I want to issue my own post-RSNA ‘Somascope’ Awards for commercial equipment, named for the original B-mode imaging device invented and built by Dr. Douglas Howry about 65 years ago. Incidentally, the first test object for the water immersion Reflectoscope (its original name) was a fluid-filled condom. I was looking forward to working with Dr. Howry at Massachusetts General Hospital, but he died shortly before I was to begin my diagnostic radiology residency there. I eventually started the clinical service for the hospital when I was a resident using an Aloka unit via the radar and nuclear medicine experts at the Raytheon Corporation in Waltham.
The Somascope for best full system goes to Supersonic Imagine. The Aixplorer, now in its eleventh major update, is best in class for image quality, Doppler sensitivity, and elastography. The Somascope for best portable unit goes to the Sonoscanner U-Lite for image quality, six interchangeable transducers, and a range of quantitative apps configured for subspecialty ultrasound indications. I happen to know those systems very well, because I was on the science board for Supersonic during its early years, and I have that position for Sonoscanner now. That is both a disclosure and a statement of pride that I was asked to contribute to these innovative and worthy companies.
I quite liked the SOFIA automated breast scan unit from Hitachi, which includes C-scan reconstruction, and the Delphinus SoftVue, which is now in phase III clinical trials. The SoftVue does a CT-type reconstruction of ultrasound propagation velocity. I intend to explore these units and their fundamentals subsequently. I also want to make a class Pre-Somatoscope award for the few companies who have commercial photoacoustic units ready for investigation.
The most recent RSNA allowed live ultrasound scanning of models. I do not think that was a WISE choice (which is my own opinion and a play on the Image WISEly campaign). The big companies will demo their finest, highly tweaked, and not necessarily commercial, equipment on movie models, slender, stunning, and fit. It is not a particularly educational or edifying exercise. As much as I like watching and admire skillful demonstrators do their thing, prolonged frivolous scanning of paid models kind of trivializes the serious nature of what we do with medical imaging on real patients every day. The ALARA principle is a good one, even considering the apparent extreme biological safety of pulsed ultrasound.
Beam Former Review
The beam former is the brain and soul of a modern ultrasound device. There are two basic kinds of beam formers, passive and dynamic. The beam former is also mated specifically to transducers, so that these are actually one functional unit like a radio and its antenna, or radar and its antenna array.
This is a brief reiteration of my pre-RSNA article. Simple, passive beam formers were state of the art in the 1980s, and are now fabricated on chips in relatively small packages. There were many new commercial units at Medica 2016 in Dusseldorf that had a passive beam former between a transducer and a separate (and physically separated) display module like a smart phone or tablet. Several of these will likely appear at next year’s RSNA.
There are relatively few, complex and technically evolved dynamic beam formers which characterize “high-end” equipment. These devices have multiple transmit options, like compounding and harmonics, and quasi-holographic focusing, and they may involve plane wave insonification or other stratagems to improve on the conventional and inefficient line-by-line form of image synthesis. There are pre-beamforming stages to handle the vast amount of raw data from all of the elements of the array. There are complex algorithms that optimize the final display from real-time analysis of signal features from different parts of the insonified field. There are compressed sensing methods, injection of statistical priors, and adaptive filtering operations, which is very demanding of hardware. The results are absolutely worth it, because of the gigantic improvement in the detail, speed, and reliability of exams.
One of the ploys of large manufacturers is to use a dynamic beam former with some of its functionality removed in a smaller, cheaper version of their flagship unit, like a platform ‘yacht’, marketed at a lower cost. I’ve talked about this before. It’s pretty fishy. My own opinion is that anyone who knowingly vends a limited performance device for diagnostic medical usage is contributing to the potential for medical negligence.
The problem for the prospective buyer is figuring out what system performance really may be.
I ambled across the massive North Exhibit Hall at McCormick place to the charming and quaint, and peripheral, French Quarter. The booths were smallish, the technology was way high, and the coffee was superb. My goal was to visit my friends at Sonoscanner. Figure 1 is the demo setup. I guess you could say that this is an example of hands-on scanning.
Typically, the higher the center frequency, the more the speckle noise, but here, noise is absent. The image is in focus throughout the field, and if you look at the vertical edges, there is no lateral resolution spread. There is a ‘parenchymal’ substructure with visible gray scale features. If Figure 1 can be magnified, have a look at the actual sponge within the beaker cup. Contrast and spatial resolution are both improved by noise reduction, like twice the bang for your center frequency and bandwidth dollar.
Excuse me for the almost bridal references. ‘Borrowed’ does not work, except maybe for trial use, and there isn’t anything blue about the U-Lite except the color Doppler display map. The new thing that I want to point out is the cable from the U-Lite to the transducer. It’s a fully flexible, weightless noodle. I do not know how it is made; I suppose it is a trade secret because there are a lot of no-loss wires for all the elements of an array. But, it is a significant and practical feature from the standpoint of the person scanning.
I am emphasizing the cable, because there has been a massive amount of hype from and about the Clarius portable unit, which is the counterpoint in this tale of two cities. A stated rationale for development for the Clarius, seems to have been an alleged need to replace a weightless wire bundle with a wireless transducer and a separate handheld display unit.
It is ironic that this ‘innovative’ development was presented at RSNA at least four years ago by Siemens as the Acuson Freestyle unit with a high capacity bluetooth transmission link from the transducers to the processor and display unit. My recollection is an excellent level of noise reduction for its three available transducers and buttons on the transducer case.
By the way, do you recall in old ultrasound systems that did line-by-line image synthesis, the tradeoff between line density and speed of imaging? There are tradeoffs between the number of elements in a transducer array, the data transmission capacity from transducer to processor, and image quality. The way to assess this, like image noise, is to view the image on a full-sized display, not a tiny screen.
I am referring to scanning by hand instead of ‘by the book’. (Word play is fine in editorials, never in clinical reports!) One of the promo pictures for the Clarius shows a proud user holding a probe in one hand and an iPhone in the other. How does he adjust the unit? Wait, he can’t. It’s easy enough to automate TGC and display grayscale assignments, which has been a standard equipment feature for a long time. But, as you know, that does not help with optimizing grayscale features for displaying parenchymal pathology, which is the core of modern diagnostic usage. I sometimes use voice activation for my smartphone camera; the promo prompted a dream of me gazing at my phone, looking at the patient beyond, and shouting ‘shoot’, ‘shoot’ as an exam was going on.
About adjustments themselves: some companies try to make a virtue out of removing most or all adjustments from the operator’s control. This is really dumb, because patients are not standardized by size, shape, or problem. There are only a few adjustments that are available in any unit; all you need to do is twiddle the knobs, slide the sliders, or whatever, while watching the image and decide what is best, even if you do not know what the controls are intended to do. There are some potential settings that almost never need resetting, and maybe only three or four that are used frequently. The idea of ‘simple enough for an idiot to use’ doesn’t seem ideal for a subjective procedure whose outcome can impact a patient’s life and happiness, nor does it send a respectful or supportive message to the health care professional who will be using the device.
The feature that I hate about the Clarius unit is the groove for holding the probe between the thumb and index finger. The basic factors for high speed imaging (that now go back 40 years) are that the probe is finger-held at its base like a pencil, the image is adjusted, as you watch it, by changing instrument settings and by repositioning the probe. The primary maneuver of ultrasound imaging starts with doing a search, finding a suspected lesion, centering it in the imaging field, and then rotating the probe to confirm that it is ‘real’, to assess its 3D structure, and to find the scan plane that demonstrates its features for documentation. The Clarius cannot be used effectively, because its transducer casing was designed for wrist motion.
A Nation Divided and Subdivided
There are plainly distinct application areas for diagnostic ultrasound. The common denominator for all applications is image quality. Image quality is not a slippery topic, it mainly comes down to elimination of structured and speckle forms of noise.
Multiple clinical applications for ultrasound is like medical and surgical subspecialization, necessitated, and made inevitable by the continuing knowledge explosion of the last century. The subspecializing physician builds upon primary competence in the general side of his or her own field. Ultrasound is special in the medical imaging sphere. Doing it requires a dual, inclusive education in technical and medical factors. It is a subjective test in which there is astounding complexity and variability in the factors related to pulse generation, propagation, tissue state, and reception. As we are entirely dependent on commercial equipment performance, one needs to build upon a common base of technical education even as we seek to master the medical intricacies of any clinical application realm.
For the last 25 years, I have seen a diaspora of ultrasound into almost all corners of medical practice in the general absence of technical understanding. Along with this spread, there has been a progressive loss of ultrasound expertise from radiology. The result has been a proliferation of ultrasound with scarcely any extension of clinical range or utility since the 1990s.
There have been several attempts to make portable equipment throughout the history of ultrasound, but they have pretty much all been commercial failures. For many, it was development when the field was more cohesive and with a better sense of image quality. Now that 25-year-old image quality is achievable in a small package and small display unit quality has improved, portable, and ultra-portable are becoming the rage. Who is setting this trend anyway?
There is the notion that someone, who does not have ultrasound yet, need only pick up one of these units to get into the game. The companies have tended to target these units to beginners for niche applications, not radiologists, cardiologists, or even obstetricians. They tend to assume that the units will be used by physicians in the course of their routine work. High performance equipment, from the same manufacturers is reserved for ‘expert’ use. That’s trendsetting by manufacturer, which is problematic, because they do not do studies themselves or share in responsibility for our clinical performance.
Is it OK to have rigid professional standards for critical care, surgery, radiology, or whatever, and revert to some loose amateur framework for ultrasound, because of its affordability, profit margin, and/or biological safety? This is important if we, as a profession, are to avoid tests and procedures that are inappropriate, wasted, non-diagnostic, or in error. In a way, this is like commercial live scanning at RSNA, as distinct from hands-on continuing education courses.
Sonoscanner has shown that it is possible to have a high performance, dynamic beam former, full-featured system in an ultraportable package. This is the way to spread ultrasound, and if this is to work, there cannot be turf issues. Radiology needs to regain its ultrasound imaging expertise globally and act locally to diffuse high performance ultrasound technology throughout its community. Manufacturers need to treat the users of ultrasound equipment with the respect; users and their educational surrogates need to judge equipment by actual performance alone; and we can all need to have free and collegial conversations about our work and difficult cases.
Now, on to a fabulous New Year.