My 2 cents worth on differences between more expensive vs cheaper ultrasounds, how ultrasound has changed over time, etc is:
* ultrasounds are getting cheaper and better. Harmonic imaging was probably the biggest advance for me.
* because of cheaper, better machines, many more doctors are using ultrasound than before. A real challenge is making sure they are appropriately trained to use ultrasound to get the most useful images, that they appropriately interpret what they see, and recognise their own limitations. The results you get from ultrasound are highly dependent on the training and experience of the operator, more than most other imaging modalities. Although some techniques are easy to learn, many require substantial training and experience. Putting ultrasound in the hands of someone who doesn't know how to interpret the findings is not risk free.
* there is a significant increase in image quality and the modalities available with more expensive machines (although this does not seem to be a linear relationship over the range of price; once you get into the top echelon of machines, differences in cost between them doesn't necessarily correlate with their capabilities).
* With more challenging assessments, these differences in image quality and modalities can mean the difference between a diagnostic and a non-diagnostic scan, and change management and outcomes for patients. However, there are some techniques (e.g. vascular access) where entry level machines are usually adequate, particularly if you have access to a better machine for challenging cases.
Although I'd love to see the quantum leap in imaging quality that some startups are apparently aiming for, I'm not holding my breath. That doesn't mean I don't see a lot of potential for technology to improve though - some things I expect or hope to see over the next decade or two are:
the application of computer vision and machine learning techniques to assist with the interpretation of scans. This is already happening a bit (e.g. strain imaging, automated 3D ejection fraction) and it's an active area of research. It has the potential to improve reproducibility of assessments. Maybe, just maybe it will get good enough to help mitigate the problem of less experienced operators using ultrasound.
multimodality fusion for interventional procedures - already, you can do a TEE (a kind of heart ultrasound) and fluoroscopy (video version of X-ray) and fuse the two images in real time to guide cardiac interventions. Maybe this will extend into other areas: e.g. guiding vascular access by fusing accelerometer / gyro / magnetic positioning / video camera data with ultrasound
virtual reality or augmented reality applications to accelerate the acquisition of ultrasound skills. There are already simulators but they're expensive; as they become cheaper and more doctors have access I'm hopeful we can improve the learning curve for image acquisition a bit!