Pattern recognition and the quantitative assessment of structures are important parts of reading diagnostic images and ultrasound is no exception to this. While expert users are able to consistently localize and characterize certain structures, it can be challenging for the less experienced sonographer. In addition, placing markers or tracing structures of interest for their quantitative assessment can be a time-consuming task, prone to errors or inaccuracies. Automated pattern recognition can therefore be useful to improve the workflow, quality, and consistency of exams for all users.

Machine Learning and particularly Deep Learning – a sophisticated AI method – can help to apply expert knowledge and skills to powerful algorithms that improve the recognizability of structures and simplify their quantification by automating the process. For this purpose, Machine Learning algorithms are trained with a large number of clinical cases, which have been evaluated by experts, in such a way that they can independently recognize or analyze certain structures. The training occurs at the factory before the algorithms are implemented in the product, so their behavior does not change.

Ultrasound users are continuously challenged by a growing number of difficult-to-scan patients, increased complexity of cases and a growing demand for standardized documentation according to the guidelines issued by their respective clinical associations. Combined with an increased demand and shortened time per patient, sonographers are looking for ways to enhance efficiency, while at the same time improving their diagnostic confidence. Improved imaging technology creates higher resolution, which drives increasingly complex measurements and analysis that require time and skill. Through accurate automatic measurements on a number of parameters, the Aplio can provide clinicians with reliable information faster and more efficiently.

“In my clinical practice, the challenges comes from the patients. Obesity is an increasingly key issue and there are short time slots for the echocardiographic examination in ambulatory practice,” remarked Professor Michel Zuber, formerly senior cardiologist at Cantonal Hospital Aarau, Switzerland. “Some years ago, it was not possible to get enough information from difficult-to-scan patients, as for example a patient weighing 125 kg, but now we can see that we have really good resolution with ultrasound, so we don’t need an MRI to obtain a simple functional and anatomical assessment, even in very difficult patients.”

“The new Aplio i-series from Canon helps me a lot, because we can get information from very difficult-to-scan patients in a short time. How? It is possible with automatic measurements due to Artificial Intelligence. That’s the keyword and the future of echo too,” he continued “We can get the workflow through the navigator, and this guides the examiner through the examination from view to view, and from mode to mode. So at the end you will not forget any measurements. We get very reproducible information from patients, such as ejection fraction or global longitudinal strain in one view or three-dimensional view, within a very short time. This guided approach is important to get always a full data set.”

Aplio achieves faster results through the use of accurate automatic measure-ments made possible by AI.

While patients become larger and more difficult to scan and evidence-based reporting of ultrasound examinations is becoming increasingly important, the number and complexity of required measurements is rising, putting increasing pressure on the examiner.

New tools that employ Deep Learning were developed for Aplio, enabling the operator to automatically identify standard scan planes and to carry out routine measurements without any user interaction. This is particularly helpful in where the relevant guidelines require a wide range of measure-ments, such as in Cardiology and Women’s Health.

“The biggest challenge we face in clinical routine is to perform examinations in crying children and across a wide range of patients ranging from fetuses to pre-term babies, children and up to adolescents with complex congenital disease,” remarked Professor Giovanni Di Salvo, cardiologist at the University Hospital of Padua, Padova, Italy. “Aplio helps us a lot to meet these challenges because the current version is equipped with Artificial Intelligence. This makes the machine extremely helpful during the normal workflow because we can get accurate information in a fast way. For instance, the acquisition of volumes or ejection fraction by using biplane Simpson or global longitudinal strain with the current series of Aplio is extremely fast and extremely effective and, of course, accurate. This is very important because those measurements are now the standards required by the guidelines.”

“Aplio also allows us to implement a lot of measurements into our daily routine that before were considered too complex and too time-consuming. Nowadays, we can implement them routinely in our practice, thanks to Artificial Intelligence. This makes a lot more measurements fast, accurate and reproducible,” he added.

Significant upgrades and the incorporation of AI to existing imaging and quantification technologies for fetal imaging into the new system has brought new and exciting possibilities for obstetricians.

“The new Aplio i-series Women’s Health model is a fantastic system with very reliable B-mode, great imaging resolution, and very precise Doppler. The new SMI (Superb Micro-vascular Imaging) brings us very exciting perspectives when imaging the fetal heart at a later stage – something that was not possible before. Now, you can use it in the late second trimester and also in early third trimester. It also gives a lot of information on the venous return system,” said Dr. Jader Cruz, fetal medicine specialist, Fetal Medicine Unit, Central University Hospital, Lisbon, Portugal.

“The new tool that I am most excited about is the introduction of Artificial Intelligence at a level that will be very helpful for education, training and even quality control of the images,” he added. “It is very interesting, and I am very glad to be working with it. I have been very satisfied working with Aplio for the past years and have fully explored it. And now Canon brings new technologies that enable us to go and learn even more. It is very exciting.”

Feedback from clinicians across the world indicates that the latest version of Aplio features technologies that enable significant improvements in the already outstanding image quality and clinical workflow truly representing a next-generation diagnostic ultrasound system. //

References:
¹ https://global.medical.canon/News/PressRelease/Detail/105723-834

Disclaimer:
Some products shown might not be available in all regulatory jurisdictions, please consult with your local Canon sales office for availability in your region.

You can also download this article as PDF