DI Europe Spring 2023

Boosting

Dear reader,

Sustainability was on everyone’s lips at ECR 2023, the European Congress of Radiology, which took place March 1-5 in Vienna, Austria. But in spite of all the in-depth conversations and innovative strategies unveiled by the industry during the meeting, artificial intelligence (AI) got the lion’s share in terms of content and novelties. By all accounts, AI has never been more present at ECR than last March. The number of scientific sessions dedicated to the technology was the highest so far at the congress, with 84 sessions tackling every aspect from education and ethics to reimbursement, and packed industry audiences in the AI Theatre. The exhibition space granted to AI vendors was bigger than ever, with booths rivalling those of established hardware manufacturers, reflecting the ever increasing number of players in the field.

AI in ultrasound: crossing the last frontier

In terms of equipment, it’s now daring to find one device that isn’t powered by AI. But the biggest news in Vienna was probably that a major frontier has been crossed. A myriad of ultrasound systems equipped with AI were launched, showing the market’s growing appetite for the technology.

Vendors and experts believe that AI will help lift some of the longtime obstacles to ultrasound’s widespread use, such as inter operator dependence and long learning curves. We’ll be sure to look at how this exciting development truly advances clinical practice and improves patient care in the coming months. Because it has taken on such a preponderant role in radiology, we’ve decided to focus on AI in our new (and my and Guido Gebhardt’s first) issue of Diagnostic Imaging Europe.

I hope you will enjoy our selection of articles, interviews and expert opinions, as we’ve ventured inside and outside of radiology to look at the impact of AI on healthcare as a whole.

Let’s dive deep into the topic and get inspired by its multiple uses in our tech savvy ecosystem.

Thanks, merci, danke, gracias, grazie, miłosierdzie, Ευχαριστώ to all of you!

IMPRINT

Editors

Mélisande Rouger

m.rouger@dieurope.com

Guido Gebhardt g.gebhardt@dieurope.com

Media Sales Manager

Marika Cooper m.cooper@dieurope.com

Senior Editor

Alan Barclay PhD a.barclay@dieurope.com

Article reprints info@dieurope.com

International Sales Offices

Europe, North America and Japan

Marika Cooper phone: + 44 7941023512 m.cooper@dieurope.com

Korea Young Media Inc

Contact: Keon Chang

phone: + 82 2 2273 4819 ymedia@chol.com

Layout & Graphic Design

Christoph Muschiol Germany christoph@muschiol-online.de

Disclaimer

The information and opinions expressed in articles and product entries published in DI EUROPE are solely those of the manufacturers / companies, their authors and contributors, for which the publisher holds no responsibility.

All trademarks, product names, company names and logos in this publication are the property of their respective holders. Users must obtain permission from those holders before copying or using the owner’s trademarks, product and company names or logos.

Diagnostic Imaging Europe Magazine is published by Mélisande Rouger

Juan Duque 7 5C

28005 Madrid · Spain & Guido Gebhardt

Adalbert-Stifter-Weg 2 b

85661 Forstinning · Germany phone +49 1511 5674833 info@dieurope.com

© 2023 DI EUROPE

All rights reserved.

Reproduction in any form is prohibited without permission of the publisher.

DI  EUROPE

6 New EU plan to boost Cancer Research with AI

The EU plans to use federated learning to help hospitals build data infrastructures and share data. Prof. Luis Martí Bonmatí told Mélisande Rouger why this is a major advance for cancer research.

8 Kotter on AI’s Present and Future

Mélisande Rouger spoke with Prof. Elmar Kotter, immediate Past-President of the European Society of Medical Imaging Informatics (EuSoMII), about trends and education in radiology AI.

10 With Ultrasound into the Future

MyLab X90 is one of the first ultrasound systems powered by AI. We spoke with Franco Fontana, CEO of Esaote, to know more about the technology behind the machine and how it will help boost ultrasound use in clinical practice.

14

Introducing the new Aquilion Serve CT System

Canon Medical introduced its new 80 / 160-slice CT scanner, the Aquilion Serve at ECR 2023. Find out how the system uses AI to streamline examinations and boost workflow.

18

Unprecedented AI Presence at ECR’s Show Floor

ECR returned to its traditional spot early March and a level of activity not seen for the past four years. That was just the time AI companies needed to take over the technical exhibiton, Guido Gebhardt found out.

20

23 Views of ECR 2023

(Re)discover the key messages and products presented at ECR 2023 through our 23 video interviews

24 Spiral Breast CT: Going Further with Photon Counting

Radiologists at University Hospital Erlangen introduce the innovative technical features of the nu:view breast CT system and discribe recent clinical findings and likely indications of this new equipment.

28 Broad Solutions & Platforms to lead Radiology AI Development

Dr. Sanjay M Parekh of Signify Research delivers his analysis and forecasts for the booming medical imaging AI market. Hint: investment will continue to grow at least until 2026.

32 Weaving AI into the Emerging Strategic Decisions in Healthcare

After the Covid-19 crisis, healthcare in Europe is being reconfigured. The process of decision-making and investment across the EU can weave AI into the changes that will happen in the immediate future, according to an eminent expert in public health.

34 deepc secures € 12 M funds to expand its AI Platform, partners with LMU

The LMU University Hospital has entered a close strategic partnership with deepc, which recently received €12M funds to expand its platform.

36 Bayer reinforces Commitment to Innovation in Radiology

Bayer‘s new cloud-hosted platform Calantic provides access to AI applications that integrate into standard medical imaging workflow to improve prioritisation, lesion detection, quantification and productivity.

38 Elevating Efficiency with AI

Guido Gebhardt spoke with Prof. Georg Langs, Founder of Viennese AI company contextflow, about integrating AI in RIS and PACS.

42 Photon counting CT: A new Challenge for Contrast Administration?

At ECR 2023, experts discussed the impact of PC-CT on diagnostic imaging and cardiac and non-cardiac applications, and the implications of the new technology on contrast medium usage.

48 United Imaging unveils Breakthrough 5T MRI System at ECR 2023

The Chinese company launched the uMR Jupiter 5T, a pioneering whole-body ultra-high field 5T MRI scanner, and new AI features in Vienna.

50 AI articles in radiology

Alan Barclay spoke with Dr. Hannah Hughes, who wrote a detailed bibliometric analysis of the current literature on AI in radiology.

“ The MRI specialist dealer that I trust”

New EU Plan to boost Cancer Research with AI

Millions of diagnostic imaging studies are being performed each year in the EU, but only 100,000 of these are being used for research purposes. A new European project powered by artificial intelligence (AI) plans to make this data available to clinicians and researchers to advance cancer care, Luis Martí Bonmatí, Professor of Radiology at La Fe University and Polytechnic Hospital Valencia, Spain told delegates at ECR 2023.

explained in a dedicated session. “We need to integrate existing repositories and projects that have generated large databases and databanks,” he told the audience. “A major issue is how to create data sharing agreements in clinical environments to incite hospitals to share their data. They’re still reluctant with GDPR and the potential risk of patient re-identification.”

been put together, and a repository with 5,000 cases has been created, that data is often forgotten as soon as the project is finished.

A European solution to tackle data privacy

By Mélisande Rouger

With nearly four billion diagnostic imaging studies performed each year worldwide (1), the amount of data that can be mined to improve outcomes in patients with cancer is dazzling. But to drill into this goldmine, systems must be made able to integrate and share the data. And these two steps have proved quite challenging so far, Martí Bonmatí

The degree of digital maturity varies from one hospital to another. But generally, it is very low. “We tend to believe that hospitals and clinical infrastructures are more technologically mature than they really are,” he said. ‘The data there is often unstructured, scattered and complicated to link. That causes many problems.”

Even when the data from radiology, pathological anatomy, radiotherapy and patient outcome has

To reuse existing data and encourage researchers to produce new information, the EUCAIM consortium and the European Commission have recently launched the European Cancer Imaging Initiative (EUCAIM), an infrastructure deployment project that feeds on federated learning, an AI technique that tackles all issues related to data privacy and security.

“Federated learning enables multiple actors to build a common and robust machine learning model without sharing data, allowing to address data privacy, security and access rights,” said Martí Bonmatí,

Image: Luis Martí Bonmatí

Luis Martí Bonmatí is Director of the Medical Imaging Department and Chairman of Radiology at La Fe University and Polytechnic Hospital in Valencia, Spain.

who serves as EUCAIM scientific director. “We will address the fragmentation of existing cancer image repositories and establish a distributed Atlas of Cancer Imaging, with over 60 million anonymized cancer image data from over 100,000 patients.”

The data will be accessible to clinicians, researchers and innovators across the EU for the development and benchmarking of trustworthy AI tools. The project builds upon the results of the “AI for Health Imaging” (AI4HI) Network, which consists of five large EU-funded projects on big data and

AI in cancer imaging – Chaimeleon, EuCanImage, ProCancer-I, Incisive and Primage.

EUCAIM brings together 76 partners from 14 EU member states, covering competences in cancer imaging and care; big data in medical imaging; FAIR data management; ethical and legal aspects of medical data; development and deployment of research infrastructures; AI and machine learning; and dissemination, communication and stakeholder outreach in biomedical imaging. “The idea is to bring the AI solutions to the data warehouses of the hospitals, so that they compute there and build the model,” he said.

EUCAIM is coordinated by the European Institute for Biomedical Imaging Research (EIBIR) in Vienna and it is the cornerstone of the European Commission’s European Cancer Imaging Initiative, a flagship

of Europe’s Beating Cancer Plan. It is a first step towards putting some order into the hospitals, Martí Bonmatí believes. “Hospitals in Europe have to consolidate a data warehouse that allows them to use the data they generate every day to optimize hospital management and to encourage investment in data,” he said.

The development of AI is a consequence, he concluded. “You need to have your hospital well organized to know how to use the data. Having an organized warehouse is the basis.”

https://digital-strategy.ec.europa. eu/en/policies/cancer-imaging

References 1 https://www.businesswire.com/news/ home/20210608005582/en/Global-MedicalImaging-Market-Report-2021-2026-Analysisby-X-ray-Ultrasound-MRI-CT-Scan-NuclearImaging---ResearchAndMarkets.com

The European Cancer Imaging Initiative (EUCAIM) is an infrastructure deployment project that feeds on federated learning to tackle all issues related to data privacy and security.

Kotter on AI’s Present and Future

DI Europe spoke with Prof. Elmar Kotter, immediate Past-President of the European Society of Medical Imaging Informatics (EuSoMII), about the current and future trends in AI and the content dedicated to the technology at ECR 2023.

By Mélisande Rouger

◾ As Chair of the eHealth and Informatics Subcommittee of the European Society of Radiology (ESR), you oversaw the AI content at ECR. Which sessions stood out?

AI has become one of the main topics at ECR because it plays an increasingly important role in radiology.

During ECR 2023, 84 sessions were labeled AI and machine learning. That’s a lot! The sessions addressed the three different levels of the European training curriculum. Besides, we had 25 industry sessions in the AI Theatre that attracted crowds of delegates.

Some sessions particularly generated interest from the audience, for example the session on data sharing fueling AI development, where Luis Martí Bonmatí, Laure Fournier, Rick Abramson, Martin Willemink, Peter Van Ooijen and I discussed how important data sharing is to AI development.

Another session tackled the EUropean Federation for CAncer IMages (EUCAIM) initiative, which aims to foster innovation and deployment of digital technologies in cancer treatment and care.

We also had content on AI for dose optimization and management; quality control; and on ethics and sustainability, to name a few. In my term as EuSoMII Presi-

dent, I initiated a partnership with the European Federation of Radiographers’ Societies, as radiographers will play an important role in the future of AI.

Another really interesting session was the one organized by the European School of Radiology (ESOR), on how to adapt education in radiology. The session was chaired by ECR 2023 President Adrian Brady and Valérie Vilgrain, Head of ESOR, and my talk focussed on how to teach radiologists about AI.

◾ How educated are radiologists when it comes to AI?

It’s very difficult to assess the knowledge of AI radiologists have. Some have very good knowledge, others do not. Most radiologists probably are on the application side, but even then, you need some basic knowledge of AI. As an analogy, you do not need to be a physicist to run an MRI scanner, but you need to understand the basic principles to reduce examination time and be able to explain the problems that may arise when using the technology.

The same goes for AI. You need to know the basic principles and what kind of errors may occur. There is no perfect system. AI, just like other systems, will make errors. You need to check the tool and that takes time. Then also comes the question on how you are going to finance AI.

We know that AI works in radiology, but how are we going to pay for it? That remains a difficult discussion in Europe as most countries don’t have any dedicated reimbursement scheme. To help radiologists get a broader picture of AI, we have just launched the ESR Master Class in AI, which comprises of 70 lectures that are organized into five modules.

◾ What should radiologists do when AI makes a mistake?

Before you buy a solution, you need to test it over a few months to see if it runs well. Once AI is running, you need to monitor its performance. That is something very difficult to implement and most institutions do not do it. And then a time will come when the radiologist does not agree with the tool and decides to override it. Then you need to tell the company about the problem.

There are only very few AI systems that allow us to deliver feedback to their developpers in a direct and easy fashion, like pressing a button. Typically, you need to call the company or write an email. This is time consuming, and we are already short on time.

AI companies should thus offer an option for users to communicate their feedback more easily. Algorithms make errors and and it is important to involve the radiologist in the loop.

◾ There was a session on return on investment (ROI) in AI in the AI Theatre. Why is that an important topic? It is very hard to show how much money can be saved buying an AI

tool. It is easier if there is no radiologist onsite, for example in the emergency department, and AI acts as a safety net, or when you have a general lack of radiologists.

But for other situations, it is an unresolved question and the answer also depends on the country. In the United States, you already have reimbursement for some applications. In Europe, there are ongoing discussions with insurance companies and the payers to know who should pay for AI. One possibility would be to prove that you can treat patients earlier, faster, and thus reduce hospital stay duration; or save time for radiologists.

Most people believe that the radiologist should be the first reader and AI should take a second look. Sometimes AI will detect nodules you have missed in a lung CT scan, but it does not allow you to work faster. If we want to save time, then maybe AI should read the scans first. There are many applications outside diagnostic imaging, to improve for example logistics in radiology and optimize work lists. Regarding reimbursement, another possibility to reimburse AI would be not through the healthcare system or the hospital, but through the pharmaceutical industry. Imagine a pharma company has a new medication and finds an AI company who identifies the pathologies for which

Elmar Kotter is Associate Professor of Radiology and Vice Chair and Director of Imaging Informatics & Quality and Risk Management at the Department of Radiology of Freiburg University Medical Center in Germany.

He is also Chair of the eHealth and Informatics Subcommittee of the ESR and the immediate Past-President of EuSoMII.

this medication works. This would save so much time and money. Bayer has recently bought Blackford Analysis and more pharmaceutical companies will buy AI developers in the future. Some AI companies are changing their orientation and do not just try to sell to hospitals but the pharmaceutical industry as well. This trend will continue in the next few years.

https://www.eusomii.org

With Ultrasound into the Future

MyLab X90, Esaote’s newest release to date, is one of the first ultrasound systems powered by Artificial Intelligence (AI) or Augmented Insight, as the Italian company puts it. We spoke with Franco Fontana, Esaote’s CEO, to know how the new system, which was launched at ECR 2023, can help improve healthcare. We also discussed the value of low field MRI and current needs in medical IT solutions.

By Mélisande Rouger

◾ AI is available on almost every modality but ultrasound was the last frontier.. How does AI expand ultrasound imaging on your new system?

There is a lot of ongoing work at the scientific and industry level, but the foundation for AI started many years ago. What has changed now is that we are able to process very large data and build up knowledge based on the data that has been collected, to train algorithms to do what we want to do. That is a big step ahead. Our focus at Esaote has been to develop solutions to streamline and facilitate workflow. It is all about letting the physicians focus on what really matters – clinical outcomes –,

by sparing them tedious or repetitive tasks.

We have spent a lot of time with our R & D team putting up together all this data and knowledge, and discussing with experts. There is not one algorithm that works for everything. You have to work on every single anatomy, every single clinical application to build up your knowledge. Ultrasound examinations are complex. When you put a probe to scan, there are things and adjustments to take care of, because patients and applications may differ widely. MyLab X90 provides automatic AI-powered setting to recognize how the probes react and improve image quality. AI also enables to reduce scan time, by extrapolating measurements’ findings.

Franco Fontana, PhD, has been Chief Executive Officer of the Esaote Group since 2019. He has more than 25 years of experience in medical imaging and healthcare IT. He co-founded the PACS vendor AETmed, which was later incorporated into Esaote, giving origin to Ebit. He served as CEO of Ebit until 2020. Under his leadership, the PACS business has grown exponentially, with Ebit becoming a market leader in Italy and consolidating its international presence around the world. F. Fontana received his PhD in Telecommunications from the University of Genoa, and his executive MBA from Bocconi Business School in Milan, Italy.

The system is useful in every application, from cardiac imaging, to extract all measurements from the cardiac cycle, to musculoskeletal, liver and women’s imaging, to name a few. You can automatically take 100 measurements in just a few seconds, and do fusion imaging for your interventional procedures with the equipment.

◾ Inter operator variability is still an issue in ultrasound imaging. How does AI help tackle the problem?

Our task is to design something that is fully reproducible and offers the same performance independently of external conditions. Measurements have to be repeatable.

With AI, systems will be more accessible to non-expert operators. MyLab X90 offers the possibility of automatically selecting the best scan plan to guide users throughout the examination. It automatically indicates if you have fully covered an organ in your examination or if you have missed something.

AI can make a big difference in interventional radiology, especially for beginners. You can match different modalities and do fusion imaging. It is one of the strong points of our system. It allows combining different data in real time on screen, to make sure you are doing the right thing.

◾ Which other technological advances does the system feature?

MyLab X90 offers premium-level components like XCrystal probes and exclusive dual-layer eLed. We have done a great work on our probes, thanks to our factories in Italy and the Netherlands, and all the feedback and discussions we had with experts. There is a direct connection between the probe and image quality to improve diagnostic accuracy.

We also put a lot of effort into user experience and interaction with the system. Design is important. It is not just about how aesthetic your solution is, but also how easy to use it is.

So we spent a lot of time with our Italian designers to build something that is both beautiful and useful.

◾ You launched MyLab X90 a few weeks ago. What has been the response so far?

The reaction has been very positive. There is a lot of interest, we can see that in the road show we organized in our key geographical areas. Being a premium system, MyLab X90 is targeted at tertiary or reference hospitals and centres that do advanced research in interventional radiology, for instance, or follow complex patients. These institutions are very receptive to new technology and we already installed the system at different sites across Europe.

MyLab X90, the new premium ultrasound system with Augmented Insight

Liver subcostal scanning with Wide view technology

Thyroid nodule with microV technology

Suprahepatic vein visualization with BrightFlow technology

Thyroid nodule with microV technology

◾ Esaote is also famous for its MRI systems. How has the Magnifico Open been doing since it was released two years ago?

The Magnifico Open, our first total body MR system, was launched in Europe in 2021 and received FDA approval in 2022. Our key markets reacted really well to the introduction of the system, which has pushed our MRI sales by 50 % last year. 2022 was actually the best year in our history, with a significant growth of our ultrasound line and a +10 % forecast turnover for the group.

The Magnifico Open is great not just for diagnostic imaging centers, but also orthopedics and surgery clinics. It is an ideal tool for musculoskeletal examinations, which we helped pioneer with the first MRI system dedicated to the limbs in 1992. MSK examinations are still the most commonly performed MRI scans today.

The Magnifico Open is a low field MR system, which offers so many benefits, starting with wider patient acceptance. It is an open technology, which means that examinations no longer have to be performed in a closed tube. So claustrophobia is not an issue anymore.

Another advantage is ease of installation. You do not need a very complex settlement to build or arrange where the system is going to be installed. Open technology also means low energy consumption, which fits our concern for delivering sustainable solutions.

In addition, the system comes with great software that helps reduce scan time and improves image reconstruction.

◾ Hospitals increasingly demand complete solutions, not just medical equipment. What do you offer in terms of IT?

IT is a growing market and we also deliver cutting edge software. We have two lines in our IT solutions. The first is called enterprise imaging, where we offer a complete inhouse solution package that we pioneered at the end of 1990s, pushing the transition from film to digital.  Digitalization saves costs and time. Everyone realizes the huge benefit of working with single electronic patient records and combining the data to improve workflows and patient outcomes. That work is done by Ebit, our IT subsidiary in Italy.

Our other IT line is in charge of everything connected with AI. That work is done by our subsidiary in the Netherlands, PIE Medical Imaging.

◾ What does the future look like for Esaote?

There is still a lot to do in terms of image accuracy to better serve our customers’ needs. This is driving us. Ultrasound is widely recognized to be the stethoscope of the future and so we may also go back to where we started and take our probes to space again. Esaote provided the ultrasound system for cardiovascular examinations in astronauts into orbit on the NASA’s Space Shuttle back in the early 1990s.

With the development of private manned space flights, we may soon be able to help in this setting as well.

https://www.esaote.com

Magnifico Open, Esaote’s total body MRI system

IMPROVED PATIENT CARE

IMPROVED PATIENT EXPERIENCE

IMPROVED WORKFLOW

IMPROVED COST EFFICIENCY

Introducing the new Aquilion Serve CT System

Canon Medical has introduced its new 80 / 160-slice CT scanner, the Aquilion Serve at ECR 2023. Designed for busy medical imaging departments, it delivers consistent imaging results, better image quality, lower radiation dose and faster throughput, creating more time for patient care.

The Aquilion Serve is an efficient solution for all routine examinations, including trauma. Its newly designed 80-cm wide bore gantry features two, easy-to-use, touch panels and inbuilt cameras that support automated one-touch patient positioning. It also introduces the next-generation INSTINX workflow solution, which combines AI-enabled automation with innovative hardware and an intuitive user experience to support fast, easy, and safe CT examinations. The new workflow also dramatically reduces training time for operators.

Preview first

The new system introduces an industry-first, 3D Landmark Scan, which is acquired at the same dose as a traditional dual 2D scanogram. 3D Landmark Scan provides a preview of the full scan range cross-sectional dataset in addition to the traditional 2D view for scan plan-

ning. This, in conjunction with Anatomical Landmark Detection (ALD), can accurately identify the anatomical structures required to perform automatic scan planning for all routine examinations. In addition, scan range and field of view can be automatically set to a position that is predefined in the scan protocol to save time, while ensuring consistent results for all CT technologists.

Sharp imaging

For image quality, Advanced intelligent Clear-IQ Engine (AiCE) harnesses the power of deep learning to distinguish signal from noise and deliver extremely sharp, clear and distinct CT images. Fully integrated into the patient-centric workflow, it also reduces dose levels significantly.

New standards in workflow efficiency

As part of the new system, Canon Medical introduces INSTINX, a total workflow experience redesigned from the ground up to set new standards in efficiency and consistency. Every detail of the workflow has been thoroughly refined based on clinical testing in medical centers around the world. Now every operation is more intuitive and can be learned faster than ever before. This ease-of-use contributes to work satisfaction, time savings and flexible allocation of resources.

“The Aquilion Serve intelligently supports a patient’s journey through a CT examination with technology that will change the way you perform a CT examination forever,” remarked Naoki Sugihara, Vice President and General Manager of CT Systems Division at Canon Medical Systems Corporation. “With an optimized workflow experience that enables consistent results to be provided more quickly while maintaining low dose, Aquilion Serve ‘simply delivers’.”

Image: Canon Medical

The

Aquilion Serve is suitable for all routine examinations. It features a new 80 cm wide bore gantry with two easy-to-use touch panels and inbuilt cameras that enable automated, one-touch patient positioning. It also introduces Canon Medical’s unique INSTINX workflow solution, which combines AI-enabled automation with innovative hardware and an intuitive user experience to support fast, easy, and safe CT exams.

IMPROVED PATIENT CARE

◾ Industry’s first 3D Landmark Scan to accurately identify the anatomical structures

◾ All-round, versatile capabilities for any type of patient or examination

◾ Low dose, consistency in image quality and contrast reduction made possible through Deep Learning Reconstruction (Advanced Intelligent Clear IQ Engine)

◾ SilverBeam filter, a beam-shaping filter to selectively remove low energy photons from a polychromatic X-ray beam, leaving an energy spectrum optimized for high contrast CT applications

IMPROVED WORKFLOW

◾ Integrated, built-in cameras to seamlessly set the patient in the iso-center

◾ Consistency thanks to a new intuitive user interface and a robust scanning process

◾ Simple scan planning and simple patient positioning

◾ Easy-to-learn user experience for both CT experts and first-time users

◾ AI-based anatomical landmark detection for patient positioning and scan range

IMPROVED PATIENT EXPERIENCE

◾ Unique flared gantry design with an opening of 80 cm providing a calming, wide-open space

◾ Increased safety with Tech Assist Lateral Slide, a feature that mechanically moves the patient up, down, left, or right to the correct position at the touch of a button

◾ In-gantry lighting and quiet gantry for optimal patient comfort

◾ No compromise on patient safety by using AI guided scan set-up and simple operation buttons

◾ Standardized workflow eliminating the risk of rescan

IMPROVED COST EFFICIENCY

◾ Designed to save space with an eco-conscious design

◾ Enables high patient throughput with standardized workflow, eliminating inconsistencies

◾ Intuitive user experience that results in shorter training requirements

Click HERE to find out more about the new Aquilion Serve

Cost-effective learning for your team

What’s included?

Educational video content divided into short chunks

Illustrations, animations, live action and interviews

Summaries and questions to test your knowledge

Final assessment and CPD certificate

Cost-effective group packages available for your team

Out now:

Radiation safety

Effective communication with your patient in imaging and therapy

MRI essentials

PET-CT essentials

Dose: Practice and principles of optimisation & safety

Requesting imaging safely: For Nurses and Allied Health Professionals

Referring your patient for imaging: For GPs and medical doctors

Coming soon:

CT essentials

Ask for a quote (state number required)

“Getting a group licence proved really cost-effective and was a perfect refresher for our staff. Everyone on the team did the course and really enjoyed it”

Unprecedented AI Presence at ECR 2023’s Show Floor

After a short intermezzo last summer, the European Congress of Radiology returned to its traditional spot early March in Vienna, Austria. Attendance reached its highest level since before the pandemic, with 17,262 participants from 122 countries, a 14 percent increase compared to ECR 2022. In the technical exhibition, which featured over 250 medical imaging companies, two topics clearly dominated the show: artificial intelligence (AI) and sustainability.

By Guido Gebhardt

Companies showed that they are not only concerned with providing their customers with energy-saving systems, but also placing a lot of emphasis on reducing energy consumption along the entire value chain, saving resources, and generating as little waste as possible.

I remember reading somewhere: “Did you know that healthcare is responsible for 4.4 % of global CO2 emissions? That is more than the entire aviation or shipping industry consumes!”

When it comes to AI, where to start? There are so many and varied individual applications. It is best if we just try to classify them. On the one hand, we have workflow AI, which is about speeding up administrative processes and boosting workflow efficiency. Workflow AI starts with making appointments, continues with registration and the filling out of examination-specific forms, and ends with the automated creation of reports. In between, so-called workflow orchestrators ensure

Image: © Sebastian Kreuzberger · ESR

that decisive information is always at the right place at the right time. Each radiologist gets to see the images that he or she can best assess. At the same time, the system makes sure that the radiologists’ work lists don’t become too long, and that patients with clear findings are given priority. All in all, the aim is to individualize, standardize and automate workflows.

Pixel AI is the area that is currently most widely associated with AI: clinical decision support, i.e. the automated recognition of findings. The question arises again and again whether this form of AI will soon replace radiologists. But those who ask this question have not yet fully understood the task for AI. In clinical decision support, AI is used to free radiologists from repetitive, error-prone tasks. Normal findings, which make up the majority of all radiological examinations, can reliably be recognized by AI. This leaves radiologists more time to deal with difficult cases, possibly also together in conferences. Unfortunately, the number of examinations is constantly increasing due to demographic developments. The global population is

aging and growing at the same time. By 2050, there will be more people older than 70 living on earth than the world population in 1950.

Furthermore, AI algorithms can increasingly be found directly on the devices, as showed at ECR 2023. Cameras evaluate the body contours to position the patient correctly inside the CT scanner by just pushing a button, and at the same time suggest the right examination protocol. In MRI, sequences are accelerated and scan times are significantly shortened. Ultrasound systems automatically recognize findings and independently optimize scan parameters. There are hardly any limits to the technology.

The possibilities for mastering the challenges of modern radiology with the help of AI are manifold. But there are still one or two downers: for example, the validation of the different algorithms. This currently takes place exclusively on the basis of clinical studies. It would be desirable to have a technical check similar to the one that is commonly done with X-ray and CT systems: put the test specimen on the table, take a picture or scan, have the image assessed by the algorithm, and that's it.

But this may take a while. Committees across Europe are still positioning themselves and must first agree on systems’ validation. But there is hardly a better time to start with AI than now –even if everything isn’t perfect yet. The path from single slice step-and-shoot to photon counting CT also took several decades.

I’m already looking forward to ECR 2024, with its exciting theme Next Generation Radiology.

https://www.myesr.org/congress

Image: © Sebastian Kreuzberger · ESR

23 Views of ECR 2023

(Re)discover the key messages and products presented at ECR 2023 through our visual narrative.

The automatic scan planning feature on CANON MEDICAL ’s Aquilion Serve CT system is pushing the boundaries in image acquisition, to let clinicians focus on what really matters: patients.

ECR 2023 saw a return to normalcy, with the highest level of activity since before the Covid-19 outbreak four years ago and nearly 17,000 delegates, Mélisande Rouger and Guido Gebhardt reported as they wrapped up the conference, which was held on March 1-5 in Vienna, Austria.

Contrast-enhanced breast imaging was a hot topic at ECR 2023, according to Benjamin Kalender, CEO of AB-CT.

offers an algorithm for lung nodules’ detection with CT. The solution integrates seamlessly into any PACS and is already used in 90 sites across Europe.

AIDENCE

Guido Gebhardt spoke with Cécile Hourquet, Marketing Manager at GLEAMER , about how far the company has come since it was created in 2017.

Lung

Once a clinical institution decides to introduce an AI solution into the workflow, that’s only the start of the journey, Franz Pfister, medical doctor, data analyst and co-founder of DEEPC , told Mélisande

on the first day of ECR

cancer screening featured predominantly in the programme of ECR 2023, Alan Barclay reported after his first day at the meeting.

Mélisande Rouger spoke with Dr. Ankur Sharma, Head of Medical Digital Affairs at BAYER , about its recent acquisition of Blackford Analysis.

The introduction of photon counting CT represents the main development in CT in the past ten years, according to Thomas Kröncke, a Professor of Radiology at AUGSBURG UNIVERSITY in Germany.

DI Europe’s publisher Guido Gebhardt talked with Antoine Jaumier, CEO and founder of INCEPTO , about the company’s approach to AI in radiology.

Rouger

2023.

Guido Gebhardt spoke with representatives of EBIT , ESAOTE ’s IT subsidiary that develops imaging software for radiology departments.

DI

New Dynamic Digital Radiography (DDR) launched by KONICA MINOLTA could be a game changer in clinical practice, according to Sharon Zollo DDR Technology Product Manager.

MINDRAY ’s new Resona R9 Platinum ultrasound system, which was unveiled at ECR 2023, integrates multi parametric ultrasound with qualitative imaging to advance clinical care and lift obstacles to ultrasound widespread use in clinical practice.

ECR 2023 highlighted the clear trend towards enterprise imaging and AI to improve radiologists’ efficiency, Guido Gebhardt reported after conducting several interviews with the industry.

Dr. Andra-Iza Iuga, Radiology Supervisor at Cologne University Hospital in Germany, explained the benefits of working with PHILIPS’ MR SmartSpeed technology.

Europe caught up with Jens Nikelski, Director of Global Sales for contrast media injectors at ULRICH MEDICAL, as the company launched its new CT motion Spicy contrast media injector at ECR 2023.

Mélisande Rouger asked Felix Nensa, Professor of Radiology at ESSEN UNIVERSITY in Germany, how AI and digital tools can help overcome current challenges in healthcare.

Sustainability was a central theme at ECR 2023. To keep abreast of developments, SIEMENS HEALTHINEERS recently appointed a sustainability officer, Dr. Maiya Shibasaki.

Structured reporting remains one of the major topics in radiology, Prof. Wieland Sommer, founder and CEO of SMART REPORTING, told DI Europe at ECR 2023.

Dr. Jusong Xia, CEO and President of International Business at UNITED IMAGING HEALTHCARE , explained how the company addresses the challenges of modern radiology.

A new frontier has been crossed in radiology as a handful of AI-powered ultrasound systems were launched at ECR 2023, Mélisande Rouger, editorial director of DI Europe, reported on the second day of the meeting.

Guido Gebhardt talked to Bhvita Jani, Research Manager at SIGNIFY RESEARCH, to check if her team’s predictions matched what happened at ECR 2023.

Guido Gebhardt talks to Tobias Anger, CTO of TELEPAXX , at ECR 2023 about the new opportunity to integrate AI-based imaging and reporting tools.

Spiral Breast CT: Going Further with Photon Counting

In this article we introduce the innovative technical features of the nu:view breast CT system and describe recent clinical findings and the likely indications of this new equipment. The

Offering significant advantages over other breast imaging modalities thanks to its innovative design and performance characteristics, dedicated breast CT is so far the only 3D breast imaging method that has been shown to reliably detect microcalcifications and mass lesions without any risk of obscuring lesions because of superimposed tissue. Contrast-enhanced imaging with the new technique enables sensitive detection of malignant mass lesions as well as of non-mass lesions. Overall, dedicated breast CT has been shown to be an appropriate modality for a number of clinical indications. Problem-solving is one such application; lesions that are ill-defined by other imaging modalities can easily be identified by 3D spiral breast CT and additional tumor characteristics assessed through contrast enhancement. Another field of application is preoperative staging of established breast cancer cases in order to determine the optimal therapeutic strategy. However, probably the largest potential of dedicated breast CT lies in screening, where currently many women drop out of standard screening mammography programs because of the discomfort and pain associated with

the breast compression involved in conventional mammography. This problem is absent in spiral breast CT, which does not involve breast compression.

Breast imaging is the unavoidable backbone technology in the diagnosis of breast cancer, with the gold standard imaging method currently being digital mammography. Several other imaging modalities such as ultrasound and breast MRI also have well-established roles and recent methods such as breast tomosynthesis are increasingly being used in breast imaging. In addition, the potential of newer imaging modalities such as contrast-enhanced spectral mammography (CESM) or breast CT is under active evaluation to determine their optimal role in the workup of breast lesions.

Recently a dedicated breast CT system, nu:view, developed by the German company AB-CT (Advanced Breast CT) has become commercially available in Europe, after obtaining CE certification in 2018. A prototype of the scanner was installed as early as 2019 in our breast radiology department at Erlangen University Hospital for preclinical studies and since then it has been also used for routine clinical examinations. So far we have carried out more than 200 clinical examinations with the new device, principally for advanced diagnostic purposes.

Current breast cancer diagnostics

The field of breast diagnostics is already supplied with many different imaging techniques, so on the face of it, a legitimate initial question is to wonder whether yet anoth-

er imaging approach is even necessary. However, even a cursory look at the various individual imaging techniques currently used in breast imaging shows that, on their own, none of them can provide optimal performance results for all of the various parameters necessary in modern breast examinations. Such parameters include the obvious ones of sensitivity, resolution and diagnostic accuracy but also encompass other issues as timely accessibility to the imaging modality in practice and, importantly, the acceptance of the technique by the women actually undergoing the examination. In modern breast diagnostics satisfactory results are mostly achieved through a combination of several individual examination modalities.

X-ray based methods: One problem with current x-ray based methods in mammography is the lack of real 3D resolution, which can lead to the masking of lesions because of the superimposition by fibroglandular tissue (FGT). This results in reduced sensitivity in the detection of mass lesions and microcalcifications, especially in dense breasts. Although digital breast tomosynthesis (DBT – a pseudo-3D technique) reduces such superimpositions, it has been found that in DBT microcalcifications appear less suspicious or can be missed completely 1 .

Non x-ray based methods: Ultrasound is very operator-dependent, and MRI is very time-consuming, costly and therefore not available everywhere. In addition, MRI and ultrasound cannot reliably identify microcalcifications, which can be the only imaging sign of breast malignancies.

These points highlight the need for a new 3D breast imaging modality that can reliably visualize both mass lesions and microcalcifications. The dedicated spiral breast CT meets these requirements and may therefore serve as good candidate for future breast imaging.

Technical characteristics of dedicated spiral breast CT

For the reliable visualization of microcalcifications, a spatial resolution of less than 0.2 mm is necessary. This is possible with the nu:view scanner thanks to its dedicated x-ray source and a newly designed single photon counting detector. The source has a focal spot size of 0.3 mm and is operated at a constant tube voltage of 60 kV. The detector has a pixel size of 0.1 × 0.1 mm and uses cadmium telluride crystals which directly convert x-ray photons into measurable electric charges, thus decreasing intrinsic noise and enabling the use of low radiation exposures. When imaging with a radiation dose equivalent to that which is typically used in conventional mammography, the dedicated spiral breast CT system provides images with a spatial resolution as low as 0.15 mm at a signal-to-noise ratio amply sufficient for image interpretation 2, 3. During a single scan of the imaging unit, up to 12,000 projections are acquired by the detector (1,000 Hz) and the total acquisition period lasts only 7 – 12 seconds, thus minimizing the risk of motion artifacts.

Similar to breast MRI, the neovascularity of tumors can be visualized by intravenous administration of contrast agent. Not only can mass lesions be better visualized by contrast media uptake, but the visualization of non-mass lesions is even possible [Figure 1]. This makes breast CT the only modality in breast diagnostics

Image: Universitätsklinikum Erlangen

The team of the multimodal breast imaging unit of the University Hospital Erlangen have accumulated extensive clinical experience with the new nu:view scanner from AB-CT – Advanced Breast-CT.

that can visualize both imaging characteristics of ductal carcinoma in situ (DCIS), namely microcalcifications [Figure 2], and non-mass enhancement.

Possible indications for Breast CT

SCREENING: Mammography screening has been shown to reduce cancer related deaths in breast cancer and is currently key for the establishment of early diagnosis. However, the level of the participation of women in screening programs leaves much to be desired. For example, the German breast screening program, only 49 percent of all eligible women actually participated in the program in 2017 4. One reason for this is the discomfort experienced by women during the mammography examination itself and in particular with the breast compression, which causes real pain for many women 5 Dedicated breast CT does not involve compression of the breast and uses a completely different body positioning from the classical upright stance of mammography. In breast CT, the woman lies prone on a patient table with the breast hanging freely in the imaging unit through an aperture in the patient table. Understandably, this procedure has been proven to significantly increase patient comfort compared to mammography by eliminating compression-related pain 6. Thus, dedicated spiral breast CT is an appropriate methodology for opportunistic breast cancer screening particularly for women who would otherwise decline participation in or drop out of conventional mammography screening. The use of dedicated breast CT for opportunistic screening has already been described and is routinely carried out at the University Hospital Zurich 7 .

PREOPERATIVE STAGING: Pre-therapeutic staging of breast cancer is important to determine optimal therapeutic strategy and whether breast conserving therapy is possible. However the determination of tumor extent, satellite lesions or DCIS can be challenging, especially in dense breasts. Breast MRI has been shown to change the previously determined extent of planned surgical procedures in up to 23 percent of cases mainly due to the identification of additional lesions 8. Like MRI, dedicated spiral breast CT shows malignancies as enhancing lesions, and has thus proved to be an excellent alternative to breast MRI, especially in situations where MRI is not readily available or in cases with contraindications to MRI examination [Figure 3]. However, it should be noted that for lymph node staging with breast CT, an additional ultrasound examination is necessary, since breast CT does not

Image: Universitätsklinikum Erlangen

cover the axilla.

Figure 1. Images from a 59-year old woman with bloody nipple discharge without any correlating lesion in two-view mammography (A, B). Contrast-enhanced dedicated spiral breast CT revealed an enhancing non-mass lesion in the lower outer quadrant (circled in image C), which was confirmed as a high-grade ductal carcinoma in situ (DCIS) on histopathology.

Figure 2. Mammography and breast CT images from a 79-year old patient with microcalcifications in the outer quadrant in the right breast.

(A) mammography in CC-view; (B) magnification of image A; (C) dedicated spiral breast CT image and (D) magnification of image C. Histopathology confirmed ductal carcinoma in situ (DCIS).

PROBLEM-SOLVING: Currently breast MRI is the modality frequently used for problem solving in breast imaging cases where analysis is difficult and precise diagnosis not yet established. Likewise, dedicated spiral breast CT can be a method of choice in clinical decision-making in cases with equivocal lesions [Figure 3]. With an advantage over MRI in that the examination takes less time, breast CT provides 3D-imaging without the need for breast compression and results in improved localization of lesions in the breast without distortions. Through contrast-enhanced imaging, dedicated spiral breast CT could aid in discriminating probably benign from malignant lesions and be an aid in the consideration of whether to take a breast biopsy or not. Breast CT also serves as a valid alternative to breast tomosynthesis or contrast enhanced spectral mammography, with in addition the advantages of real 3D imaging and increased patient comfort.

Conclusion

Dedicated spiral breast CT is a new imaging modality with striking and proven performance characteristics; the technique has already found its way into daily clinical practice. It is not only suitable as an imaging option in tertiary care institutions, but also for use in radiology practices who want to offer their patients an alternative for breast screening and as a powerful problem-solving modality.

https://www.uk-erlangen.de

REFERENCES

1 Poplack SP, Tosteson TD, Kogel CA, Nagy HM. Digital breast tomosynthesis: initial experience in 98 women with abnormal digital screening mammography. AJR Am J Roentgenol 2007;189(3):616–623. http://dx.doi.org/10.2214/AJR.07.2231

2 Wetzl M, Wenkel E, Balbach E, Dethlefsen E, Hartmann A, Emons J, Kuhl C, Beckmann MW, Uder M, Ohlmeyer S. Detection of Microcalcifications in Spiral Breast Computed Tomography with photon counting Detector Is Feasible: A Specimen Study. Diagnostics (Basel). 2021 May 9;11(5):848. http://dx.doi.org/10.3390/diagnostics11050848

3 Germann M, Shim MS, Angst F, Saltybaeva N & Boss A. Spiral breast computed tomography (CT): signal-to-noise and dose optimization using 3D-printed phantoms, Eur Radiol (2021) 31 (6):3693-3702. http://dx.doi.org/10.1007/s00330-020-07549-3

4 Jahresbericht Evaluation 2017. Deutsches Mammographie-Screening-Programm. Kooperationsgemeinschaft Mammographie, Berlin, Oktober 2019. https://fachservice.mammo-programm.de/download/evaluationsberichte/ Jahresbericht-Evaluation-2017.pdf (accessed July 2nd, 2022)

5 Rutter DR, Calnan M, Vaile MS, Field S & Wade KA Discomfort and pain during mammography: description, prediction, and prevention BMJ. 1992 Aug 22; 305(6851):443-5. http://dx.doi.org/10.1136/bmj.305.6851.443.

6 Wetzl M, Wenkel E, Dietzel M, Siegler L, Emons J, Dethlefsen E, Heindl F, Kuhl C, Uder M & Ohlmeyer S. Potential of spiral breast computed tomography to increase patient comfort compared to DM. Eur J Radiol. 2022 Dec;145:110038. http://dx.doi.org/10.1016/j.ejrad.2021.110038

7 Interview, Prof. A Boss Diagnostic Imaging Europe 2021; 37(2): 16

8 Perono Biacchiardi C, Brizzi D, Genta F, Zanon E, Camanni M & Deltetto F. Breast cancer preoperative staging: does contrast-enhanced magnetic resonance mammography modify surgery? Int J Breast Cancer. 2011; 2011:757234. http://dx.doi.org/10.4061/2011/757234

Figure 3. Example of the potential of breast CT for both problem-solving and also preoperative staging in a single examination. 63-year old woman with a palpable breast carcinoma in the lower inner quadrant of the right breast (indicated by stars in A, B, C, E). Mammography and ultrasound revealed a second lesion in the lower outer quadrant of the right breast (arrowhead in A, B, D), which was suspected as hamartoma. Dedicated breast CT confirmed the benign nature of the lesion through the lack of contrast enhancement (arrowhead in E). Incidentally, breast CT revealed a second contrast-enhancing lesion in the lower inner quadrant of the right breast (arrow in E), which proved to be multifocal cancer that had been missed on ultrasound and in initial mammography (arrow in A).

Broad Solutions & Platforms to lead Radiology AI Development

Platforms that host various AI applications and solutions focusing on a whole range of diseases offer much promise for medical imaging AI, according to Dr. Sanjay M Parekh, Senior Market Analyst at Signify Research, who shared his insights and forecasts for the booming market.

By Mélisande Rouger

◾ What are the current trends on the medical imaging AI market?

The market is still nascent, but it’s growing fast with many developments. Radiologists better understand the technology and the potential it offers, and as such are becoming more discerning when considering AI. The question has shifted from “do I need AI” to “how is AI going to benefit me?” Vendors are responding, and this is reflected in the product developments and broader strategies. It’s not just about offering an algorithm, but rather delivering an AI solution that offers greater value to the end user. For lung cancer, for example, having

an algorithm that detected nodules was enough to garner attention several years ago. Now, vendors need to offer a more complete AI solution, which in addition to detection also measures the nodule size and volume, as well as provides characteristics of the nodule, such as the risk of it being malignant, to improve a patient’s prognosis. Such solutions offer great-

Dr. Sanjay M Parekh, PhD, is Senior Market Analyst at Signify Research, an independent supplier of market intelligence and consultancy services to the global healthcare technology industry. Dr. Parekh’s areas of expertise include medical imaging and AI, and he leads Signify Research’s coverage of the medical imaging AI market, including its flagship AI in Medical Imaging Market Intelligence Service, and AI in Medical Imaging Company and Product Database. Dr. Parekh has published numerous market research reports, presented in several webinars, and authored many articles delivering thought leadership related to the use of AI in radiology.

er value to a clinician than detection alone and are therefore more likely to be considered for purchase.

Another approach that some AI vendors have adopted is to mimic and replicate the way radiologists work. Rather than focussing on a particular aspect or finding, these AI solutions can identify a myriad of findings from a single scan.

Yet another approach is to create a workflow around a particular condition, for example stroke. Some vendors have done so by creating a solution that addresses the broader workflow around the condition, improving the turn-around-time from detection to treatment.

Another important trend in the medical imaging AI market is AI platforms. These solutions address the last-mile challenges faced by radiologists, including the back-end deployment, front-end integration, and orchestration of AI solutions. AI platforms are paramount as radiology AI scales, enabling efficient orchestration to deliver the image to the right application for analysis, before delivering the results back to the radiologist in a timely and seamless manner. Although the traction around AI platforms remains limited, their value proposition will increase as AI begins to scale and is increasingly adopted in radiology.

◾ Which solutions have had more traction so far?

Many types of AI solutions have had differing success to date. AI solutions for stroke imaging, particularly those with a broader workflow component, have realized very good traction to date. AI solutions for fractional flow reserve (FFR), measuring blood flow in the coronary arteries

from a CT, have also been very successful, including receiving regulatory approval and even reimbursement in multiple regions. Such is the success of these vendors that many AI vendors from China have also developed such solutions as part of their expansive product portfolios, and several of which have received NMPA (China) clearance.

AI solutions for breast imaging –primarily screening – have also benefitted from commercial traction, although AI solutions for digital breast tomosynthesis (DBT) will become increasingly popular compared to AI solutions for 2D mammography, due to the complexity and time it takes to read DBT images. Further, as DBT is increasingly adopted across Europe for screening, the demand for such solutions will also increase.

Chest imaging is another clinical segment that has benefitted from AI, and comprehensive AI solutions, whether for chest X-ray or CT, have gained the most traction due to the value they confer, includ-

ing identifying incidental findings beyond the primary read.

◾ What is the market worth today?

As per our latest report published in July 2022, the medical imaging AI world market is forecast to reach $ 1.4 Bn by 2026, up from an estimated $ 400 M in 2021. However, several barriers need to be addressed for this potential to be realized. The lack of reimbursement, including the limited number of CPT codes, remains one of the most significant barriers to date. Other obstacles include the lack of real-world clinical validation demonstrating the generalizability of AI solutions, and lack of health economic studies demonstrating the return-on-investment (ROI) of AI solutions. Furthermore, other global headwinds and the deadline for the European Union Medical Device Regulation (MDR) may also hold back this market.

However, as outlined above, progress is being made thanks to continued investment into the market. Medical imaging AI companies

Image: © Who is Danny stock.adobe.com

The AI market will start to mature as artificial intelligence becomes more ubiquitous in radiology. Further growth may be spurred on by tailwinds, including new CPT codes, and reimbursement for a greater range of tools.

have raised close to $5Bn in venture capital funding since 2015, which is a very healthy amount.

The challenge remains for vendors to generate commercial traction by delivering a ROI for investors. A significant proportion of this funding is skewed towards companies from China. However, in contrast to the US, for example, the revenues generated by these companies for their AI solutions remains low.

◾ Where do you see the market going next?

In the short term, we will see a greater influence on the market from companies from China. As the country increasingly adopts AI, but remains relatively closed to nonnative vendors, it provides a great opportunity for native vendors to realize commercial traction. Other regions where commercial traction will also grow, despite some regions lacking the necessary infrastructure to adopt AI, include India and Brazil. In the mid to longer term, this market will start to mature as AI becomes more ubiquitous in radiology. However, further growth may be spurred on by significant tailwinds in the market, including new

CPT codes, and reimbursement for a greater range of tools, especially beyond the US.

Investment in the market will taper off as investors become more discerning and place a greater emphasis on companies which they perceive to offer greater value to providers or are likely to receive reimbursement for their solutions. When reimbursement in medical imaging AI becomes more commonplace, it will likely prompt a renewed enthusiasm from investors given the defined return-on-investment that such companies will benefit from. However, this is unlikely in the near term.

◾ Has the Covid-19 pandemic had an impact on the market?  Covid-19 did not have lasting impact, but it has modestly slowed the pace of commercialization of some solutions. Some countries such as the US were far less affected compared to other regions, e.g. Asia, where more stringent lockdowns persisted for a longer period. However, one of the positive effects of the pandemic was highlighting the broader need for digitalization, including the need for the healthcare industry to

follow and adopt newer technologies such as AI at a faster rate.

◾ Bayer has recently acquired Blackford Analysis. Why are pharmaceutical companies taking an interest in AI?

There has been a growing interest from companies outside of algorithm developers in the AI market, including pharmaceutical companies. One of the reasons for this is to mitigate the risk of contrast agents being less used in radiology. Another is that such companies may want to expand their reach within this market, especially as commercial traction ramps up. As I highlighted in the Signify Premium Insight addressing this topic, the move will ultimately enable such vendors to be well positioned as demands in radiology change (e.g., less use of contrast agents, greater adoption of AI), rather than playing catch-up later.

Other similar developments include Guerbet investing in Intrasense – with the view of potentially acquiring it –, and Tempus acquiring Arterys.

https://

According to Signify Research, the medical imaging AI market is forecast to grow to almost $1.4 billion by 2026, from an estimated $ 400 m in 2021.

Image: Signify Research

Weaving AI into the Emerging Strategic Decisions in Healthcare

After the Covid-19 crisis, healthcare in Europe is being reconfigured. That momentum has already started. These policy initiatives should not be strategic policy level exercises separate from the strategic changes related to the technology agenda, writes Rafael Bengoa, an expert in management and public health, who kindly accepted our invitation as a guest columnist.

By Dr. Rafael Bengoa

It is well known that all European countries follow a model focused on rescue in the management of acute episodes. This model addresses care for acute diseases with quality, but it must be complemented with a care model focused on chronicity, which is responsible for 90 percent of deaths in Europe and for 70 percent of the demand in healthcare. This occurs both in Bismark-type insurance assistance models (Germany and France,

Belgium) and in tax-based NHS-type models (Spain, UK and Italy).

Simultaneously to this strategic redefinition of healthcare, artificial intelligence (AI) emerges showing great potential. So it is convenient to understand the complementary potential of the technology to achieve these changes, as well as to understand how the process of decision-making and investment across the EU can weave AI into the changes that will happen in these next few years.

In Europe, best practice examples for the use of AI in healthcare can be identified, but in general, the services are not yet oriented to use AI to proactively detect health or social problems. In general, health systems are more passive than proactive. They are frequently faced with the need to react to medical crises in patients whose pathologies may have been effectively managed initially, but whose therapeutic regimes no longer function optimally.

Image: © greenbutterfly stock.adobe.com

Data management powered by AI would allow for a proactive approach in the early detection of pathologies when patients are home, before they arrive to the emergency room and occupy a hospital bed. All European systems share the challenge of transitioning towards another, more integrated and proactive model, providing clinical services in an innovative way and, in many cases, turning the home into a delivery hub.

AI can be a tremendous ally in this effort to detect problems early. AI tools are already being used to monitor patients in home care models, and within healthcare infrastructures, they enrich diagnostic imaging tests in screening for the most common cancers, such as breast or lung cancer. In this decade, all countries will tend to shift their system from reactive to proactive, by identifying situations before they become crises.

The transformative effect of AI and the productivity paradox

It is necessary to use the potential of AI to transform care models and not to root the current models. This approach will be particularly important given the arrival of post-pandemic recovery funds in Europe. It is a strategic moment that will not happen again.

Many of these funds will have to be directed to AI projects in the healthcare sector and many decision-makers will demand more evidence for such investments in AI. In this context, it is worth noting the “productivity paradox”.

The expert panel – The Watcher Review, 2016. UK – indicated that, although it is natural to ask for a

Dr. Rafael Bengoa is Co-founder and Co-director of the Institute for Health and Strategy (SI-Health) in Bilbao, Spain.

He is Vice Chairman of the Horizon 2020 Advisory Group on Health, Demography and Wellbeing at the European Commission, and a senior leadership fellow at Harvard University, Boston, USA.

Dr. Bengoa served as Director of Health Systems at the World Health Organiza-

short-term return on investment for investments in information technology, experience shows that this short-term return comes more in the form of quality improvements and clinical safety than in terms of raw financial returns.

In fact, investment savings can only be seen ten years later because of what is called the “productivity paradox” – the observation that, as more investment is made in informa-

tion (WHO) and Minister of Health and Consumer Affairs for the Basque Government in Spain. More recently, he has been advising national governments of Northern Ireland, Spain and the United States of America on health policy.

Dr. Bengoa trained in medicine & surgery and healthcare management in Spain and the UK.

tion technology, worker productivity may go down instead of up.

This has become evident in other sectors of the economy, in which the digital push has been key to rise to the challenges of the 21st century. This strategic effort does not imply that all the investment needs to come from the public sector, but rather via the identification of public and private partnerships that could add value to these policies.

Image: Dr. Rafael Bengoa

“We already have many carefully selected applications from leading AI vendors worldwide fully integrated into our department’s existing IT infrastructure.”

deepc secures € 12 M funds to expand its AI Platform, partners with LMU

Clinical adoption of artificial intelligence (AI) in medicine is rapidly increasing worldwide. Platforms are giving healthcare systems the flexibility to adopt and integrate third-party AI applications and solutions into existing workflows and ultimately improve patient outcomes.

Prof.

Clemens Cyran, Vice Chair at the Department of Radiology at LMU

Germany-based AI platform provider deepc has recently raised a € 12 million Series A funding round led by Sofinnova Partners, a venture capital firm in life sciences that specializes in healthcare and sustainability. The new capital will be used to propel the commercialization of the deepcOS AI operating system, and to further develop streamlined services to improve user experience for radiologists. Bertelsmann Investments and existing investor Winning Mindset Ventures also participated in the round.

deepc’s core technology is a cloud-native, vendor-neutral platform that integrates seamlessly with existing radiology workflows. It provides approved third-party AI vendors with an efficient way to commercialize their solutions while offering hospitals and clinics a secure, “one-stop shop” for their AI-assisted radiology needs.

Partnership at the LMU University Hospital

Munich-based LMU University Hospital has been working on AI solutions for clinical practice and research applications to continuously improve clinical workflows and patient care. For its radiology department, the hospital has entered a close strategic partnership with deepc to systematically

apply AI in clinical practice on a broader scale. Through the innovative AI platform, LMU radiologists will gain effortless access to numerous regulatory-cleared AI solutions. Medical workflows can thus become more efficient, and patient treatment quality can be further increased. Prof. Clemens Cyran, Vice Chair of the Department of Radiology at LMU, is already planning further. “In the future, we would like our colleagues to be alerted in realtime about critical emergency findings that the AI detects in the image data to support prompt patient care,” he said. “We will also focus on the deep integration of various solutions into the existing clinical workflows, such as the automatic transfer of results into report texts or integration with the RIS / PACS,

within the framework of work list prioritization.’

With its two sites, LMU University Hospital Munich is one of the largest hospitals in Europe, treating over 500,000 patients annually. LMU’s radiology department pursues a clearly defined strategy to personalize and optimize diagnostics-guided therapy, which requires using AI in daily practice.

LMU radiologists have recently started working on routine diagnostic operations with the deepcOS AI platform developed by the Munichbased medtech company deepc.

“We believe that AI has the power to revolutionize the way we approach healthcare,” said Dr. Franz Pfister, Co-Founder and CEO of deepc. “Our AI platform is a game changer for radiology, allow-

ing for faster and more accurate diagnoses, and ultimately improving patient outcomes. We are thrilled to welcome Sofinnova Partners and Bertelsmann Investments on board. Their extensive networks in the healthcare ecosystem will propel deepc forward.”

The new funds will expand deepc’s geographic coverage and help the company develop additional features within its solution. Radiology teams will benefit from new AI-assisted options that are seamlessly integrated into their workflows. At the same time, AI software vendors will get low-friction access to a large number of potential customers, the company wrote in a communiqué.

“Our investment in deepc underscores our conviction that digital

“ The new capital will enable the company to develop new features and expand commercialization of deepcOS, its proprietary operating system that powers AI integration within radiology workflows.”

medicine holds enormous promise and that enabling technologies will be at the forefront of the digital revolution in healthcare, precisely the focus of our digital medicine investment strategy,” said Simon Turner, Partner at Sofinnova Partners. deepcOS enables wide adoption of AI into a major part of the patient journey. The team has demonstrated the benefit that they can provide to both healthcare professionals and patients.”

https://www.deepc.ai

Julia Moosbauer, COO and Dr. Franz Pfister, CEO of deepc

Image: deepc

Bayer reinforces Commitment to Innovation in Radiology

In healthcare, and particularly medical imaging, innovation powered by artificial intelligence (AI) is needed more than ever. Dr. Konstanze Diefenbach, Head of Research and Development at Bayer, and Dr. Ankur Sharma, Head of Medical Affairs, Digital Radiology, explained to DI Europe how Bayer is helping to shape innovations to benefit patients and their treating physicians.

AI applications for medical imaging improve prioritization, lesion detection, quantification and productivity. Bayer‘s new cloud-hosted platform Calantic provides access to apps that integrate into standard medical imaging workflow.

„Medical imaging plays a key role along the entire patient journey, in facilitating diagnosis, treatment decisions and monitoring the patients’ response to treatment,“ said Dr. Konstanze Diefenbach, Head of Research and Development for Radiology at Bayer. „Increasingly, radiology departments are under enormous pressure with a significant rise in the medical imaging data they need

to assess, and innovative solutions are urgently needed.“

Pressure and workload are increasing along with the amount and complexity of imaging data triggered by technological advances and the growing demand for imaging procedures, driven by aging populations and changing lifestyles.

In this context, AI has the potential to be part of the solution with the value proposition to support

Developed by radiologists for radiologists, Calantic takes previous fragmented and unrelated app solutions and coordinates them into a single convenient marketplace. Vetted with Bayer‘s expertise, to ensure the feasibility and reliability of third-party solutions.

Image: Bayer

diagnosis and increase the throughput of radiological examinations.

Serving as a springboard for innovations, radiology AI is the fastest growing market segment within the overall global radiology market. Radiology AI is expected to continue growing dynamically, with a compounded estimated annual growth rate of more than 26 percent through 2026. As AI is making its way into clinical radiology practice, it is crucial that it gains the trust of experts.

Prof. Roland Wiest, Deputy Director of the Institute of Diagnostic and Interventional Neuroradiology at Bern University in Switzerland, explained the importance of education to help drive adoption of AI in clinical practice.

„Appropriate training is key to successfully govern AI,“ he said. „Radiologists, not AI, remain in charge for patient care and need to acquire new skills to do their best for patients in the new ecosystem.“ Prof. Wiest is leading several educational initiatives aiming at driving medical, engineering, and practical capabilities in radiology to support the wider use of AI-based technologies.

Dr. Ankur Sharma, Head of Medical Affairs, Digital Radiology at Bayer is convinced that ‚radiology has always been at the forefront of digital innovation, and artificial intelligence in particular, has vast potential in medical imaging. Bayer‘s long established medical expertise and commitment to innovation across a multitude of diagnostic and therapeutic areas is what we will bring to the AI space in radiology, as a point of significant focus. „With Calantic our aim is to help shape the future of radiology, transform treatment outcomes, and patient care.“

Aquiring Calantic, a suite to help radiologists

With its introduction of Calantic Digital Solutions, Bayer underlines the leadership in key radiology segments and its deep medical understanding across a multitude of diagnostic and therapeutic areas.

Calantic Digital Solutions is a suite of digital radiology AI-enabled applications which assist radiologists and their teams at critical steps within a patient‘s treatment journey. The vendor-neutral, cloud-hosted platform includes a growing number of applications designed to prioritisation, lesion detection and quantification as well as apps that automate routine tasks and measurements, improve the radiology suite‘s workflow, and free up time for radiologists and their teams.

„Our aim is to help shape the future of radiology, transform treatment outcomes, and patient care.“

The offering is orchestrated by body region, and procedure and initially focused on thoracic and neurological diseases, such as pulmonary nodule detection and triage of potential intracerebral hemorrhage (ICH) and large vessel occlusions (LVO). ICH and LVO can be associated with stroke. Bayer recently announced to partner with app developers Quantib, ClariPi, EXINI Diagnostics, Mediaire, Coreline, and ScreenPoint Medical to broaden the already available options on the platform, by including tools which aim to aid in neuro, cardiac, prostate and breast imaging. First launch markets include the US and several European countries.

https://www.bayer.com

„Increasingly, radiology departments are under enormous pressure with a significant rise in the medical imaging data.“

Dr. Konstanze Diefenbach, Head of Research and Development at Bayer

Dr. Ankur Sharma, Head of Medical Affairs, Digital Radiology at Bayer

Image: Bayer

Image: Bayer

Elevating Efficiency with AI

Prof. Georg Langs teaches at the Department of Radiology and Nuclear Medicine‘s Computational Imaging Research Lab at Vienna Medical University. He is also one of the founders of contextflow, a company that develops artificial intelligence (AI) software.

Guido Gebhardt spoke with the scientist about the integration of AI in RIS and PACS.

By Guido Gebhardt

◾ Professor Langs, how do you assess the current situation of artificial intelligence in radiology?

While there was initially a lot of enthusiasm for solutions that addressed only one specific problem, it soon became clear that radiology practices or clinics cannot and do not want to deal with procuring, integrating and maintaining 30 different algorithms for different application scenarios. At the same

time, there is still a need to use AI to improve diagnosis, prognosis and treatment decisions. Very exciting developments are taking place in the field of predictive models and integration of diagnostic data. In practice, users are asking for comprehensive integrated solutions or platforms or marketplaces that make it easier for them to install and integrate several AI algorithms, even from different providers. We can already see that, in addition to the classic providers of

AI algorithms for different organs, companies are also establishing themselves that deal with the seamless integration of algorithms in RIS and PACS. My feeling is that we are currently in a start-up phase. At the moment, the first sales are coming in and I am curious to see how the market will look in two years and how the landscape between AI platforms and individual integrations will develop by then. The race for the best AI algorithms is also far from decided – there will still be a lot going on. I assume that comprehensive solutions that cover entire areas and improve diagnostic possibilities in addition to efficiency will prevail.

contextflow is a spin-off of the Medical University of Vienna (MUW) and the European research project KHRESMOI, which is supported by the Vienna University of Technology (TU). The company was founded in July 2016 by a team of AI and engineering experts

and has already received numerous awards; only recently contextflow was recently selected by GE Healthcare Canada for the Edison AI Orchestrator Accelerator. ADVANCE Chest CT is CE marked and available for clinical use in Europe under the new MDR.

◾ There are supposedly over 200 providers of AI solutions for radiology and that number is rising. Will the market consolidate soon?

From the radiologists‘ point of view, it is not directly important how many providers there are at the same time. In my opinion, the offer will be reduced to a few in the future. At the same time, their portfolio will

„The first approach of machine learning was the prediction and segmentation of findings. Tumors and certain disease patterns should be recognized. But now we want to know if there are other patterns in the image data that have predictive value for diseases. And the question is: Can I go beyond what I know so far and expand my diagnostic vocabulary for better predictions?“

Prof. Georg Langs, Co-founder of contextflow

◾ This means that radiologists have to participate more intensively in integration than they are used to with modalities?

increase in order to cover the widest possible range of applications, because the installation and integration costs of individual solutions should not be underestimated.

◾ How do you think the market will continue to develop?

When we talk about artificial intelligence in radiology, we are not just talking about an algorithm that delivers measurements. The added value of AI solutions also depends on how they are integrated into the workflow: from registration to sending findings and their use in interdisciplinary case conferences. This is about interfaces and how the individual measured values of

the highly developed algorithms are transferred via the RIS and the PACS to the structured findings and then to the referring physicians. I see enormous development potential in this. The competition will take place in the quality of the AI algorithms and the integration into the workflow, all of which requires joint iterative development work with the radiologists. In order to achieve a gain in efficiency and to exploit the diagnostic contribution of imaging to treatment decisions, one has to deal intensively with the user interface in addition to the medical questions and involve all process participants in the workflow integration.

Yes, that is what it means. Not only radiologists are affected by data processing, but also referring physicians and other disciplines – the keyword being integrative diagnostics. If all players in the healthcare system are networked in the future, the data flow in radiology will begin with the registration, continue via the modality and the AI-supported reporting into RIS and PACS and from there in the form of a structured report of findings to the referring physicians or the physicians who continue treatment. At the same time, the integration of information from disciplines such as radiology, pathology or laboratory medicine also plays a key role here. The added value of such solutions lies in the seamless processes and the gain in information. All persons involved in the treatment process get immediate access to quantitative values that are decisive for further treatment and the success of the therapy. The challenge is therefore to handle the data correctly and to present the information in such a

way that it offers added value. This means that all those involved in the process chain should really engage with artificial intelligence.

◾ As an AI provider, how does contextflow ensure that their data, once it is integrated from the platform into the PACS, is presented correctly, as users would like it to be?

So far, no matter how the integration is carried out, the team at contextflow spends a lot of time with the radiologists to coordinate and optimize the solution – this is a culture that has certainly been taken along from its origins as a spin-off of a medical university and university hospital. It doesn‘t matter whether it‘s a single integration or a digital marketplace. There is also an ongoing very close cooperation with all integration partners, such as PACS manufacturers, to jointly optimise the integration into the routine. The aim is not only to show the user isolated findings, but to understand exactly where and when which

information makes a contribution to the diagnosis and further decisions.

◾ What do you think is the right time to start with AI? Should you start right away or is it better to wait?

In my opinion, now is a good time to enter, as one is still involved in the further development. At the moment it is being decided where the journey will go. Although almost all manufacturers have mature products, the interesting thing now is the fine-tuning and further development – to further improve the solutions, to shape the role of AI and to optimally adapt it to the respective application on site. And that is happening right now.

◾ What about the validation of algorithms in the future? Do users want to know why and how the algorithm produces this or that result?

I think the demand for validation of AI solutions is important and right. Algorithms must of course be vali-

dated. This takes place within the framework of benchmarks and is proven by numerous scientific studies. At the same time, it is important that the results can be explained. Comparative cases, for example, play an important role here, as they make the reasoning transparent. A key point here is that manufacturers communicate very clearly what the software can and cannot do. Users confirm that they not only work more efficiently with the help of AI, but also achieve better results because, for example, observations become quantifiable. Quite independently of this, all manufacturers must approve their respective systems as medical devices within the framework of the MDR (Medical Device Regulation). Another important aspect is the feedback loops that manufacturers need in order to con-

Evaluating chest CTs can be a complex and timeconsuming process. Advanced Chest CT AI solutions give qualitative and quantitative insights to objectively report on suspected lung cancer, ILD, and COPD cases.

Image: contextflow

tinuously develop the systems and improve their integration into clinical routine together with the users. AI algorithms become more accurate the more they are trained on real and diverse data. Manufacturers need feedback from confirmed examination results from everyday use to improve the algorithms and adapt them to changing care –there is a lot happening in the field of continuous learning.

◾ Are there any prerequisites that have to be fulfilled if I, as a radiologist, am involved in the integration of AI algorithms?

As already described, one must think about the use of AI beyond one‘s own departmental boundaries and include those who collect

diagnostic data as well as those who are supplied with data and information by radiology in decision-making processes. Technically, manageable computing power is sufficient to create interfaces between the individual IT systems. The situation is different when it comes to the requirements for automated image analysis. With on-premise solutions, it is advantageous to have a powerful GPU, but cloud solutions can be more efficient overall and also become practicable with the establishment of data protection framework conditions. With the integration of AI, we are in principle already preparing for the next development step: Integrated Diagnostics. This involves intensive, intersectoral, interdisciplinary and probably also

international exchange, and the integration of images and findings that we already know from tumor boards, together with values provided by AI. To achieve this, there is still work to be done within the hospitals and also medicine in terms of digital infrastructure and processes. This joint development at the interface between medicine and AI is, in my opinion, one of the most exciting aspects of the current development. In practice, it is about harmonising technologies and processes to simplify the integration of IT systems. Together with interdisciplinary research, this enables the further development of AI in medicine and to tap into the enormous development potential.

Photon Counting CT

A new Challenge for Contrast Administration?

By Alan Barclay

The recently introduced photon counting CT (PC-CT) system has been widely acclaimed thanks to its significantly improved performance characteristics, such as higher spatial resolution, improved contrast-to-noise ratio, markedly reduced radiation exposure and the possibility of performing spectral imaging. Despite these considerable advantages, administration of an iodinated contrast medium is still mandatory, and adaptation of contrast injection protocols to optimize iodine delivery while minimizing waste remains important.

At the recent ECR meeting, Bracco sponsored a symposium, chaired by Prof. T Kröncke, on the impact of PC-CT on diagnostic imaging and on cardiac and non-cardiac applications in particular, and addressed the implications of the new technology on contrast medium usage

This article summarizes the principal points presented at the symposium.

Photon

Counting CT:

a Technical and Diagnostic Advance

Prof. Kröncke began his presentation with a reminder of the enthusiastic press release issued nearly two years ago by the FDA on the occasion of the 510 (k) clearance of the new photon counting CT device, the Naeotom Alpha photon counting CT system from Siemens Healthineers. The FDA proclaimed that new system was “the first major new technology for computed tomography imaging in nearly a decade,” a comment all the more remarkable in that the normally staid regulatory body only rarely issues press releases and even then usually in a flat, factual style. However, the fact is that, since then there have been more and

more studies and reports from users of the new system, which confirm that the FDA’s enthusiasm for photon counting CT is justified. The main objective of the speakers at the Bracco-sponsored symposium at ECR was to share the results behind the enthusiasm for the new technology and the implications for contrast media. Prof. Kröncke reminded the audience that technological innovations can be categorized in three classes, namely incremental innovations; radical innovations (which involve breakthrough developments) and finally disruptive innovations which can be counted as

Prof. Thomas Kröncke, Chair of Department of Radiology, Augsburg University Hospital, Germany

Image: Augsburg University Hospital

compared to “conventional” or energy-integrating detection (EID).

A: EID. In a two‐step conversion process, the absorbed X‐rays are first converted into visible light in a scintillation crystal. The light is then converted into an electrical signal by a photodiode attached to the rear of each detector cell.

B: In photon counting detection X‐ray photons are directly transformed into electrical signals in the Cadmium Telluride semi-conductor. With a photon counting detector being able to count the charges created by individual X‐ray photons as well as measuring their energy level, the detector has intrinsic spectral sensitivity in every scan. Images courtesy of Siemens Healthineers

game-changers. Prof. Kröncke invited the audience to decide into what innovation category they would place photon counting CT after they had heard the presentations. Most of the audience could guess that Prof. Kröncke had already decided (game-changer!).

Prof. Kröncke then went on to briefly describe the principles behind the new photon counting detector (PCDs) and its comparison with the conventional or energy-integrating detectors (EID) used in standard CTs [Figure 1]. There is a lot of high-quality physics, engineering design and manufacturing know-how involved in photon counting detectors, but in practice

the result of the technology has several advantages for the user, namely: PCDs are more sensitive than EIDs, so enabling reduced radiation doses; they give higher spatial resolution; in addition since PCDs measure the energy of individual X-ray photons, they open up the possibility of multi-energy imaging and material decomposition. Prof. Kröncke then illustrated these with practical examples. For spatial resolution, he showed PCD images of small skeletal structures in the mouse which had clearly improved resolution compared to conventional CT. As for multi-energy resolution or spectral iodine enhancement, it is well known that when lower keV

levels are used, for example in CT angiography, improved contrast can be achieved; the advantage the PCD system offers is that the precise energy level for the optimal creation of virtual mono-energetic images (VMI) can be easily determined. Recent studies have shown that low keV VMI reconstructions of run-off computed tomography angiography (CTA) scans on a PC-CT system result in substantially higher signal to noise ratio (SNR) and contrast to noise ratio (CNR) than 80 kVp and 100 kVp EID CT acquisitions without significant changes in subjective image quality, showing this new detector technology’s potential for saving contrast material usage. As regards material decomposition or spectral material differentiation it is now possible to remove materials such as calcium from the images, not just in large vessels abut also for example in the coronaries. To finish his presentation, Prof. Kröncke evoked the possibility that the use of the new PCD CT system could enable the use of dual-contrast imaging processes and cited preliminary studies of orally administered bismuth and intravenous iodine being used in the same examination, although there is still a lot of work to be done before these possibilities of PCD CT are realized.

TAKE-HOME MESSAGES

PCD CT will

◾ define a new level of resolution in CT

◾ change the way of looking at the composition /make-up of imaged tissue

◾ significantly improve contrast-enhanced vascular imaging and open up the possibility of multi-contrast imaging and new contrast agents

Figure 1. The principle of photon counting detection

Cardiac Imaging with PC-CT

With his 25 years’ experience in cardiac imaging, Prof. Cademartiri has seen significant developments in the field. He pointed out that the requirements of cardiac imaging have been major drivers for most of the technological developments introduced to CT by industry over the period, with the technology innovations first introduced in cardiac imaging then spreading out to other non-cardiac uses of CT.

Prof. Cademartiri, wanted to focus on the impact of the new photon counting technology on the whole aspect of contrast media usage. First, he reminded the audience of the central role of the use of the Iodine Delivery Rate (IDR), which is at the core of coronary imaging and cardiac imaging.

When there is a high attenuation inside the coronary arteries, there is a correspondingly high diagnostic accuracy. So over the years there has been a trend in conventional CT scanners towards lower kV imaging and higher temporal resolution scanners, Prof. Cademartiri showed examples of cardiac images acquired using the advanced technology of current CT scanners equipped with conventional detectors.

Against this background, the advent of photon counting CT technology raises the question of how to

best determine the optimal direction of future developments making use of the innovative properties of the photon counting detector (PCD). As the name suggests, the essential feature of the PCD is that it counts individual x-ray photons, and, crucially, also determines the precise energy (or energy range) of the photon detected. This applies to any element detected in the image, including calcium, gadolinium, iodine, bismuth, manganese and several other metals that could be suitable for imaging. The ability to determine the energy of the detected photon means that any noise can effectively be filtered out simply by setting an appropriate filter threshold. Yet another advantage of the new detector is that the size of the element is much smaller than that in conventional energy-integrating detectors, so dramatically increasing the spatial resolution.

Prof. Cademartiri then showed several exam-

Figure 2. The high spatial resolution of the new PC-CT means that calcified plaques in the coronaries can be clearly seen and are not accompanied by “blooming” effects which are common when EID- CT is used. The calcified plaques can be seen not in the lumen as is the case in EID CT but in the wall of the coronary artery.

Prof. Filippo Cademartiri, Chairman of Radiology, Fondazione Monasterio, CNR Pisa, Italy

Image: Prof. Filippo Cademartiri

ples which illustrated the practical effects of the characteristics of the new detector on the final acquired image. These included cases of atherosclerosis where several calcified plaques in the coronaries could be clearly seen [Figure 2]. CTs with EIDs have detectors lower spatial resolution and frequently show the presence of blooming artefacts in which the calcified plaque often appears large and intruding into the lumen of the artery, the plaques seen on PCD images are smaller, are clearly contained in the wall and not in the lumen and are not affected by any blooming artefacts. All-in-all, the use of PC-CT affords much better characterization of coronary atherosclerosis.

One of the biggest challenges in cardiac / coronary CT imaging is when there are stents in place. Prof. Cademartiri showed several striking examples of stented arteries, where

TAKE-HOME MESSAGES

For Prof. Cademartiri, the technological advances brought about by photon counting CT are so great that there is no doubt that in the future all CT systems will be photon counting-based. However, for optimal cardiac imaging, Prof. Cademartiri had some questions on photon counting CT that for him remain open:

◾ Reducing Contrast Media as far as possible. Although superficially this seems beneficial, Prof. Cademartiri nuanced his opinion – to get the maximum out of the new applications that are possible, such as delayed enhancement, it will still be necessary to give sufficient iodine.

◾ A diversified approach will be necessary. There is no single protocol for cardiac CT or coronary CT. If for example it is desired to look for scars in the ventricle a diversified protocol will be needed.

the high spatial resolution given by the PCD system enabled even the internal metal structure of the stent to be visible.

Turning to the potential impact of the new spectral imaging possibility on the new PC-CT, Prof. Cademartiri reminded the audience that Dual Energy CT imaging has been available for some time now, but when used in cardiac imaging, frequently involved trade-offs, whereas spectral imaging with the new PCD system avoids any such trade-offs. Prof. Cademartiri illustrated this with a series of projections of coronary arteries with severe calcifications, where routinely the 55 keV, 40 keV and the Pure lumen images could be seen as well as virtual non contrast images showing calcium alone and finally an iodine-only map. These are available at the push of a button for all patients.

PC-CT in non-cardiac applications

◾ Is iodine alone as contrast agent enough? Prof. Cademartiri thinks not. The new system enables combination of contrast agents to be administered to the same patient at the same time, so in the future, combinations of gadolinium and iodine could be envisaged. Of course all the necessary studies/regulatory challenges will first have to be met.

◾ A deeper knowledge of the pharmacokinetics and pharmaco-dynamics of contrast agents will be beneficial

◾ New carriers for iodine are theoretically possible as are other nuclei.

Overall such innovations effectively mean that with photon counting CT, a totally new imaging modality has emerged, covering not just anatomic structure, but increasingly functional and also molecular aspects.

The objective of Prof. Tamandl’s talk was to illustrate the performance of the new PC-CT system on non-cardiac applications by showing examples of images acquired under real-life clinical conditions. The majority of the images were acquired in his institution over the last year and a half. But first Prof. Tamandl briefly summarized the advantages of PC-CT over conventional EID CT as being

◾ Increased spatial resolution, with no dose penalty.

The in-plane resolution is 0.151 mm × 0.176 mm at the isocenter.

◾ Improved Contrast to Noise Ratio, with improved iodine sensitivity. Less noise and the possibility of lower dose of contrast agents.

Prof. Dietmar Tamandl, Department of Biomedical Imaging and Image-Guided Therapy, Medical University of Vienna, Austria

Image: Medical University of Vienna

◾ The possibility of spectral imaging in every scan, enabling virtual mono-energetic reconstructions, iodine maps and virtual non-contrast reconstructions.

A) Lung Patients

Prof. Tamandl showed the comparison of images acquired by PC-CT and EID CT from a patient with interstitial lung disease. In this study the PC-CT examination was carried out applying a straightforward protocol — no special tweaking / set-up was required. The EID images were acquired using a high-end conventional EID CT scanner, which used to be the flagship scanner in the hospital. The images for PC-CT were much clearer than those from EID CT at approximately the same dose. When quantitative analysis of image quality is carried out, the PC-CT images show statistically significant improvement in terms of quality compared to images acquired by EID CT.

In long-COVID patients, i. e., patients with persistent Covid symptoms, it was found that, when compared with EID CT, PC-CT detected additional findings, e. g. bronchio -

ectasis in about 50 % of cases, for examinations carried out at approximately the same radiation dose. Such findings have direct clinical significance

Systemic sclerosis – Interstitial lung Disease. In a study from Dr. Alkadhi’s group in Zurich, comparisons were made between by conventional EID CT and PC-CT, which for this study was carried out at lower doses, namely 100 %, 66 % and 33 % of the dose used in EID CT. Even at 33 % of the dose used in the conventional scanner, the images were equivalent in terms of diagnostic performance. Efforts are on-going to reduce the dose even further with no loss of diagnostic performance.

B) MSK patients

It was found that PC-CT was clearly superior to EID CT in detecting osteolysis in patients with multiple myeloma in studies carried out at approximately the same dose.

Temporal bone imaging. In a young patient with recurrent middle ear infections who had a stapes prosthesis inserted, it was found that the PC-CT images were less noisy, and clearer than EID CT even at half

the dose. More extensive studies confirmed these findings.

C) Oncology patients / abdominal imaging

Patients with hepatocellular carcinoma were imaged, using the spectral imaging capabilities of the new photon counting CT system which can provide virtual mono-energetic images (VMI), e.g. at 40 keV, 50 kEV, 60 keV, 70 keV & 80 keV [Figure 3]. Such mono-energetic images do not have to be set up prior to the PC-CT examination; they can be extracted from the data during or after the examination. At low keV levels, the iodine attenuation is of course much greater. The price to pay for this, namely increased noise, is still acceptable, i. e. the contrast-to-noise ratio is increased compared to EID CT. Several broader studies have been carried with the same conclusion, namely that with low-keV VMI, PC-CT yields significantly improved objective and subjective quality of arterial phase oncological imaging compared with EID CT. This advantage may translate into higher diagnostic confidence and lower radiation dose protocols. Interestingly in such stud-

Figure 3. Effect of varying keV: a patient with Hepatocellular Carcinoma (HCC) having had an administration of 90 ml of IOMERON 400. It can be seen that as the chosen keV level is decreased, e. g. to 40 keV or 50 keV, the attenuation and clarity of smaller structures are improved.

Image: Bracco

ies, although the average dose with PC-CT was 18 % lower than that with EID CT, there were nevertheless some patients who actually received higher dose than with EID CT. However, this might be due to ongoing protocol optimizations.

Obese patients. The question of noise in the imaging of obese patients is an issue — the usual tradeoff using EID CT is either accepting very noisy images at reasonable radiation doses or significantly increasing the dose to reduce the noise. With PC-CT there is still an increase in noise with increasing BMI of obese patients but studies have shown that the dose used is 25 % less than that with EID CT. Thus the use of PC-CT is preferred with obese patients.

One reason behind the high quality of the PC-CT images is that a new iterative reconstruction algorithm, specially designed for PC-CT is incorporated in the new system. Known as quantitative iterative reconstruction (QIR), this algorithm was first evaluated in phantoms where it was found that the loss of accuracy as dose is reduced was much less than the loss of accuracy with even high level EID CT. Such findings have been confirmed in human patients where it was shown that the use of QIR decreased noise by 44 % and increased the CNR by 74 % in portal venous phase examinations.

Iodine quantification. It was found that the error in iodine quantification was roughly equivalent to that obtained with even high-level EID CT systems, but as the size of the patient or phantom increases, the increase on the iodine error quantification was much less than that with EID CT If it is desired to go in the opposite direction, i.e. removing the iodine, so creating Virtual Non Contrast (VNC) images, there has always

been a small residual iodine attenuation in VNC images compared to true non-contrast (TNC) images. This difference is constantly being reduced with development of the software. Although the quality of VNC images in photon counting CT is still slightly less good than in TNC, they are nevertheless still acceptable for qualitative purposes.

One main use of VNC is in steatosis assessment. It was found that the VNC images from the PC-CT were reliable for the assessment of hepatic steatosis, although there may be a tendency to slightly overestimate the number of patients with steatosis, because of a small error of approximately 8 – 10. Hounsfield Units between the VNC and the TNC. Successive software updates are reducing this error.

TAKE-HOME MESSAGES

Prof. Tamandl’s main conclusions were:

◾ Higher resolution with less noise in non-contrast images was shown in:-Interstitial Lung Disease patients; Post-Covid investigations ; Patients with Temporal bone prostheses; MSK patients

◾ Spectral information similar to dual energy CT but available all the time

◾ Higher Contrast-Noise-Ratio less noise: Better performance in low dose scans; Virtual mono-energetic reconstructions are now standard Iodine quantification and virtual non-contrast images easily available

MORE INFORMATION

A video recording of this Braccosponsored symposium is available at: https://connect.myesr.org/course/ photon-counting-ct-a-new-challenge-for-contrast-administration/

Breakthrough 5T MRI System and AI features unveiled at ECR 2023

United Imaging Healthcare, a global provider of advanced medical imaging and radiotherapy equipment, life science instruments and intelligent digital solutions, has just launched a pioneering whole-body ultra-high field 5T MRI scanner, the uMR Jupiter 5.0T, at ECR 2023.

The system was unveiled for the first time internationally along with United Imaging Healthcare‘s comprehensive flagship products across all modalities, including PET / CT, PET / MR, MR, CT, and DR, on March 1 in the ACV, the traditional home of the ECR in Vienna, Austria.

The uMR Jupiter 5T features an 8-channel volume transit coil, gradient strength of 120 mT / m and a peak slew rate of 200 T / m / s. The 5T magnet has the same footprint as 3T system and can be installed in a 3T-sized room, making it more accessible to healthcare professionals.

The uMR Jupiter 5.0T unveiling ceremony

Thanks to its higher resolution and SNR, radiologists can now visualize anatomical structures that were

previously unseen on 3T MRI. “This feature enhances diagnostic confidence and facilitates precise preoperative planning,” the group wrote in a communiqué.

The uMR Jupiter 5T, which is not CE marked yet, is “an entirely new species, setting itself apart from the pack,” the group also wrote. It has successfully mastered the technical challenges of ultra-highfield MRI to achieve whole-body clinical applications, the firm added.

Other featured products

United Imaging Healthcare also presented a number of cutting-edge technologies, including the vendor’s artificial intelligence (AI) for imaging (uAIFI) Technology Platform, which is dedicated to empowering the group’s MR (uMR) scanners and increasing access to AI.

The platform is based on a series of advanced AI applications and hardware innovations that make high-quality imaging more attainable, faster, and easier, while offering modern solutions for patient care, according to the company.

The uAIFI has been designed to expand the potential for uMR to offer more powerful imaging capabilities and an improved examination experience, the communiqué also read.

“Our constant drive for innovation and attention to both clinical and patient needs give our products an edge even in one of the most competitive regions,” Dr. Jusong Xia, President of International Business at United Imaging Healthcare, said. “As one of the largest medical meetings in Europe, ECR (…) is a great opportunity to showcase our capabilities to industry stakeholders and raise brand awareness in this industry,” he added.

In 2019 Poland installed its first digital PET / CT (uMI), representing United Imaging’s kick-off in the European market. In 2022 the vendor installed its first digital PET / CT uMI780 in Italy, along with sequential installations of the whole portfolio products, including the first 75 cm ultra wide bore 3T uMR Omega in Europe.

“Europe has a very mature healthcare market with fierce competition. We have put much investment in the European market. In 2019, for example, we established

our subsidiary in Warsaw and now have a well-equipped local team including marketing, training, and after-sales service for hundreds of clients across Europe,” added Vice President and General Manager for Europe, Lukasz Mizerka.

United Imaging Healthcare was established in 2011 with a commitment to provide high-performance medical imaging products, radiotherapy equipment, life science instruments, and intelligent digital solutions to global customers. With a mission to “Bring Equal Healthcare for All” and a vision to lead healthcare innovation, United Imaging Healthcare continuously works to create more value for its customers and improve the accessibility of high-end medical equipment and services worldwide.

United Imaging Healthcare also embraced ECR 2023’s theme, ‘The Cycle of Life’,  highlighting its commitment to provide equal healthcare for all. “United Imaging is certain that medical imaging can improve people‘s lives,” Dr. Qiang ‚Al‘ Zhang, Chairman & Co-CEO of United Imaging Healthcare, said. “Through promoting technological innovation, United Imaging is dedicated to achieving its mission of Equal Healthcare for All. Our products are utilized in more than 10,400 hospitals and research facilities in 59 nations and regions, including China, the U.S., Europe, Japan, India, Southeast Asia, Africa, South America, and the Middle East,“ he concluded.

United Imaging Healthcare unveiled the uMR Jupiter 5.0T, a pioneering whole-body ultra-high field 5.0T MRI, at ECR 2023 – along with along with a series of other groundbreaking technologies.

The top 100 most cited AI-Articles: a bibliometric Analysis in Radiology

The number of papers published in the field of Artificial Intelligence (AI) and its applications in radiology continues to increase at an exponential rate, to such an extent that it is very difficult to keep up with the developments and emerging trends.

Against this background a recently published paper [1] describing a detailed bibliometric analysis of the current literature on AI in radiology and which enabled the identification of the top 100 most cited articles in the field is particularly welcome. We wanted to find out more, so we spoke to the lead author, Dr. Hannah Hughes, Specialist Registrar in Radiology at St Vincent’s University Hospital in Dublin, Ireland.

By Alan Barclay

◾ To start with, let’s set the scene. What was the basic rationale for carrying out the bibliometric analysis in the first place?

It was predominantly curiosity on my part that drove me to write this paper. AI is a burgeoning field in medicine in general and radiology in particular, with a significant increase in the number of papers published on the topic in recent years. When I began to look into it myself, the scope of the available literature was vast and therefore rather difficult to synthesize at first glance. I have seen bibliometric analyzes used to great effect in other fields of medicine and surgery, such

as “The top 100 most cited articles in medical artificial intelligence: a bibliometric analysis” by Sreedharan et al. in 2019 [2] or “The 100 Classic Papers of Pediatric Orthopaedic Surgery A Bibliometric Analysis” by Kavanagh et al. [3]

I felt that the approach could be particularly useful in radiology; having a central summary to refer back to would be helpful.

◾ What are the main principles behind bibliometric analyzis?

As the name itself suggests, bibliometric analysis is the process of evaluating published literature on a particular subject. It involves the collation of data in order to determine the most influential or most

cited publications in a particular subject. It can also be used to identify trends in certain research fields and to determine potential gaps where further research is required. However a possible, but important limitation of bibliometric analysis the phenomenon whereby findings from older seminal papers may become assimilated into general knowledge within a speciality and therefore may not be cited in contemporary literature. [4]

◾ Is citation count the best measure of an article’s impact?

This is a particularly interesting question with regard to published medical literature. I published a paper related to this topic in 2017. [5]

Image:

Dr. Hannah Hughes

Dr. Hannah Hughes Department of Radiology, St. Vincent’s University Hospital, Dublin, Ireland

Although impact factor has been used since the 1960s as a quantitative measure of a journal’s output quality, citation count may not be the most reliable metric in evaluating the overall influence of published literature in the modern era. The influence of an article on clinical practice is largely reliant on its importance to practicing physicians, some of whom do not produce published work as part of a dedicated research agenda. As a result, this key demographic of end-users is not being captured via impact factor and citation metrics alone [6]

The use of social media (SoMe), in particular amongst medical professionals, has grown rapidly over the last two decades and it therefore provides a platform where an article’s impact can be more broadly assessed [7]. Since the advent of SoMe, many scoring systems, or ‘alternative metrics’, have been developed to try and combine traditional citation metrics and online metrics (to establish to what degree published literature is interacted with online). For example, Altmetric (www.altmetric. com) is an online software that collates data from multiple sources; from mentions, clicks and views on social media to citations via online sites like Wikipedia in order to calculate an ‘Altmetric Score’. The higher the score, the greater the degree of online interaction with that material. Another alternative metric in use is the ‘Klout Score’ (www.klout.com). This is generated by the aggregation of online reactions and activity stimulated by a SoMe user or profile (clicks, likes, comments, reshares, purchase behaviour). Again, a higher Klout Score corresponds to greater influence across SoMe.

While alternative metrics are likely to play a key role in literature evaluation going forward, their use

is not without limitations. Some caveats to mention are the fact that the ‘weight’ of a mention is not quantified; i.e. whether an article is simply mentioned or is the topic of an in-depth discussion. Furthermore, the validity and reliability of alternative metrics is hindered due to the potential for online data manipulation and relative lack of quality control compared to traditional citation methods.

◾ What about citation count divided by years since publication?

A potential source of bias in bibliometric analysis is the use of citation count as the sole measure of an article’s impact. The longer a paper has been published, the more likely it is to have been cited over time and thus have a higher citation count overall. In order to mitigate this source of bias, we used a method employed in prior bibliometric analyzes whereby the average citation count per year is used as an adjunct means of evaluating an article’s overall impact [2] .

◾ Now turning to your analyzis of the top 100 most cited articles on AI in radiology, what were the clinical fields most often represented? Were clinical articles in general more recent than the non-clinical? If so, could this indicate that the basic IT development phase is giving way to a phase of application of the algorithms?

We classified clinical studies as those that included human participants or those related to clinical applications of artificial intelligence. Out of the top 100 articles, 52 were categorized as clinical studies and more than half of those studies (34 articles) were published within the last ten years.

This is not surprising given the increasing demands for quality

healthcare delivery worldwide coupled with physician shortages in many countries. As a result, the demand for diagnostic radiology provision is growing at a rate that is disproportionate to the number of available radiologists. This combination has lead to an increased need for productivity and efficiency in order to maximize output; this is where artificial intelligence theoretically steps in. Analyzis performed by Accenture has estimated that artificial intelligence will reduce healthcare costs in the United States (US) by up to 150 billion US dollars per year by 2026 [8]. Thus, it is clear why efforts to safely integrate artificial intelligence into clinical radiological workflow while minimizing errors has largely dominated this sphere of research over the last 10 years.

◾ Your analyzis showed that although neuroimaging applications were the most cited in absolute terms, when the number of citations was expressed as a function of the time since publication, respiratory applications topped the list. Do you think this is just a “blip” caused by the Covid-19 epidemic?

Yes, this is very likely. The Covid-19 pandemic saw an unprecedented increase in the number of publications worldwide; in 2020. Four per cent of published literature was on the topic of Covid-19 alone, with a significant increase seen in the publication of studies on all subjects that same year. [9] .

Our study indeed found that, while the majority of the clinical studies included in the top 100 most cited articles on artificial intelligence in radiology were on the topic of neuroimaging, the average number of citations per year for articles relating to respiratory imaging

ARTIFICIAL INTELLIGENCE

was significantly higher. Two out of the seven clinical respiratory articles focused on the topic of Covid-19 and both of these articles were published in 2020.

◾ Which were the most prolific countries, and centers in terms of contributing to the top 100 most cited articles?

The majority of publications in the top 100 list originated in the United States. That the United States is prolific in research output is well known across many fields of research. Studies have found that American authors are more likely to cite other American authors, thereby increasing the prevalence of US papers amongst the most cited papers in the literature. This may be partially due to the fact that authors writing in English have been shown to be more likely to cite papers in English [10]. China and Europe also contributed to the top 100 list. However they were each less than half of that made by the US.

◾ Let’s now broaden out from the bibliometric analysis, into AI in radiology as a whole. Several years ago many observers (mostly computer scientists – not radiologists!) came out with dire warnings about the future of radiology being completely overtaken by AI, making radiologists unnecessary. What in your opinion is the current status of the debate over AI being an existential threat or an indispensable help?

Experts from various fields have highlighted concerns that the incorporation of artificial intelligence into medicine will render the physician, and indeed the radiologist, redundant in terms of a patient’s journey to diagnosis [11]. We know that there are a number of essential human tasks that physicians carry

out on a daily basis that cannot be replicated, automated or assigned an economic value, from empathetic human interaction to clinical judgement and participation in multidisciplinary discussions. Studies have estimated that in one eight-hour-workday, a radiologist is required to interpret, on average, one image every three to four seconds. This bolsters the case for the automation of appropriate functions. However, rather than replacing the clinician, it is believed that the role of artificial intelligence will be to support those essential human functions.

Collaboration between the fields of radiology and computer science will therefore be critical over the coming years in order to promote collective artificial intelligence research and put it to its best use. The primary goal is to have artificial intelligence safely integrated into clinical practice in such a way that it strengthens the radiologist’s or physician’s role in providing those essential human components of patient and professional interactions.

◾ To come back to your bibliometric analyzis. Given that the number of published papers on AI in radiology looks like it is continuing to grow, do you anticipate that you will have to repeat the bibliometric analysis of the AI literature on an on-going basis?

Each bibliometric analysis is essentially a snapshot in time and is therefore vulnerable to, and under the influence of, continuous fluctuations in citation counts. One of the potential gaps in the literature that this bibliometric analysis, and indeed our conversation today, has highlighted is the question of where alternative metrics fits into the

bibliometric analyzis framework. This is something I am currently working on and I look forward to sharing the results.

References

1 Hughes H et al. The top 100 most cited articles on artificial intelligence in radiology: a bibliometric analysis Clin Radiol 2023 Feb;78(2):99106.

https://doi.org/10.1016/j.crad.2022.09.133

2 Sreedharan S et al. The top 100 most cited articles in medical artificial intelligence: a bibliometric analysis. J. Med Art Intell. 2020; 3: 25.

https://doi.org/10.21037/jmai.2019.11.04

3 Kavanagh RJ et al. The 100 classic papers of pediatric orthopaedic surgery: a bibliometric analysis J Bone Joint Surg Am. 2013; 18; 95(18).

https://doi.org/10.2106/JBJS.L.01681

4 Crockett MT et al. 100 classic papers of interventional radiology: A citation analysis, World J Radiol. 2015; 7(4): 79-86.

https://doi.org/10.4329/wjr.v7.i4.79

5 Hughes H et al. The Use of Twitter by the Trauma and Orthopaedic Surgery Journals: Twitter Activity, Impact Factor, and Alternative Metrics. 4. Cureus. 2017 ; 9(12): e1931.

https://dx.doi.org/10.7759/cureus.1931

6 Saha S et al. Impact factor: a valid measure of journal quality? J Med Libr Assoc. 2003;91(1):426. PMID: 12572533

7. Wilkinson SE et al. The social media revolution is changing the conference experience: analytics and trends from eight international meetings. BJU Int. 2015; 115(5): 839-46.

https://doi.org/10.1111/bju.12910

8 Bohr A & Memarzadeh K. The rise of artificial intelligence in healthcare applications. Artificial Intelligence in Healthcare. 2020: 25–60.

https://doi.org/10.1016/B978-0-12-818438-7.00002-2

9 Else H. How a torrent of COVID science changed research publishing – in seven charts. Nature. 2020 ; 588(7839): 553.

https://dpi.org/10.1038/d41586-020-03564-y

10 Di Bitetti MS & Ferreras JA. Publish (in English) or perish: The effect on citation rate of using languages other than English in scientific publications. Ambio. 2017; 46(1): 121-127.

https://doi.org/10.1007/s13280-016-0820-7

11 Liew C. The future of radiology augmented with Artificial Intelligence: A strategy for success. Eur J Radiol.2018; 102: 152-156.

https://doi.org/10.1016/j.ejrad.2018.03.019

12 McDonald RJ et al. The effects of changes in utilization and technological advancements of cross-sectional imaging on radiologist workload. Acad Radiol. 2015; 22(9): 1191-8.

https://doi.org/10.1016/j.acra.2015.05.007