Imaging in medicine

2025:5 top digital trends for medical imaging in 2025

Artificial intelligence tools have an increasing impact on the medical imaging sector, helping to improve the efficiency and speed of image processing. However, there are also certain restraining factors that prevent the faster introduction of AI in the relevant area. In a material dated January 15, 2025, Signify Research specialists identified five main trends in the field of digital technologies for visualizing medical data.

1. Generative AI (Genia)

The introduction of Genia is expected to accelerate in 2025. Analysts believe that IT visualization providers will focus on operational options for using generative AI. It is about helping to improve the efficiency of medical institutions and radiologists in the short term. GIs can be used, for example, to summarize patient data or generate medical 3D images based on 2D data. This will provide specialists with a more complete picture of the features of the disease and the anatomy of patients. GeneAI can help in making the most informed decisions on a case-by-case basis.

𝓲

Фото: Freepik

2. Multimodal AI

We are talking about models of artificial intelligence that perform user operations and are trained on several modalities: text, images, video, 3D, speech, sounds, tables, graphs, code. This approach expands the capabilities of AI systems. It is noted that by the beginning of 2025, most medical AI imaging platforms use only images of different types with limited integration of additional information, such as clinical notes, patient histories and laboratory results. However, the addition of additional data sources will improve diagnostic accuracy and open the way to more personalized treatment. The authors of the study believe that the emergence of specialized large language models (LLMs) in healthcare opens up new opportunities for multimodal AI. LLMs allow you to extract data from natural language sources (unstructured), such as all kinds of medical records and reports. Overall, multimodal AI is one area in which AI system providers are expected to invest heavily in 2025.

3. AI regulation

The regulatory framework covering medical devices, and therefore AI for medical imaging, is one of the main obstacles to the larger adoption of artificial intelligence technologies. Current regulatory requirements can lead to significant delays in bringing innovative AI products to market. Therefore, the authors of the study believe, there is a significant risk of stopping innovations in the field of medical AI imaging.

4. Market consolidation

Due to the difficulties in raising funding, as noted in the report, the question arises: "Who will pay for AI?" Only a small number of industry participants boast breakeven status, so mergers and acquisitions are expected in 2025, it said.

5. Increase the size and complexity of deals

The conclusion of multibillion-dollar IT contracts in radiology, according to the authors of the study, will become more commonplace during 2025. At the same time, not only the scale of contracts increases, but also their complexity. These larger deals will begin to shape a new competitive dynamic.^[1]

2021

Beginning of using the method of 3D imaging of hard-to-distinguish brain states

In early May 2021, researchers from Matai Medical Research Institute the Stevens New Zealand Institute of Technology and other scientific institutions presented a new advanced method for 3D imaging of hard-to-distinguish brain conditions such as obstructive diseases brain and aneurysms. 3D-MRI with enhancement allows evaluating physiological movement of the brain in all directions in real time. More. here

Start of using the AI1 system diagnosing CT and X-ray in the Russian Federation

In March 2021, the European Medical Center (EMS) announced that it was the first in Russia to use the AI1 medical imaging system (All-In-One) developed by the Israeli company Zebra Medical Vision. Read more here.

2020: The dose of radiation during medical visualization has decreased by 20% over 10 years

In mid-March 2020, the US National Council on Radiation Protection presented a report according to which the dose of radiation in medical imaging fell by 20% over the past decade.

The new paper covers trends in diagnostic and interventional medical imaging in the U.S. and tracks radiation doses used to produce adequate images. The committee found that radiation exposure in medical treatments decreased from 2.9 mSv per person in 2006 to 2.3 mSv in 2016. The total number of X-ray examinations also fell from about 877 million in 2006 to 691 million in 2016.

In mid-March 2020, the US National Council on Radiation Protection presented a report according to which the dose of radiation in medical imaging fell by 20% over the past decade

It is known that in the previous period, from 1980 to 2006, there was a sixfold increase in the same indicator due to the widespread introduction of new medical imaging methods, mainly in the form of computed tomography and nuclear medicine. The authors summarized the trends identified and discussed the potential reasons for the sudden change in radiation exposure of patients in the United States. The report used data from a wide range of sources, including American Medicare Part B applications, professional society data registers and surveys of several federal and commercial agencies.

One of the main reasons for the overall reduction was a significant reduction in the number of nuclear medicine procedures - from 17 million in 2006 to 13.5 million in 2016. This is mainly due to insufficient reimbursement and the introduction of new, safer and equally accurate screening methods such as echocardiography. The dose of CT radiation is likely to have decreased due to new methods of dose modulation, as well as the efforts of public campaigns to optimize the dose and reduce unnecessary prescriptions.^[2]

2019: Unnecessary MRI contrast injection can be avoided through CAD

In late May 2019, 3D radiology specialists used a computer-aided design (CAD) system to assess whether patients with multiple sclerosis should be injected with contrast agent in brain MRI scans. The results of the study are presented online in the Journal of the American College of Radiology.

A team of researchers led by Dr. Jeffrey Rudie of the University of Pennsylvania Hospital in Philadelphia has tested a new MRI protocol for multiple sclerosis. Although patients with multiple sclerosis usually undergo contrast studies to assess response to therapy, the lack of new disease activity in non-contrast imaging suggests that the introduction of contrast will be uninformative. In an attempt to reduce the risk of gadolinium contrasts, the team of researchers decided to use CAD to weed out patients without new lesions.

3D radiologists used CAD to assess whether patients with multiple sclerosis should be injected with contrast agent in brain MRI

After obtaining a non-contrast image in free water signal suppression MRI (FLAIR) mode, the researchers used CAD software to assess the new disease activity. A specialist in the institution's 3D laboratory reviewed the CAD results and informed the analyst within a few minutes if a gadolinium-based contrast agent was required to continue the examination.

CAD evaluated non-contrast images over 10 minutes, and preliminary data were then compared with the neurorengenologist's conclusion. During the two-month pilot trial, it appeared that unnecessary contrast administration could have been avoided in 87% of patients. This approach has significantly reduced the costs to the health care system associated with the contrast agent itself, the time to introduce contrast, and the time to obtain and interpret images.^[3]

2018: Investment in AI technology developers for medical imaging doubles to $580 million

In early February 2019, Signify Research presented an analysis of funding for companies that develop solutions for medical imaging based on artificial intelligence. According to analysts' calculations, in 2018 the volume of investments in this market reached almost $580 million, which is twice the amount received by startups in 2017.

The report shows that funding for developers of AI solutions for medical visualization has moved to a new stage. Newly created companies began to go beyond initial research and development and began to look for their place in the market. Investments in later stages of startup development are aimed at financing clinical validation research, developing additional products, entering the international market and expanding the operational aspects of the business.

In 2018, developers of AI technologies for medical visualization received more than $0.5 billion in investments

By the end of 2018, more than 120 startups were developing AI for medical imaging, and the total amount of funds invested in them since 2014 exceeded $1.2 billion. The average amount of investments in one company amounted to $14.4 million in 2018.

HeartFlow received the most support, in which investors invested $476.6 million to develop cardiac applications. It is followed by Voxelcloud - a US company with $80.5 million in funding, and Chinese startup Infervision with a $73.1 million fund. At the same time, 42% of investments of more than $240 million settled in Asian companies (if excluded from the HeartFlow analysis). This fact may be partially explained by the expansion of Asian developers in other countries and regions.

Analysts note that such investments are highly likely to justify themselves. In a market with a shortage of radiologists with experience with artificial intelligence, it will become one of the significant tools for improving the efficiency of work processes. This will provide cost-effective diagnostic pathways, such as the use of non-invasive CT instead of angiography.^[4]

2017: Medical Imaging Trends Forecast

In December 2017, Ambra Health CEO Morris Panner presented a forecast for technology trends that he believes will change medical imaging in 2018.

The expert believes that digital technologies should take a special place in the healthcare industry. Changing consumer preferences, merging supplier companies and the emergence of tech titans such as Apple, Amazon and Microsoft in the industry will fundamentally change the market for medical devices and services themselves.

The main priority will remain value-oriented service, and although we are unlikely to receive an AI doctor in an application on a smartphone next year, several technological trends can already be predicted, Panner said.

Open networks of recommendations are becoming more important for the growth of the health care system

Imaging studies are becoming more popular thanks to a wide network of recommendations in patient communities. According to a survey by Ambra Health, patients are still looking for doctors on recommendations, which means that health professionals will look for effective, time-saving and high-quality ways to process images.

Digital technology should take a special place in the healthcare industry, experts say

The Big Four is being introduced into the healthcare sector

Google, Amazon, Apple and Microsoft continue to be introduced into a new industry for them in various ways: they are already engaged in venture financing, research and the creation of new applications in the field of health care. The Big Four received special attention from information technology. Now the real race has begun to dominate the field of transmission and storage of information, including data from imaging studies. Google has already attended the 2017 annual conference of the Society of Radiologists of North America (RSNA) and announced a collaboration with several medical imaging device providers, including Ambra Health.

Cloud image archives will dominate information

Providers are looking for independent cloud archives to create a shared store of information. Centralizing the data archive will give doctors access to all the patient's necessary information, which will increase the efficiency of their work. Some forms of data do not correspond to the traditional electronic information storage system. Single cloud storage provides the ability to organize such data in a rigorous and convenient way.

Patients will be able to influence the decision on data privacy

AI can radically transform healthcare with huge data archives that are used to test new ideas. However, this raises problems regarding the use of confidential data. In 2018, patients will be able to play a more active role in this matter and influence the decisions of large companies - a new era is coming for the field of medical ethics, and it should be based on openness for all participants in the process.

Informatics develops by leaps and bounds

Value-based healthcare is beginning to prevail over the conventional fixed-fee-for-service model. Patient-centric means that information technology, especially those that harness the potential of deep learning and artificial intelligence technologies, will be increasingly used to optimize and streamline data flow and imaging research results.

Suppliers are ready to meet the changed requirements of consumers

The 21st century healthcare consumer is already familiar with cloud and mobile technology. However, health care providers are changing their priorities much more slowly, and in the U.S. so far, about 44% of patients receive research data on CDs. In 2018, thanks to the widespread adoption of cloud-based storage technologies, providers will be able to provide patients with access to all the information they need at any time convenient for them.

Moving from data protection to threat prevention

The transition to centralized storage requires appropriate information protection. Thanks to the latest AI and machine learning security technologies, data protection is reaching a new level: now programs do not respond to a threat, but prevent it from appearing. Artificial intelligence and deep learning technologies will play an increasingly important role in predicting security threats, which will undoubtedly require serious investment from the health sector.

AI is becoming increasingly important in document automation

Modern machine learning and document automation technologies can speed up data processing, improve diagnostic accuracy and reduce errors in radiology and other areas. AI became the most talked about topic at RSNA's annual conference in November 2017. Many medical practitioners admire the new opportunities to transform everyday workflows, while other professionals are skeptical of AI's advances.

The job of AI is to process a huge array of patient data and get comparison algorithms. There are undoubtedly many potential uses for AI to improve radiology workflows. One example is the use of algorithms to automatically compare the results of new and previous studies.

Implementation of virtual and mobile services

2018 will open a new level of mobile services. Already, companies are providing applications for primary care, preventive screening and even treatment. The cost-effectiveness and efficiency of new technologies contribute to the expansion of the range of services and their application.

Ethnic and sociocultural diversity among radiologists benefits patient

Thanks to new technologies, radiologists are increasingly communicating with patients, and interaction with the patient and his relatives is an extremely personal and emotional part of the doctor's work. Therefore, it is so important that the society of doctors reflects the real situation in the world, and information technology will allow new specialists of various nationalities to contribute to the common cause.^[5]

Notes