top of page
  • DIHs_P4Medicine Consortium


Updated: Jun 27


  • One of the promising applications of artificial intelligence in the healthcare sector is the use of AI in medical imaging, which helps professionals to detect and quantify different diseases and pathologies, thus considerably improving diagnostic effectiveness and efficiency and assignment of treatment.

  • Another important effect of this technology in medical imaging is freeing professionals from slow and repetitive tasks, and the early detection of diseases.

  • Population ageing, increased use of medical imaging tests, the need to improve work flow and efficiency, and the need for greater precision in detecting and diagnosing diseases are the principal grounds for expecting the global market for AI in medical imaging to continue expanding, to reach $1200 million in 2025.

  • In countries such as Mexico or India, where the number of expert radiologists is significantly lower than in other countries such as China or the United States, artificial intelligence in medical imaging may have a major impact as well as reducing treatment costs owing to automation and early prevention techniques.

  • At present, there are numerous use cases of this technology in medical imaging, and this trend is expected to expand into other branches of medicine, from fields such as oncology, radiology, neurology, cardiology, surgical interventions, to studies in fractures, or complications of the thorax, to give a few examples.

Medical imaging in the field of health and medical care has facilitated diagnosis and administration of treatments thanks to creating and processing images of the human body or its parts. This increasingly digitised market is also constantly growing. In 2021 the size of the global medical imaging market was valued at $37,970 million, and it is expected to reach $56,530 million by 2028, as per data by Fortune Business Insights.

According to this study, the introduction of advanced technology such as artificial intelligence in medical imaging is one of the main drivers of growth in this market, as this technology potentially processes these images more rapidly, thus enabling faster diagnosis and treatment for patients.

The use of Artificial Intelligence (AI) in processing medical imaging, such as X-rays and CAT scans, leads to more effective medical care. Owing to the vast amount of visual data generated within the healthcare sector, its analysis and review is potentially highly time-consuming for staff, and one of the major advantages of AI is the capacity to analyse and organise large quantities of data in great detail in seconds or minutes. In accordance with data by GE Healthcare, more than 90% of healthcare data are extracted from medical imaging, and over 97% of medical imaging is not analysed2. Statistics also show that, on average, radiologists currently read approximately 12 magnetic resonance images per minute, in contrast with 3 images one decade ago3.

From September 2020 to August 2021, the British National Health Service (NHS) reported having performed 40.3 million medical imaging tests in the United Kingdom4.

This increase in the amount of medical imaging, while endeavouring to improve workflow and efficiency, and to achieve more precise detection and diagnosis of diseases, are the principal reasons for expecting the global market for artificial intelligence in medical imaging to continue its expansion to reach $1200 million by 20255, an increase of more than $800 million with respect to the size of this market in 2020, according to Signify Research. Moreover, as per data by Research and Markets, it is estimated that the global market for artificial intelligence in medical imaging will reach approximately $3200 million by 2027 (6).

Other factors driving market growth and the implementation of artificial intelligence in medical imaging is population ageing (in 2001, 16% of the population was 65 years of age; in 2020 this percentage was 21%7) as well as different countries' investments in this technology. Population ageing causes growing numbers of medical imaging diagnostic tests, as mentioned, hence the need for enhancing and streamlining patient assessment and diagnosis processes.

In addition, in countries such as Mexico, where there are only some 4000 radiologists attending a population of 130 million(9) or India, where just 10,000 trained radiologists must care for a population of 1,300 million8, AI in medical imaging may have a great impact and help to compensate this scarcity of professionals in the sector9. Indian startups such as Predible Health, based on artificial intelligence in medical radiological imaging, were created for this purpose and this one in particular supports professionals in diagnosing and treating any affection of the chest, including lung nodules, pulmonary fibrosis, emphysema, and even in detecting COVID-19 and to determine its severity, much more quickly and efficiently(10).

However, in counties like China, where in 2017 some 158,072 radiologists were counted for 1411 million inhabitants(11), the application of AI to medical image processing has become the second largest market segment for medical applications of AI in the healthcare sector, and is forecast to reach $2,500 million by 2024. This growth is accompanied by the fact that each year, overall data for medical imaging increases by 30% in the country, representing 90% of the volume of digital data at hospitals (12).

In the United States, the increasing number of patients receiving a diagnosis and treatment thanks to medical imaging analysis has boosted market growth to reach $11,300 million in 20201, which has established the context for integrating artificial intelligence in these systems. According to a survey by KLAS Research conducted on medical care executives in the United States in autumn 20201, around 48% of participants stated that they were currently implementing artificial intelligence in diagnostic processes using medical imaging(13).


Until now, health professionals had to draw patterns for reaching a diagnosis and applying treatment in a long and tedious process, as they had to record and analyse all the patient's data using traditional methods. Artificial Intelligence accelerates this whole process analysing great quantities of data and forming more highly detailed patterns, so that professionals only need to interpret these and issue a diagnosis, which is far more precise thanks to this process and to having extracted the clinically significant information. This leads to faster and more accurate decision-making regarding the treatment administered to patients.

The implementation of AI in medical imaging also contributes toward relieving the workload of healthcare professionals who, with the automation of tedious and repetitive tasks, gain time to focus on the more important and urgent cases, while the risk of possible errors and inefficiency is reduced through optimised use of resources. According to data published by Vitech, this type of AI applications have a precision factor of 97% to 99%, and enhance workflow efficiency by reducing the time taken to read reports by 34%(13).

An example of this is set by Cognitiva and ChileRad, that feature what is referred to as “Análisis Cognitivo de Imágenes Médicas [Cognitive Analysis of Medical Imaging] (ACIM)”, a system supporting medical diagnostic tests trained through over 300,000 anonymous X-ray images of the chest taken from the Chilerad databases and Stanford University. According to Cognitiva, "its learning capacity automates the image reading and interpreting process to improve the time taken in analysis and drafting medical reports, since radiologists only need to load the patients' X-ray images on the system for an analysis and convey to the treating doctor an accurate pre-diagnosis of the pathology observed"(14). The system is capable of distinguishing up to 16 different medical pathologies. In addition, an algorithm developed by Stanford ML Group, known as CheXNeXt, reads chest X-rays for 14 different pathologies, with the same precision as expert radiologists but with greater efficiency, since radiologists took an average of 4 hours to conclude their readings and the algorithm performed the task in less than 2 minutes(15).

Training AI with thousands of videos and images and teching the software to distinguish different diseases, it can build a recognition pattern capable of detecting some of these diseases with far greater precision and immediacy than a doctor, in addition to which it can differentiate among variations that are not discernible by specialists, to the extent of performing early detection of a disease, offering patients not only better results but also reduced treatment costs. For instance, early diagnosis and treatment of many types of cancer will potentially reduce costs by more than 50%(16).

In a recent study published in The Journal of Nuclear Medicine, researchers discovered that on implementing AI in medical imaging and combining this with clinical data, doctors are able to enhance predictive models indicating the risk of heart attack in patients with an established coronary arterial disease(17).


The potential and benefits of artificial intelligence in medical imaging has led to its implementation in several fields of medicine, resulting in new forms of support to professionals for faster and more efficient detection of different types of disease.

Analysis of cerebral/morphometric structures

Siemens Healthineers has launched a series of AI-driven solutions for addressing problems in patient diagnostics. Ons such solution is AI-Rad Companion Brain MR, which helps professionals to view, analyse and evaluate Magnetic Resonance – MR images. This provides a quantitative analysis of individual brain structures, and a quantitative comparison of each brain structure with the normative data of a healthy population. The outcome is time-saving for radiologists, thanks to performing segmentation, volume measurement and automatic highlighting of over 30 different regions of the brain. The system measures the grey matter, white matter and cerebro-spinal fluid volume in said 30+ different regions. This allows the neurology department to obtain more detailed results and analyses in a far shorter time.

Alma Medical Imaging have signed collaboration agreements with several companies for the integration of artificial intelligence algorithms in their Alma Health Platform. One such agreement is established with Qubiotech, whose combination of algorithms with tools for nuclear medicine imaging visualisation and analysis results in early and objective detection of disorders and diseases of the brain.

Chest CT

AI-Rad Companion Chest CT is another of Siemens Healthineers' products supporting doctors' decision-making processes for the radiological assessment of computerised tomography imaging (CT) of the thorax. It supports radiologists' interpretation of these images, rendering the process faster and more accurate, thus considerably reducing the time taken to document their findings. This system includes the capability for lung lobule segmentation, detection and measurement of pulmonary lesions, segmentation of the heart and detection of calcium, segmentation of the aorta and measurement of its diameter, and segmentation and measurement of vertebral bodies.

Early detection of tumour cell death

Researchers at Massachusetts General Hospital (MGH) have shown that nuclear magnetic resonance and artificial intelligence can be used to detect early signs of tumour cell death in response to a new therapy against cancer based on a recent virus that selectively eliminates cancer cells without harming healthy tissue, and which has given hope for the treatment of aggressive brain tumours. To maximise the effectiveness of this approach to treatment, the researchers developed a method involving the application of artificial intelligence for detecting the death of tumour cells caused by said virus. This has allowed the early and speedy detection of tumoral regions that respond to the treatment. Recently, these researchers have implemented this method to quantify the pH of cells and molecular composites in the healthy human brain. This is likely to enhance healthcare, enabling treatments more individually adapted to each patient.

Wrist fractures

In 2018, the FDA approved one of the first artificial intelligence algorithms designed to assist clinical decision-making in relation to wrist fractures. This software, OsteoDetect, uses an AI algorithm that analyses images of wrist X-rays to detect fractures of the distal radius, one of the commonest lesions to this joint. The software marks the fracture location, helping the doctor to detect it.

Forecasting cardiac problems

Researchers at the Society of Nuclear Medicine and Molecular Imaging have developed a Deep Learning (DL) network that can predict heart problems, such as heart attacks, or death. The study included over 20,000 patients and the DL algorithm was capable of detecting adverse heart events using myocardial perfusion imaging (MPI) obtained through single-photon emission computed tomography (SPECT).

Detection of colorectal cancer

Researchers to the University of Tulane have discovered that AI can accurately detect and diagnose colorectal cancer by means of tissue scan analysis. According to these researchers, pathologists regularly evaluate and label thousands of histopathology images to identify whether a patient has cancer. However, their workload has increased significantly, which may lead to undesirable erroneous diagnoses. The Machine Learning (ML) algorithm can recognise colorectal cancer among thousands of images with superior results.

Detection of foetal cardiopathies

Researchers at RIKEN Center for Advanced Intelligence Project (AIP) have conducted studies testing artificial intelligence for improving diagnoses of foetal congenital heart disease. They concluded that the diagnostic results were more accurate when using a graphic interface representing the AI analysis of the foetal cardiac ultrasound detection videos. This may assist in prenatal early diagnostics to improve the probability of survival.

Pneumonia related to COVID-19

AI applied to medical imaging has also helped to detect and provide better treatment for complications deriving from COVID-19 during the pandemic. To give an example, Siemens Healthineers has developed an assistant for doctors to interpret chest X-rays: AI-Rad Companion Chest X-ray. The algorithm is able to detect pulmonary lesions, pleural effusion, pneumothorax, consolidation and atelectasis, assisting radiologists in their diagnostic and clinical decision-making processes. This technology has been instrumental in the handling of patients and in clinical decisions made by professionals during the COVID-19 pandemic, as the system, in addition to detecting pneumothorax, pleural effusion and nodules, can indicate consolidations and atelectasis. The latter may be signs of pneumonia caused by COVID-19.



1. FORTUNE BUSINESS INSIGHTS. Medical Imaging Market Size, share & COVID-19 impact análisis, by application, by end user, and regional forecast, 2021-2028. January 2022 [consulted 05-04-22]. Available at:

2. GE HEALTHCARE. Artificial Intelligence in Healthcare: 10 Questions: AI and machine learning technology are the rising stars of healthcare, empowering providers to deliver faster, more precise care. 2019 [consulted 05-04-22]. Available at:

3. VITECH. Benefits of implementing AI in medical imaging: Most significant use cases. September 2021 [consulted 05-04-22]. Available at:

4. NHS. Diagnostic Imaging Dataset Statistical Release. December 2021 [consulted 05-04-22]. Available at:

5. SIGNIFY RESEARCH. Medical Imaging AI Market Projected to Reach $1.2 Billion by 2025. July 2021 [consulted 05-04-22]. Available at:

6. RESEARCH AND MARKETS. Artificial Intelligence In Medical Imaging Market Research Report by End-user, by Application, by Region - Global Forecast to 2027. January 2022 [consulted 05-04-22]. Available at:

7. INE. Una población envejecida. 2021 [consulted 05-04-22]. Available at:

8. HEALTH ECONOMIC TIMES INDIA. There is a surfeit of Indian doctors globally, but we have a shortage of doctors within India: Dr. Sunita Maheshwari. January 2019 [consulted 05-04-22]. Available at:

9. MORDOR INTELLIGENCE. AI MARKET IN MEDICAL IMAGING - GROWTH, TRENDS, COVID-19 IMPACT, AND FORECASTS (2022 - 2027). 2021 [consulted 05-04-22]. Available at:

10. PREDIBLE. LungIQ. [consulted 05-04-22]. Available at:

11. CHINESE MEDICAL DOCTOR ASSOCIATION. 中国放射医师面临的挑战和机遇. 2017 [consulted 05-04-22]. Available at:,%E7%9A%847.5%25%E3%80%82%E2%80%9D%E4%B8%AD%E5%9B%BD%E5%8C%BB%E7%A7%91

12. MINISTRY OF FOREIGN AFFAIRS OF DENMARK. China AI healthcare. 2020 [consulted 05-04-22]. Available at:

13. STATISTA. Percentage of health care applications for AI use in the United States as of 2020. October 2020 [consulted 05-04-22]. Available at:

14. COGNITIVA. El impacto de la Inteligencia Artificial en el análisis de imágenes médicas. July 2020 [consulted 05-04-22]. Available at:

15. STANFORD ML GROUP. CheXNeXt: Deep learning for chest radiograph diagnosis. [consulted 05-04-22]. Available at:

16. BCG. Chasing Value as AI Transforms Health Care. 2019 [consulted 05-04-22]. Available at:

17. THE JOURNAL OF NUCLEAR MEDICINE. Machine Learning with 18F-Sodium Fluoride PET and Quantitative Plaque Analysis on CT Angiography for the Future Risk of Myocardial Infarction. January 2022 [consulted 05-04-22]. Available at:

7 views0 comments
bottom of page