ARTIFICIAL INTELLIGENCE IN MEDICAL IMAGING
Artificial Intelligence (AI) is a branch of science that simulates human intelligence in computer like machines that are programmed to think like humans and exhibit traits like learning and problem-solving. AI uses different digital inputs to gather multiple arrays of data. The data is then processed by AI analytical models for fast and effective output of information with human-like intelligence. Data processing is usually executed with the assistance of two sophisticated technological subsets of AI, which are Deep Learning and Machine Learning that help the machines to constantly learn and develop its knowledge.
With the advent of AI, the collective output of machines has greatly improved. Some of the most primitive and tedious machine based functions are largely simplified with the use of complex programming algorithms used in AI. Currently, AI has impacted almost every industry sector. In the past few years, the Healthcare sector has also seen a gradual upsurge in the adoption of AI based technology for the different healthcare divisions. AI healthcare is broadly employed for segments like patient care, drug discovery, and personalized treatment. Medical imaging has emerged as one of the most potential healthcare segment with a major technological AI upgradation. This upgradation has proved effectiveness in improving the quality of internal imaging using various radiological imaging techniques such as X-ray radiography, magnetic resonance imaging (MRI), medical ultrasonography or ultrasound, computed tomography (CT), and nuclear medicine functional imaging techniques like positron emission tomography (PET). Integration of AI with such imaging techniques can facilitate reviewing of an image and identify potential findings within it by searching a patient’s history or other parameters related to the particular anatomy scanned.
AI in the healthcare sector has not only improved the accuracy of disease detection but has also reduced the treatment time. The potential of AI in improving medical imaging lies in:
- Improved Automation: The work flow can be synchronized, i.e., the individual radiological instruments can be collectively regulated.
- Better image interpretation: Well taught machines can analyze minute details more effectively.
- Effective diagnosis: AI can predict about the diseases more accurately. Studies show high competence of AI in predicting early stages of cancer.
- Help doctors: AI can represent only the relevant parts in brief. This helps the doctor conclude quickly.
Some really interesting case studies where AI was employed to improve the diagnostic efficacy through medical imaging are as follows:
- Harvard Medical School’s deep learning system can diagnose breast cancer with an accuracy of 97 percent compared to 96 percent of a radiologist. Aided by the diagnostic system, the radiologist’s accuracy improved to 99 percent.
- McMaster University’s deep learning system detects Alzheimer’s disease with an accuracy of 98 percent to 99 percent by using magnetic resonance images, compared to 84 percent accuracy of previous computer vision algorithms.
The case studies above are indicative towards a collective approach, where AI can assist the radiologists to increase the value they provide. This is usually done by training the AI system to recognize normal anatomy through typical scans from CT, MRI, ultrasound or nuclear imaging. The AI algorithms used in machine automation help the machines to read medical images by identifying patterns within the image the way radiologists do. Patent trend analysis of the AI in medical imaging field highlights patent filing growth in the field, wherein companies (Philips, Siemens, Smith & Nephew) originating from USA has been top patent filers in the domain. Other countries like India, Germany, Great Britain are also active in the domain. It has been observed that the companies are filing patents in collaboration with academic institutions, like Yale University and British Columbia University.
The market for diagnostic imaging equipment and devices is relatively mature. The market is controlled by four global conglomerates with a combined market share of 80 percent in diagnostic imaging: GE Healthcare, Siemens, Philips, and Toshiba. These four companies dominate the market in the field of Radiological Information System and Picture Archiving and Communication System (RIS/PACS), and Advanced Visualization (AV). Toshiba is endeavoring to expand to the PACS/image analytics market through collaborations with other companies. The AV market remains highly fragmented, with a large number of small companies competing in niche markets. AI is consistently improving both (RIS/PACS and AV) the approach and general access to reliable and accurate medical image analysis, with help from digital image processing, combined with pattern recognition and machine learning AI platforms. For example, a start-up called Butterfly Network has developed a handheld 3D-ultrasound tool that creates 3D images of the medical image in real time and sends the data to a cloud service, which then identifies image characteristics and automates a diagnosis. This kind of clinical support from AI is expected to have a significant impact on the overall medical imaging diagnosis market and its growth. In further instance, Arterys developed an AI algorithm using MR images to draw up the contours of the heart’s four chambers, measuring precisely how much blood they move with each contraction. Cardiologists usually need 30 to 60 minutes to calculate the volume of blood transported with each pump, but Arterys’ AI comes up with the answer within seconds.
The implementation of AI technology into the healthcare sector has following challenges:
- a) Availability of Structured and Standardized Data: The lack of sufficient quantities of high quality structured and standardized data, as the data in the healthcare industry comes from different sources such as electronic medical records, laboratory and imaging systems, physician notes, and health-insurance claims.
- b) Eating Away Jobs: As Artificial Intelligence gains deeper access to work and personal life, it demonstrates that the biggest threat to mankind is the replacement of humans with machines.
- c) Patient Hesitation: There is a degree of patient pushback against being “forced” to engage with a “machine,” whether the AI system assists with patient engagement/communication or provides high-level clinical related services.
- d) Data security: Security and data privacy is also a significant challenge for Artificial Intelligence. e) The structures of the human body present great variation in terms of normal dimensions and textures, and this variation potentially masks pathological conditions.
The current business strategy among many large companies in diagnostic imaging is to leverage licensing agreements and work collaboratively with technology suppliers, rather than to acquire these companies outright. In order to make up for the lack of commercial funding available from traditional venture capital resources for imaging technology, most of the key imaging OEM’s have established corporate venture funds. For example, Siemens Venture Capital Healthcare, Philips Healthcare Incubator, and the GE Healthymagination Fund.
An important trend in diagnostic medical imaging is a growing interest in fusion and multimodality imaging. As the market for diagnostic imaging equipment matures, new opportunities are emerging for imaging modalities that can be used by mobile doctors or health-care workers in the field. Another major trend is the idea of smaller, portable imaging machineries. The global medical imaging market is facing increasing competition from refurbished systems due to the high cost of devices and installation in developing markets. IBM/Merge, Philips, Agfa, and Siemens have already started incorporating AI into their medical imaging software systems.
The growth of the global Healthcare AI market is directly correlated with the existing economic conditions and health status (e.g. COVID-19 pandemic situation) across the globe. The rising level of disposable income has propelled the spending trends on healthcare. In addition, the improving global economy is expected to take a step further in the years ahead and catalyze the growth of AI in healthcare industry. Increased global investment in the healthcare sector has facilitated AI based developments world-wide. AI has emerged as a potential alternative to the traditional healthcare sector by improving and standardizing the primitive technologies. It is believed that the advent of AI in medical sciences has revolutionized its basis by inculcating a cooperative and synchronized association between machines and doctors.