Artificial Intelligence and the future of Healthcare delivery

Written by Gandoul Abdessamiaa

Doctor in the making , Rennes University

Machine Learning Engineering Nanodegree Student

This article provides a glimpse at the potential uses of AI technology in clinical practice and considers the possibility of AI replacing physicians in some medical fields:

What is Artificial Intelligence?

Artificial intelligence (AI) is by definition the branch of computer science concerned with making computers behave like humans. The term derives from the Czech word” robota”, meaning biosynthetic machines used as forced labour. In this Context, Leonardo Da Vinci’s lasting heritage is today’s flourishing use of robotic-assisted surgery, named after him, for complex urologic and gynecologic procedures.

AI, described as the science and engineering of making intelligent machines, was officially coined in 1956 John McCarthy at the Massachusetts Institute of Technology.

The term is applicable to a broad range of items in medicine such as robotics, medical diagnosis, medical statistics, and human biology — up to and including today’s “omics”.

Artificial Intelligence in Medicine :

AI in medicine, which is the focus of this article, has two main branches: virtual and physical:

The virtual branch includes informatics approaches from deep learning information management to control of health management systems, including electronic health records, and active guidance of physicians in their treatment decisions.

The physical branch is best represented by robots used to assist the elderly patient or the attending surgeon. Also embodied in this branch are targeted nanorobots, a unique new drug delivery system. Likewise, the societal and ethical complexities of these applications require further reflection, proof of their medical utility, economic value, and development of interdisciplinary strategies for their application.

We all know that we, doctors, diagnose diseases based on personal medical histories, individual biomarkers, simple scores, and our physical examination of individual patients. In contrast, AI can diagnose diseases based on a complex algorithm using hundreds of biomarkers, imaging results from millions of patients, aggregated published clinical research from PubMed, and thousands of physician’s notes from electronic health records (EHRs). In fact, Physicians in everyday clinical practice are under pressure to innovate faster than ever because of the rapid, exponential growth in healthcare data.

“Big data”is by definition : ‘’an extremely large data sets that cannot be analyzed or interpreted using traditional data processing methods.’’ Actually, big data itself is meaningless, but processing it offers the promise of unlocking novel insights and accelerating breakthroughs in medicine which in turn has the potential to transform current clinical practice.

What are applications Of AI in Healthcare?

Let us look at the emerging uses of AI in medicine :

1/ Image recognition technology :

Using The same technology that Facebook use for image recognition : (analyzing and contextualizing the content of images in detail and compares them to identify similarities and thereby determine what is being shown in an image. ) image recognition technology can analyze data from imaging studies such as EKGs, EEGs, X-rays, MRIs, and CT scans, compiling millions of interpretations of results by expert physicians to make more accurate diagnoses and might make predictions or recognize diseases as effectively as or even better than doctors :

- A group of searchers from Google performed an AI technique called convolutional neural network (CNN) machine learning and demonstrated that AI achieves image-level scores above 97% on both the Camelyon16 test set (metastasis detection of lymph nodes) and an independent set of 110 slides, compared to a human pathologist who achieved 73.2% sensitivity.

- Another group of Scientists has tested CNN performance against 21 board-certified dermatologists on biopsy-proven clinical images, with two binary case studies: ( keratinocyte carcinomas VS benign seborrheic keratosis, and malignant melanomas VS benign nevi. They found that CNN performs skin cancer classification at a level of competence comparable to dermatologists.

2 / Motion Recognition Technology

Scientists have developed supervised machine learning algorithms for complex motion phenotypes obtained from cardiac MRIs and found that the patterns of cardiac motion were associated with increased survival rates in patients with PH, compared with conventional parameters.

3/ Electronic prescriptions

AI may facilitate communication between physicians and patients by decreasing processing times, thereby increasing the quality of patient care. Electronic prescription as it exists today may be problematic since sometimes patients cannot get their medications due to mismatches between prescribed medications and insurance company rules or system errors.

4/ Scheduling conflicts solution

AI could prioritize appointment scheduling based on the risk of readmission and overall severity of an illness in order to reduce readmissions.

5/Assisting patient triage based on symptoms

AI may assist patient triage based on symptoms. For example, the digital health firm HealthTap developed “Dr. A.I.,” which operates based on past medical history and knowledge extracted from experienced physicians, and asks patients to specify symptoms to triage whether they should go to the ED, urgent care, or a primary care doctor.

6/Body censors Technology

In the near future, body censors for blood sugar, hematocrit, oxygen saturation, HbA1C, lipids, infection, and inflammation biomarkers, which are signs of volume overload or dehydration, will also be integrated into AI technology.

7/Computer assisted surgery

Robotically-assisted surgery was developed to overcome the limitations of pre-existing minimally-invasive surgical procedures and to enhance the capabilities of surgeons performing open surgery. As a future doctor, I can assure you that a third hand for a surgeon would open up incredible horizons for all its advantages: Shorter hospitalization Reducing pain and discomfort, Faster recovery time and return to normal activities, Smaller incisions, reducing the risk of infection Reducing blood loss and transfusions Minimalizing scarring. Further advantages are articulation beyond normal manipulation and three-dimensional magnification

AI in medicine: major catalysts in the business of artificial intelligence :

We are obviously living in the Fourth Industrial Revolution, which is characterized by a range of new technologies that are fusing the physical, digital and biological worlds, impacting all disciplines, economies and industries, and even challenging ideas about what it means to be human.

Therefore, In my humble opinion, healthcare will be the lead industrial area of such a revolution and one of the major catalysts for change is going to be artificial intelligence : AI healthcare market is expected reach $6.6 billion by 2021, representing a compound annual growth rate (CAGR) of 40 % over the next five years, according to a new report from ‘’ Accenture’’. Many Companies have to Anticipate The Wave: IBM’s Watson Oncology, Microsoft Hanover project And Google’s Deep mind platform .

Could AI Really Replace a Doctor?

In my Conceit, The human Mind is the most sophisticated machine which means that artificial intelligence cannot, in any case, replace a doctor. In fact, It’s a myth, (except for image recognition and other fields ).On the contrary, it is a great opportunity to be assisted in our work in order to minimize the risk of medical malpractice and perfect Our medical procedures, given its limitations :

- First, AI cannot engage in high-level conversation or interaction with patients to gain their trust, reassure them, or express empathy, all important parts of the doctor–patient relationship.

- Second, AI sensors may extract valuable information ( for example volume status or inflammatory cytokines) to help diagnosis, but traditional physical exams performed by physicians are still needed, especially in areas such as neurology, internal medicine, and rheumatology, which requires high-level interaction and critical thinking.

- Third, AI may reach the point where it can conduct real-time CT scans or other physical scans to detect diseases, physicians will still be needed for interpretation in ambiguous conditions to integrate medical histories, conduct physical exams, and facilitate further discussion.

Last but not least AI in medicine remains very promising, although still in its infancy. Despite the advancement of big data, further research and clinical trials are necessary to confirm the feasibility and validity of clinical AI before it actually Takes a Big place in a large area of clinical practice, particularly in image recognition which should push us, to reconsider our way of thinking as physicians.

References :

Artificial_intelligence_in_healthcare.Retrivedfromhttps://en.wikipedia.org/wiki/Artificial_intelligence_in_healthcare

Herper M. MD Anderson benches IBM Watson in setback for arti- ficial intelligence in medicine. Available at: www.forbes.com/sites/ matthewherper/2017/02/19/md-anderson-benches-ibm-watson-insetback-for-artificial-intelligence-in-medicine/#63077d5f3. Accessed august 18, 2017.

Holmes JH, Peek N. Intelligent data analysis in biomedicine. Journal of Biomedical Informatics 2007;40(6):605–8.

Rosenkrantz AB, Lepor H, Taneja SS, Recht MP. Adoption of an integrated radiology reading room within a urologic oncology clinic: initial experience in facilitating clinician consultations. J Am Coll Radiol 2014;11:496–500.

Trepetin S. Privacy in context: the costs and benefits of a new de-identification method. Ph.D. dissertation (computer science). Cambridge, MA: Massachusetts Institute of Technology; 2006