- Feb 14
- 7 min read

Artificial Intelligence in healthcare

Modern medicine is continuously evolving. Most of the patients now visit a doctor after reading, what they think is important, from internet (google) articles about his/her complaints or abnormal findings. These people think themselves to be clever, are opinionated, and try to influence the physician wanting treatment of their own choice. They forget the fact that physicians are educated about the various problems for which they visit and try to act smart during the interactions. On this day of the internet, internationally accepted management guidelines are published periodically and patients or the patient parties forget that a physician possesses the uncanny ability to relate various complaints with the presenting signs and symptoms to arrive at a diagnosis. The physician also can clinically examine a patient and logically arrange the various findings, disregarding the unnecessary way, in deduction of the ailment. Reading between the lines is one other attribute physicians come to learn with experience and this sharpens with time. However, artificial intelligence (AI) is now making headlines and impacting medical practice significantly.

Alan Turing proposed artificial intelligence in the middle of the 20th century coincident with the gradual upgradation of the analytical machines used in WWII for deciphering encrypted mathematical German codes. The name of John von Neumann is also mentioned for theorizing AI along with Turing. Turing proposed an analytic machine that could learn and correct itself independently from experience.

These analytic machines, now better known as computers, formed by the arrangement of electronic circuitry, can understand 0 and 1 only signifying whether or not there is flow of current through a particular portion of the circuit. This is the basic binary language and further machine and intuitive learning followed.

These machine learning and self-problem-solving attributes form the basis of AI. 1956 saw the acceptance of the concept of AI as a separate speciality in computer science. John McCarthy was the unwritten father and the works of some dedicated people like Minsky, Holland, Simon, Newell, etc., infused imagination in AI. Algorithms, or a specific set and arrangement of codes, were specifically developed with AI in mind and LISP was the first such code. However, LISP was not popular and the interest waned temporarily in the 70s only with a gradual renewal with an increase in practical application.

From the beginning computer scientists were mystified by the complex interaction of the nerve fibres and the cells of the brain with the maintenance of an order, to activate the response of organs or systems properly, and with precision. Mimicking the activities of the human brain is well-nigh impossible. First, artificial intelligence is man-made, where all the data is fed by man and two human beings are never equal. Computers are provided with the data that is considered most appropriate and free from any influence by a group of computer technologists who have been specifically assigned this task. Hence, computer-created intelligence can be different between individual parties. To overcome this and bring universal acceptance, computer-based search engines are employed and the data that is acceptable to everyone in the literate community is introduced. Various algorithms like support vector machine (SVM), gradient boosting, K-nearest neighbour, Random forest (a neural network) that shows the accuracy rate), etc, are used and the server fed with an enormous amount of data analyzed from prospective, retrospective, meta-analytic, and blinded studies. Further modifications occur with deep learning and problem-solving abilities of the central server. AI algorithms became a choice in the development of robotics, voice assistants, games, agriculture and health care. Coded language was developed to represent every human noticed sign exemplified by the recent innovation of the Generative Pre-trained Transformer or GPT allowing search and message response within a very short time. Intuitive learning, larger memory, and the development of genetic algorithms (so-called because the codes resemble the base-pair sequencing arrangements in chromosomal DNAs which are a congregation of genes).

In today's world AI has infiltrated almost every aspect of our lives and somehow we the common people believe almost everything that is churned out by the computer. Only recently articles on matters about deception, deep fake, and simulated con acts are making rounds enthusing the common man to use the computer responsibly and stop forwarding incoming messages without verifying the source and authenticity.

Incorporation of new programming languages like Java, C, C+, C++, C#, Python, etc., enable appliances to display live images of internal structures. Foray into the world of health started with the transcription of medical documents in an orderly fashion, and the impact of AI in the medical world is different and varied. AI-generated anticipated suggestions help type prescriptions and reports faster. All the recording devices use AI algorithms, intuitive and genetic learning for interpretation followed by generation of printed reports, especially for computer storage, and preparation of reports for automated investigations. What was done previously by manual means, is now replaced by AI-powered machines where problem-solving and self-learning capabilities have reached the desired level. Pattern recognition and counting algorithms allow these appliances to deliver consistent, error-free, and rapid results. Most of the cell counters, Holter monitors (where continuous electrical activity of the heart as ECG is recorded), hormone and drugs level monitors, serum Na & K detectors, CRP & pro-cal level assessors, lipid levels, liver and lung function y, are examples. Automated chromatography and spectroscopy are converted samples by coding for computer language and then are fed for analysis and generation of a report. A micro-level chemical reaction is induced, in many situations, with the introduction of a small amount of the blood sample that travels forward across various stations containing reagents. The reaction produces products that are further made to react in a stepwise fashion. Identification and quantification of substances in the sample according to the advice of the physician can now be programmed.

X-ray machines are no longer bulky like before and can be linked to a computer with artefact removal codes. The resultant picture appears sharp. Subtraction angiography or skeletal reconstructions with volume rendition provide a life-like image. AI and advanced coding processes even allow virtual reconstructed images of the interior of the air-filled respiratory tree - virtual bronchoscopy or bronchography. Aided by AI, the processed images are sharper and detailed. Early pathologic changes with personalized planning are now available to a surgeon who is ready to take the plunge. Even densitometric studies aiding bone absorption of required minerals are possible.

Ultrasound was first introduced for visualizing the gravid uterus (obstetric ultrasound). 2-D ultrasound is nowadays a common examination. It can be a still film, a 2-D film, or a motion mode. The examination is subjective, depending on the operator and his/her interpretation. The 2-D imaging has been further improved by depth perception, as in 3-D imaging, and reconstructive imaging. Specialized algorithms are needed that interact with 2-D images taken from multiple angles to create an image with depth perception. Moving internal structures produce a cine effect, and this live streaming during ultrasonography is called 4-D ultrasonography. This modalíty allows a real-time cine imaging of internal organs and a better perspective of the working principle of especially internal moving organs, 4-D ultrasound finds best use in --

i) assessing cardiac structures, such as contractility, valvular defects and leaflet coaptation, intracardiac defects, etc., and
ii) obstetrics & gynaecology, e.g., assessment of abnormal vascularity in suspicious growths, assessment of intrauterine fetal movement & development, baby facing, pelvic assessment, etc.

Computerized tomographic scans and magnetic resonance scan reports are more attractive and always accompanied by a volume rendition or 3-D reconstructed photograph. The formal text report is also AI-generated and smarter.

Endoscopic visualization of cavitary organs and pathological pattern recognition enabled faster lesion detection. Intestines and the airway were ideally suited. As there are both oral and abnormal openings, almost the whole of the intestines can be visualized. Integration of a high-definition camera makes clearer pictures. The normal oral to aboral intestinal movements are utilized for a capsule endoscopy which allows internal photography of the total length of the intestine without any discomfort. AI helps in pattern recognition and the generation of reports.

Interpretation of data, derived calculations, pattern recognition, and incorporation of high-definition cameras are the hallmarks of these investigations.

AI has enabled human genetic coding, considered impossible due to the sheer size at one time. Only a part of the code may need to be written and the next part is anticipated and written automatically by the machine using AI. This not only increased our genetic knowledge but also developed treatment methods for diseases that had no cure before. Certain genetic disorders that were not compatible with an acceptable quality of life could be terminated in utero sparing the lady of the turmoils of progression of pregnancy.

DNA splicing by the CRISPR-CAS technology enabled localizing genetic localisation and the laboratory multiplication (polymerase chain reaction, PCR) of these genetic loci by mRNAs helped in the mass production of vaccines against virulent rapidly spreading diseases. Machine and intuitive learning have made such epochal procedures available to common people. Artificial intelligence optimizes vaccine production and helps in tailoring the product with patients.

The role of AI in the mass production of pharmaceutical products is well understood. Manufacturing targeted drugs and drug delivery systems, both monoclonal antibodies and other agents, which can destroy or prevent further multiplication of unnecessary cells, is a recent interest. It has not only modified cancer chemotherapy but also made certain chronic diseases, hitherto untreatable, curable now.

Minimal access procedures, endoscopic operations and remote robotic procedures appear attractive with liberal application of AI in one form or another. Surgical procedures and new methods, liberal use of prosthetics, use of soft tissue protectors, bone-sparing access, and simple approach have all contributed. AI has made possible pre-procedural planning of a surgery meaningful and the use of robotics in selected cases has ushered improvement further.

The application of AI is seen in all aspects of medicine. Modification of modern-day health care has been possible by the influence of AI, minimisation of appliances, adoption of simple and less invasive techniques, and Incorporation of virtual simulation of real-life scenarios altogether. Thus today's healthcare delivery has become prompt, efficient, and eye-catching. Experience, empathy, thinking capacity, farsight, skill, and the ability to make a prompt decision are the only human touches left in modern scientific health care.

References :

Artificial Intelligence in healthcare

Recent Posts