Status of artificial intelligence in Pakistan and its implications in anesthesiology

Muhammad Imran Khan 1 , Rana Shahzad Murtaza 2 , Muhammad Haider Ali 2 , Muhammad Sulaman Ashraf 3 , Sahar Yazdani 4 , Anum Yaseen 5
Author affiliation:
  1. Postgraduate Resident Anesthesiology, Choudhary Pervaiz Elahi Institute of Cardiology, Multan, Pakistan.
  2. Medical officer, Anesthesia, Choudhary Pervaiz Elahi Institute Of Cardiology, Multan, Pakistan.
  3. Medical officer, Intensive Care, Choudhary Pervaiz Elahi Institute Of Cardiology, Multan, Pakistan.
  4. Clinical Fellow, Tameside Hospital, Manchester, UK.
  5. Medical Officer, Iqbal Hospital, Bahawalpur, Pakistan.
Correspondence: Dr Muhammad Imran Khan; E-mail: imrkhan803@gmail.com; Phone: +923326366121

Abstract
Artificial intelligence (AI) refers to the technology that enables machine-based intelligent behavior. This descriptive article gives an overview of the implementation of artificial intelligence in the following disciplines within the field of clinical medicine: Anesthesiology, Cardiology, Pulmonology, Endocrinology, Nephrology, Gastroenterology, Neurology, Histopathology, and Radiology. In this article, we will discuss the employment of artificial intelligence in anesthesiology and provide an analysis of the several areas in which it is being utilized. We will also evaluate the current status of AI and its future in Pakistan. The importance of this manuscript is to highlight the benefits of artificial intelligence so that the benefits can be realized in Pakistan.

Key words: Artificial intelligence; Anesthesiology; Clinical medicine; Endocrinology; Gastroenterology; Histopathology; Nephrology; Neurology; Pakistan; Radiology

Citation: Khan MI, Murtaza RS, Ali MH, Ashraf MS, Yazdani S, Yaseen A. Status of artificial intelligence in Pakistan and its implications in anesthesiology. Anaesth. pain intensive care 2022;26(1):110-114;

DOI: 10.35975/apic.v26i1.1776
Received: September 12, 2021, Reviewed: November 24, 2021, Accepted: December 10, 2021


Introduction
Artificial intelligence (AI) has altered the way many firms operate by assisting them in doing their tasks more effectively.1 AI is basically a computer program that is capable of performing tasks typically associated with human intelligence, e.g. complicated problem solving, decision making, object detection, etc. AI makes use of the complex blend of mathematics and computer science. AI can provide outcomes that are more accurate than that of human beings.2  AI performs more convenient work because it doesn’t need any break and refreshment time compared to the human beings. It constantly performs its work for long time duration without facing any problem. There are no factors influencing the performance of AI as opposed to factors which influence the work of human beings.3
Background
Alan Turing created the "Turing test" in 1950 to determine if a computer could do a task indistinguishably from that of a person. After the 1990s, several AI approaches were used in the healthcare sector, including Bayesian networks, fuzzy expert systems, artificial neural networks, and deep learning, etc.4 After the successful deployment of AI methods in a variety of sectors including healthcare, a significant investment was made specifically for the development of AI-based technologies for the healthcare sector.

The use of AI in medicine is divided into two parts. One part is physical and includes all robotic applications which can assist during surgeries, early diagnosis, and treatment of handicapped patients.5 While the other element is virtual, which includes all programs that assist experts in making treatment decisions, such as massive recorded data systems analysis using neural networks.

Implementation of AI in healthcare system
The implementation of AI-based techniques gives a power to healthcare professionals to search for appropriate solutions for emerging clinical problems.6 Healthcare sectors have a huge amount of medical data that are gathered from several areas of the medical sector, data that can easily be utilized by a deep learning algorithm. Currently, there are some departments of clinical medicine that use AI-based techniques and get advantages from them.7 Some examples include: use of ECG in the detection of atrial fibrillation, detection of electrical activity in brain (like during epileptic seizures), detection of the sugar level in the blood, and detection of disease, related to deep tissue examination through imaging techniques in radiology, anesthesia, intensive care we well as interventional pain management.8 This advanced technology helps healthcare professionals to indulge in the concept of personalized treatment.9
Application of AI in medicine
The following section discusses current uses of AI in medicine.10
Cardiology
AI transformed medicine for the first time in 2014 when it was used for the identification of atrial fibrillation using digital ECG monitoring equipment. These remote device applications made it possible to identify arrhythmias. Apple's wristwatch capable of detecting atrial fibrillation has been cleared by the FDA. However, despite this, there are some limitations to this approach, such as the false positive rate that is shown to the public9. Through AI-based techniques like deep learning and neural network, machines are able to detect a risk of cardiovascular disease in people with high accuracy.10
Pulmonology
In pulmonary function testing, AI plays a critical role by allowing software specialists to more accurately interpret the test results and proceed with subsequent treatment procedures.11
Endocrinology
Numerous businesses have obtained FDA permission to produce a smart monitoring gadget for diabetes patients that enables frequent monitoring of blood glucose levels.12
Nephrology
AI also helps in the treatment of kidney disease by detecting the rate of filtration in the patient and also predicts the risk of mortality related to nephrology.13
Gastroenterology
Deep learning and neural network approaches of AI system help in the endoscopy and ultrasound procedure by aiding in the diagnosis of precancerous and cancerous lesions so as to facilitate the patient’s treatment by earlier detection of problematic lesions.14.
Neurology
Electroencephalography that use AI to identify epileptic episodes in patients. This gadget utilizes the most advanced technology available for improving seizure therapy via reliable monitoring. Gadgets like BIS, NIRS etc. powered by AI are critical for objectively detecting problems in neurological patients.15
Histopathology
There is a big challenge for a healthcare professional to recognize cancer cells in patients. By the implementation of an AI-based algorithm, the professionals are able to find the cancerous tissue with a high accuracy rate and successfully remove the tumor margin.16
Radiology
Radiologists used a deep learning approach in imaged-based diagnosis.  Studies are done on making appropriate models through CT scanner, MRI, and x-ray for next-generation radiologist tool development.17
Role of AI in anesthesia
The term “Anesthesiology” refers to all measures that are concerned before, after, and also during the patient surgical operations. In its perioperative preparations includes giving anesthesia, critical care medicine, emergency treatment, and pain management. In today's advanced technological period, when many areas profit from AI-based solutions, AI computers can oversee the job of anesthesiologist autonomously.18 There are different areas of anesthesiology where AI techniques are utilized.

Measurement of depth of anesthesia
Several studies have shown improvement in the measurement of depth of anesthesia by using AI. Especially in electroencephalography testing, it shows brain activity when the brain cells communicate through electrical impulses to find out the depth of anesthesia at the time of BIS monitoring. The use of machine learning AI helps to analyze a complex range of electroencephalography signals to evaluate the effects of certain drugs in patients.19
Standard control over anesthesia delivery
AI helps to construct a standard closed-loop anesthesia delivery system with real-time monitoring. According to this, the anesthesiologist used an automated medication administration strategy to ensure the patient's safety by maximizing clinical indicators or markers that indicate the level of anesthesia. Apart from blood pressure as a vital indicator, BIS measurement using AI is being used to establish control over anesthetic administration. To facilitate the endotracheal intubation a deep machine learning approach is used to deliver the neuromuscular blockade automatically.20-22
Complication event prediction
There are many approaches based on AI, such as machine learning, neural networks, and fuzzy logic that are being used to assess the possibility of danger occurrence through hypnotic effect assessment. Throughout surgery, a fuzzy neural network can automatically assess the patient's awareness after general anesthetic, assisting in decision-making and adjusting anesthetic during the procedure. There are two locations where one must identify event occurrences: the intensive care unit (ICU) and the operating room.21 Here, machine learning is used to evaluate a vast database in order to forecast any event, such as morbidity, mortality, or sepsis.

Assist ultrasound-based practice
The neural network approach from an AI-based network is most commonly used to assist the ultrasound-based practice by its accurate imaging features.22 By using this technique professionals easily distinguish between the two images' differences. Additionally, these approaches aid in identifying anatomical levels for the administration of anesthetic from the spinal cord insertion in ultrasound imaging.

Pain management services
AI specialists control discomfort during preoperative measurements in order to aid patients. By using a deep learning method, professionals were able to evaluate MRI data in order to distinguish between painful and non-painful stimuli.23 Machine learning is also used to analyze the photo plethysmography and skin conductance wave data of patients. Machine learning also applies for detecting the patient who responds to acute pain for therapy.23
Management of operating room
AI also plays a role in evaluating the aspects affecting operating room logistics, such as time management in OT or sensing an anesthesiologist's activities. This technique used large amounts of electronic data and a trained neural network method to forecast the overall duration of the surgical operation, as well as the kind of surgery procedure24. Tags were examined using machine learning to monitor anesthesiologists' interactions with operating room equipment.

Current status of AI in Pakistan
Numerous industries use AI in their operations on a worldwide scale. Pakistan is also developing into a significant market for AI-based solutions. Pakistan uses AI technology in a variety of industries to provide better answers to existing and emerging challenges and to boost demand in the new age.25 Pakistan enhanced its self-security systems as a result of this, including missiles powered by AI, cyber security, and very effective cameras. There has been a noticeable increase in the spectrum of AI applications in Pakistan during the last several years.

PIAIC project in Pakistan
To involve the youth generation the Pakistan government initiated the PIAIC project for the development and progress of the country's economy. The government develops national-level AI centers around the country.26 Currently, commercial companies in Pakistan are using AI in their operations by engaging top AI professionals. The following AI laboratories are located in Pakistan:
  • The National University of Sciences & Technology (NUST) – National Center of AI, Islamabad.
  • Artificial intelligence research lab (AIRL) Alkhawarizmi Institute of Computer Science(KICS), University of Engineering & Technology (UET), Lahore.
  • The Center of Intelligent Systems and Network Research (CISNR) University of Engineering & Technology (UET), Peshawar.
  • The Artificial Intelligence Lab, The Institute of Business Administration IBA (The AI Lab), Karachi
Pakistan has to restructure its faculty in the area of medicine to include AI-based learning and also take initiatives to improve the appeal of new students to high-technology learning approaches aimed at transforming the medical sector via technological services. Implementation of these approaches would catapult Pakistan's economy to new heights.26
The Future of AI in Pakistan
AI has a significant influence on the economic growth of several nations worldwide. Pakistan, on the other hand, is making relatively modest progress in implementing AI applications in a variety of disciplines, including medical. Several educational institutions and corporate companies, for example, NUST University, have introduced this AI course and department in order to compete in the market. Pakistan's agricultural department also employs a machine learning strategy to deal with natural disasters in the country.27
Conclusion
In conclusion, the deployment of AI-based approaches improves the precision of executing diverse tasks. AI assists healthcare providers by transforming the field of medicine with its accurate prediction capabilities. In clinical medicine, AI-based approaches are applied in a variety of fields, including anesthesiology. This article discussed numerous areas of anesthesiology in which AI-based solutions are being used to improve patient outcomes. Finally, we noted that Pakistan is a long way from reaping the benefits of this industry. Pakistan's healthcare industry is unlikely to benefit from this breakthrough in the future without further developing an expertise in AI.

Conflict of interest
The author declares no conflict of interest.

Authors contribution
The author is solely responsible for literature search and drafting and review of this manuscript.

References
  1. Davenport TH, Ronanki R. Artificial intelligence for the real world. Harvard Bus Rev. 2018;96(1):108-116. [Free Full Text]
  2. Acemoglu D, Restrepo P. 8. Artificial Intelligence, Automation, and Work. In: Agrawal A, Gans J, Goldfarb A (ed.) The Economics of Artificial Intelligence: An Agenda. Chicago: University of Chicago Press; 2019. p.197-236. DOI: 7208/9780226613475-010
  3. Wu H, Han H, Wang X, Sun S. Research on artificial intelligence enhancing internet of things security: a survey. IEEE Access. 2020;8:153826-153848. DOI: 1109/ACCESS.2020.3018170
  4. Amisha, Malik P, Pathania M, Rathaur VK. Overview of artificial intelligence in medicine. J Family Med Prim Care. 2019 Jul;8(7):2328-2331. [PubMed] DOI: 4103/jfmpc.jfmpc_440_19
  5. Lorkowski J, Grzegorowska O, Pokorski M. Artificial Intelligence in the healthcare system: an overview. Adv Exp Med Biol. 2021;1335:1-10. [PubMed] DOI: 1007/5584_2021_620
  6. Morley J, Floridi L. An ethically mindful approach to AI for health care. Lancet. 2020 Jan 25;395(10220):254-255. [PubMed] DOI: 1016/S0140-6736(19)32975-7
  7. Panch T, Szolovits P, Atun R. Artificial intelligence, machine learning and health systems. J Glob Health. 2018 Dec;8(2):020303. [PubMed] DOI: 7189/jogh.08.020303
  8. Briganti G. Augmented Medicine: Changing Clinical Practice with Artificial Intelligence. In: Advances in Artificial Intelligence, Computation, and Data Science. Springer, Cham; 2021. p. 333-339. DOI: 1007/978-3-030-69951-2_14
  9. Becker A. Artificial intelligence in medicine: What is it doing for us today? Health Policy Tech. 2019;8(2):198-205. DOI: 1016/j.hlpt.2019.03.004
  10. Dorado-Díaz P, Sampedro-Gómez J, Vicente-Palacios V, Sánchez P. Applications of artificial intelligence in cardiology. the future is already here. Revista Española de Cardiología. 2019;72(12):1065-1075. [PubMed] DOI: 1016/j.rec.2019.05.014
  11. Das N, Topalovic M, Janssens W. Artificial intelligence in diagnosis of obstructive lung disease. Curr Opin Pulm Med. 2018;24(2):117-123. [PubMed] DOI: 1097/MCP.0000000000000459
  12. Dvey-Aharon Z, Huhtinen P. Screening for diabetic retinopathy in endocrinology clinics by using handheld cameras and applying artificial intelligence algorithms. J Endocrine Soc. 2021;5(Suppl 1):A419-A425. DOI: 1210%2Fjendso%2Fbvab048.856
  13. Niel O, Bastard P. Artificial intelligence in nephrology: core concepts, clinical applications, and perspectives. Am J Kidney Dis. 2019;74(6):803-810. [PubMed] DOI: 1053/j.ajkd.2019.05.020
  14. Yang YJ, Bang CS. Application of artificial intelligence in gastroenterology. World J Gastroenterol. 2019 Apr 14;25(14):1666-1683. [PubMed] DOI: 3748/wjg.v25.i14.1666
  15. Raghavendra U, Acharya UR, Adeli H. Artificial intelligence techniques for automated diagnosis of neurological disorders. Eur Neurol. 2019;82(1-3):41-64. [PubMed] DOI: 1159/000504292
  16. Sultan AS, Elgharib MA, Tavares T, Jessri M, Basile JR. The use of artificial intelligence, machine learning and deep learning in oncologic histopathology. J Oral Pathol Med. 2020 Oct;49(9):849-856. [PubMed] DOI: 1111/jop.13042
  17. Y Yasaka K, Abe O. Deep learning and artificial intelligence in radiology: Current applications and future directions. PLoS Med. 2018 Nov 30;15(11):e1002707. [PubMed] DOI: 1371/journal.pmed.1002707
  18. Hashimoto DA, Witkowski E, Gao L, Meireles O, Rosman G. Artificial intelligence in anesthesiology: current techniques, clinical applications, and limitations. Anesthesiology. 2020 Feb;132(2):379-394. [PubMed] DOI: 1097/ALN.0000000000002960
  19. Lönnqvist PA. Asleep or awake: is paediatric regional anaesthesia without general anaesthesia possible? Br J Anaesth. 2020 Aug;125(2):115-117. [PubMed] DOI: 1016/j.bja.2020.04.071
  20. Wingert T, Lee C, Cannesson M. Machine learning, deep learning, and closed loop devices-anesthesia delivery. Anesthesiol Clin. 2021 Sep;39(3):565-581. [PubMed] DOI: 1016/j.anclin.2021.03.012
  21. Xia M, Xu T, Jiang H. Progress and perspective of artificial intelligence and machine learning of prediction in anesthesiology. J Shanghai Jiaotong Uni (Sci). 2021;1-9. DOI: 1007/s12204-021-2331-3
  22. Bowness J, El-Boghdadly K, Burckett-St Laurent D. Artificial intelligence for image interpretation in ultrasound-guided regional anaesthesia. Anaesthesia. 2021 May;76(5):602-607. [PubMed] DOI: 1111/anae.15212
  23. Wang R, Wang S, Duan N, Wang Q. From patient-controlled analgesia to artificial intelligence-assisted patient-controlled analgesia: practices and perspectives. Front Med (Lausanne). 2020 May 22;7:145. [PubMed] DOI: 3389/fmed.2020.00145
  24. Bellini V, Guzzon M, Bigliardi B, Mordonini M, Filippelli S, Bignami E. Artificial intelligence: a new tool in operating room management. role of machine learning models in operating room optimization. J Med Syst. 2019 Dec 10;44(1):20. [PubMed] DOI: 1007/s10916-019-1512-1
  25. Shahid MZ, Li G. Impact of artificial intelligence in marketing: a perspective of marketing professionals of Pakistan. Global J Manag Bus Res. 2019;19(2-E). [Free Full Text]
  26. Jamil S. Artificial intelligence and journalistic practice: The crossroads of obstacles and opportunities for the Pakistani journalists. Journalism Pract. 2020;15(10):1400-1422. DOI: 1080/17512786.2020.1788412
  27. Abid S, Awan B, Ismail T, Sarwar N, Sarwar G, Tariq M, et al. Artificial intelligence: medical student s attitude in district Peshawar Pakistan. PJPH. 2019;9(1):19-21. DOI: 32413/pjph.v9i1.295
 

My Most Memorable Patient
If you can, you must
B.M. Munasinghe
Acting Consultant Anaesthetist, District General Hospital, Mannar, Sri Lanka

I was an anesthesia registrar in 2016, and had completed the cardiothoracic anesthesia rotation. It was Sunday, when I arrived early morning in ICU. All of the patients were stable and many were about to be shifted to HDU. A postoperative patient arrived in the ICU in the early hours following an emergency coronary artery bypass graft surgery. He was a 79-year-old gentleman with diabetes, hypertension and recent extensive anterior ST-elevation myocardial infarction. He had left main coronary stenosis with 99% block on coronary angiography. The surgery had been performed at night while the patient was on dual antiplatelet drugs and subcutaneous enoxaparin in therapeutic doses. Intraoperatively, four vessels were grafted and he was sent to ICU, an hour before me. He was on ventilation.

I noticed that the central venous pressure was 25 cmH2O and his arterial lactate level increased over the last one hour from 3.1 to 6.5 mmol/L. His BP was maintained with noradrenaline infusion (0.15 µg/kg/min) which needed gradual escalation. I immediately checked the mediastinal and intercostal drains but there was minimal output. He was quite pale. When drain tubes were ‘milked’, fresh blood poured from the mediastinal drain. My suspicion of cardiac tamponade due to hemopericardium was justified. Blood transfusion was immediately commenced. I alerted the surgeon, the anesthetist and the OR. Within 15 min, the patient was back on the operating table. 500 ml of blood was evacuated from the pericardium. The grafts were intact. The bleeding was from raw areas around the pericardial surgical incisions which were cauterized and packed. Following a 2-h surgery, the surgeon was satisfied with the hemostasis and routine surgical closure was done. The patient was retransferred to ICU to be closely monitored. During the next 2 h, the same clinical picture was noticed. I performed a transthoracic echo and it confirmed the collapsed right ventricle by hemopericardium. He was rushed to the OR, while blood and blood products were being pumped. For the second time, around 450 ml of blood was evacuated. He was again sent back to ICU. By now, transfusion medicine opinion was sought and point of care testing for coagulopathy (ROTEM) was performed. Platelet transfusion was suggested. Repeat test after platelet transfusion suggested correcting coagulopathy. The surgeon who operated upon the patient was highly skilled, dedicated and experienced. He wearily told me that he had done everything possible for the patient.

Alas! The bleeding again recurred. We transfused factor VII as a last resort. I barely managed to maintain his hemodynamics with blood and blood products and titrated noradrenaline and dobutamine infusions. I insisted the surgeon on one last-ditch effort to which he agreed. Within 8 h, exploration was done for the third time and cardiac tamponade was relieved. We inserted an intra-aortic balloon pump as the blood pressure was in the lower range. His metabolic parameters including serum calcium, potassium, acid-base status and temperature were normalized during the next few hours following the surgery. ROTEM was satisfactory. I managed to wean him off the inotropes and subsequently the intra-aortic balloon pump support over the next 72 h. His ventilatory support was also gradually weaned. By day 7, he was stable hemodynamically and extubated. Fortunately, he did not develop any organ injury during this whole episode. Three days later, he was sent to HDU and subsequently home.

This incident taught me to be vigilant and take evasive action in any medical emergency. By never giving up on potentially salvageable patients, all the hard work done is repaid by a simple smile of a patient. To this day, I am doing my very best medically to revive even the very bad patient, even though at times the outcome might be poor.

Consent: All appropriate informed, written consent was taken from the patient prior to preparation of the manuscript