|
 
 |
| SPECIAL ARTICLE |
|
| Year : 2021 | Volume
: 9
| Issue : 1 | Page : 145-150 |
|
Artificial intelligence in medicine and health sciences
Parameshwar R Hegde1, Manjunath Mala Shenoy2
1 AI Design Engineer, Yenepoya Technology Incubator, Yenepoya (Deemed to be university), Deralakatte, Mangalore, Karnataka, India
2 Department of Dermatology, Venereology & Leprosy, Yenepoya Medical College, Yenepoya (deemed to be university), Deralakatte, Mangalore, Karnataka, India
| Date of Submission | 06-Dec-2020 |
| Date of Decision | 15-Mar-2021 |
| Date of Acceptance | 16-Mar-2021 |
| Date of Web Publication | 26-Jun-2021 |
Correspondence Address:
Dr. Manjunath Mala Shenoy
Department of Dermatology, Venereology and Leprosy, Yenepoya Medical College, Deralakatte, Mangalore
India
 Source of Support: None, Conflict of Interest: None  | Check |
DOI: 10.4103/amhs.amhs_315_20
Artificial intelligence (AI) is being used in almost all aspects of life. The AI can assist medical consultants, primary health workers, and hospital administrators to understand analyze and interpret the medical data. The concepts such as data mining, image and signal processing, computer graphics, and machine learning are being utilized or being tried to implement in different health science areas such as dermatology, radiology, anesthesiology, psychiatry, surgery, and medical records . This article throws lights on some of the AI-related works that were carried out, especially in dermatology and radiology. This review has discussed its utility in other fields and the advantages such as faster execution, reliable results, and advantage over the manual process in certain aspects of healthcare. There are challenges, but further research and advances in technology in AI are likely to enormously benefit mankind.
Keywords: Artificial intelligence, data mining, health sciences, machine learning, medicine
How to cite this article:
Hegde PR, Shenoy MM. Artificial intelligence in medicine and health sciences. Arch Med Health Sci 2021;9:145-50 |
| Introduction | |  |
Human being and animals are capable of displaying natural intelligence that involves consciousness, problem-solving, and emotionality. It is now possible to make machines exhibit these human traits and simulate human intelligence by emplying an artificial neural network in a computing system. This is called artificial intelligence (AI). Utility of AI has already been explored in day-to-day life. Spam filters and auto reply in E-mails, autopilot in commercial flights, analyzing speed of traffic moment by Google Map, mobile banking apps, and product recommendation by amazon and other retailers are some of the examples.
Over the decade, the medical field has been incorporated several technologies to improve diagnosis and patient care.[1],[2] That includes high-quality imaging, robotic surgery, virtual reality (VR), telemedicine, and most importantly machine learning to assist the doctors in decision-making.[3] Machine learning is one of the fields of computer science that predict the results based on mathematical models such as statistical theories, information theory, and certain other fields.[4] It can also be described as it is a subset of AI that allows the machines to learn from the existing data and find a pattern in that.[5]
| Why Artificial Intelligence? | | |
It is not an uncommon to notice social media pages displaying matters that one interested as “recommended for you,” which can save time for the person who is browsing. That is the example of a behavioral study of an individual done by a machine. Browsing history by an individual has been utilized as the data for it.[6] The same mechanism is used in the medical field to study the health-related issues of an individual. The data gathered include information such as gender, age, height, weight, family history, medical history, and treatment history. This can be stored for future references and when the patient revisits the hospital, these stored data will be helpful to overview the medical conditions and further evaluate and recommend medication.[7]
AI is nothing new in medicine. It is being regularly utilized and they are evidenced by several studies in many disciplines of medical sciences mainly in dermatology, radiology, oncology, and even in medical record maintenance [Figure 1]. Relevant aspects of AI in medical sciences are summarized in [Table 1]. To give an overview of utility of AI in medical science and the published literature, this review has been written. | Table 1: List of medical disciplines, activities and respective artificial intelligence techniques
Click here to view |
| Dermatology | | |
Skin diseases are identified based on the symptoms such as rashes, redness, and scaling , and the common findings are elevations, depressions, nail defects, and so on.[8] All these features can be recognized and extracted from an image using image processing algorithms.[9],[10] An application can be designed which utilizes machine learning classifier for pattern recognition of a fresh image and thus aid in the diagnosis. In dermatology, it can be utilized to identify primarily skin malignancies and inflammatory skin diseases. Clinical and dermoscopic images can be utilized for this purpose. Dermoscope is an optical instrument that captures images with a visualization of deeper skin pathology.
In 2019, ALEnezi[11] collected 20 normal, 20 melanoma, 20 eczema, and 20 psoriasis images from different skin disease websites. Convolution Neural Network's (CNN) model was used to classify the images and identify the lesions. Hegde et al., 2018[12] studied chronic eczema, lichen planus, and plaque psoriasis based on color and texture features. Four different classifiers were compared in three feature combinations, i.e., color, texture, and color and texture. They concluded that performance in the Linear Discriminant Analysis along with color features had the better accuracy. Bajaj et al.[13] collected 813 images from five diseases namely eczema, psoriasis, impetigo, melanoma, and scleroderma to diagnose it based on the color feature and achieved good accuracy using the Artificial Neural Network classification method. Herpes, paedures dermatitis, and psoriasis using ten standard samples and twenty test samples were tested based on the texture patterns that were extracted by the Gray Level Co-occurrence Matrix method. Later support vector machine classifier was implemented to identify the diseases.[14] Li and Shen, in 2018,[15] designed a study to detect melanoma skin cancer in its early stages based on dermoscopic features. The international skin imaging collaboration database's 2017 edition images were used. Study showed good system accuracy for image processing steps such as segmentation, feature extraction, and classification. A study has also been performed utilizing the region of interest segmentation from the healthy skin using L*a*b color space and K-means clustering.[16]
There are other utilities of AI in dermatology such as dermoscopic analysis of pigmented skin lesions and hair diseases (trichoscopy). Images obtained can be subjected to a computer-assisted digital dermoscopic image analyzer for the detection of melanoma and other pigmented lesions.[17] Dermoscopic images can also be improved by the removal of noise (unwanted structures) using appropriate software. Psoriasis and lichen planus dermoscopic images were collected from the patients and the noise (presence of hair, ruler markings, and air bubbles which impede the diagnosis) removal was achieved using image preprocessing and segmentation techniques.[18] Trichoscopy has been a new advent where the machine can recognize the telogen (resting) and anagen (growing) hair and has been successful in replacing the conventional trichogram where hairs are pulled from the scalp and analyzed microscopically [Figure 2].[19] Trichogram has hence become noninvasive with the advent of AI.
| Radiology and Oncology | | |
In recent times, the scope of machine learning approaches in the field of medical imaging and radiology has tremendously increased.[20],[21] In cancer diagnosis, radiology image preprocessing, tumor segmentation, diagnosis using image features, symptom-based disease prediction model, and treatment model design and prediction have been frequently used.[22]
In 2019, a system was developed in Iran to detect brain tumors using MRI images.[23] Including normal and tumor classes a total of 1321 images were collected from 153 patients and accuracy for the test data using CNN and decision tree algorithms were very impressive. A visual and explainable case-based reasoning model was designed for breast cancer (BC) with the help of a large sample group from Breast Cancer Wisconsin. A total of 830 cases from the mammographic mass and 286 cases from BC dataset were considered in the study.[24] This study examined three diverse algorithms, namely, K-Nearest Neighbor (KNN), Weighted KNN , and Red/Blue Intersection Algorithm with all three datasets and different combinations. Another study used deep learning techniques to identify BC on screening mammograms.[25] The digital dataset of screening mammograms (CBIS-DDIM) gave high predictions in different models.
Machine learning can ease the major objectives of oncology such as identifying high-risk cases and early diagnosis. It can have major impact in future considering the aging population, increasing prevalence of cancers, availability of accurate diagnostic tests, and demand for precision medicine.
| Medical Data Mining and Artificial Intelligence | | |
Currently, everywhere in the world, hospitals are adopting Electronic Health Record (EHR) system,[26] and the same model is practiced in many healthcare facilities in India. The major advantage of this approach is rapid and accurate data collection and storage. The concepts such as Optical Character Recognition or Natural Language Processing and speech recognition techniques of AI can be used as a platform to convert medical records into digital format.[27] As Leo Anthony Celi, Principal Research Scientist, Laboratory of Computational Physiology, Massachusetts Institute of Technology once said “Health data is like crude oil. It is useless unless it is refined.” As mentioned earlier, AI can be utilized to study the patient data and design the appropriate treatment strategy based on the available details.[28] Woldaregay et al.,[29] discussed how mining of blood glucose data could be helpful in the prediction of Type 1 diabetes. Singh and Thakral examined BC data that were collected from Wisconsin Breast Cancer dataset.[30] This data were processed to classify benign and malignant types using decision tree and Bayes classifiers and achieved higher accuracy with the decision tree classifier. Another study from Gaza predicted the cancer patient's survivability based on the Gaza strip 2017 cancer patient's dataset.[31]
| Artificial Intelligence in Other Disciplines | | |
Applications of AI in anesthesiology including monitoring depth and control of anesthesia, prediction of risk, ultrasound guidance, and operating room logistics have been proposed and utilized.[32] In psychiatry, AI techniques continue to be refined and improved to re-define mental illnesses more objectively and to identify illnesses at an earlier stage when interventions are likely to be more effective.[33] In ophthalmology, several studies have been done to diagnose diabetic retinopathy (DR) using image processing algorithms.[34] In 2019, a set of 21 normal fundus images and 379 DR fundus images (162 nonproliferative DR and 217 proliferative DR) images were used to identify nonproliferative and proliferative DR.[35] Its utility in medical and surgical training in skill labs and simulation centers has been explored for decades. Mannequins utilize the AI principles and can simulate the disease symptoms and signs, and also can respond to dummy treatments [Figure 3]a. They are also utilized as virtual endoscopic and ultrasound training in medical practice [Figure 3]b. | Figure 3: Simulation mannequins; (a) human-patient simulator (b) ultrasound simulator
Click here to view |
| Strengths and Challenges | | |
AI has a potential to be a major partner in healthcare industry. AI is currently the best option to study a large dataset and summarize within a shorter time.[28] Speech recognition methods can assist to convert the data into digital format and to store it on the server for future references. It can smartly acquire the data for a physician by which time can be saved. It can also help in scientific publications by accurately acquiring the medical data and smart literature search. AI robots can handle the tiniest and most accurate movements, which is unique and great assistance in the robotic surgery. The AI-powered VR models can help in the therapy of mental illnesses. The AI-assisted Internet of Things technique can be used to develop an advanced remote healthcare system.[36] Health applications are already being licensed and used in various aspects of health education such as diet guidance, medication adherence, detection of arrhythmias in wearable devices with or without sensors. AI has a capability to reduce human errors if properly designed and trained. It can save time in decision-making in certain clinical scenario. Image processing can be utilized as a screening tool for the disease diagnosis in remote areas where there is a dearth of qualified professionals. AI can work round the clock which can be an advantage where there is a lack of manpower.
Incorporating AI in health systems is a challenge and requires a thorough interdisciplinary discussion among the experts in the field of healthcare and engineering. This involves high initial cost and time. Most AI techniques require a large dataset to train and develop a tool. Improper data model will yield poor results. The AI systems are completely logical and programed, hence it cannot have that emotional connection with the patient; it cannot violate some rules to save a life.[37] There is a threat of misutilization of data and the technology by patients. Fear of the self-utilization by patients can have untoward consequences of erroneous diagnosis and self-treatment. There is fear among people in healthcare industry of losing jobs that instigate them not to adopt AI or reluctance in regularly utilizing the tools.[38] Government agencies also look at them as a technology that is likely to reduce job creation. Training the healthcare professionals in AI is also a big a challenge in country as big as ours with lakhs of professionals. These are the immediate challenges and disadvantages that need to be addressed while implementing AI in healthcare.
| Regulatory and Ethical Issues | | |
Technological, ethical, and regulatory concerns are expressed by healthcare professionals while using AI applications in healthcare.[39] Like any major advances in the field of medicine and engineering, AI is not free from legal and ethical issues.[40] Since the technologies utilize large number of images ethical concerns are mandatory and are the priority. Ethical concerns are centered around the informed consent, safety, and privacy of data. It is physician's duty to sensitize and educate on the AI principles, advantages, and limitations to the patient. Devices and apps can give unsafe and incorrect recommendations which shall be highlighted.[41] Efficacy of all systems depends on the quantity and quality of the data fed into it. Regulation of intellectual property, reliability of the techniques, cybersecurity, and data protection are the major legal issues.[40] Effective law and its implementation are the responsibility of the government. If these issues are not duly addressed by the concerned, there can be backfire of the technologies.
| Future of Artificial Intelligence in Healthcare | | |
Major task for the future lies in validating the AI tools to improve skill and utility. There is a need to develop a mutually beneficial collaboration between AI and health care professionals. AI will aid the medical professionals a greater efficiency or cost-effective AI tools where clinicians can provide the data and essential clinical exposure to develop smarter models.[42] Several tools in dermatology, radiology, and laboratory medicine have proven to be more accurate than the experts in the respective fields. The Ministry of health and family welfare of India has provided the national digital health blueprint, where there is recommendation on deployment of AI and Machine learning technologies.[43] However, technology adoption in public health sector cannot be same as with banking, tourism, and transport sectors which generally involves profit.[44] It is interesting to watch to what extent and when replacement of personal expertise of human will takes place with AI-enables systems.
| Conclusion | | |
The major advantage of AI is it is faster, unbiased, fairly accurate, and reliable compared to the traditional methods in many situations. The open-source platforms and packages such as Python, OpenCV, Scikit-image, Scikit learn, and TensorFlow., are available to develop the AI tools, hence the investment to develop a system is considerably low. AI is only as good as the human train or programs the system. It will make medical care more efficient and effective if it is deployed scientifically with attention to legal and ethical issues. Since AI has already been a part of medical science and is going to be utilized more frequently. There are several concerns, but further research and advances are likely to overcome them.
Acknowledgment
The authors acknowledge Advanced Comprehensive Clinical Training and Simulation Centre – Yenepoya (ACTS-YEN) for permission to use the mannequin images.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
| References | | |
| 1. | Le Nguyen T, Do TT. Artificial intelligence in healthcare: A new technology benefit for both patients and doctors. In 2019 International Conference on Management of Engineering and Technology (PICMET); Portland. 2019 Proceedings of PICMET'19. 2019, pp. 1-15.  |
| 2. | Rong G, Mendez A, Assi EB, Zhao B, Sawan M. Artificial intelligence in healthcare: Review and prediction case studies. Engineering 2020;6:291-301. |
| 3. | Meskó B, Görög M. A short guide for medical professionals in the era of artificial intelligence. NPJ Digit Med 2020;3:1-8. |
| 4. | Dey A. Machine learning algorithms: A review. IJCSIT 2016;7:1174-9. |
| 5. | Amisha, Malik P, Pathania M, Rathaur VK. Overview of artificial intelligence in medicine. J Family Med Prim Care 2019;8:2328-31. |
| 6. | Bohr A, Memarzadeh K. The rise of artificial intelligence in healthcare applications. Artificial Intelligence in Healthcare. 2020:25–60. |
| 7. | Khare A, Jeon M, Sethi IK, Xu B. Machine learning theory and applications for healthcare. J Healthc Eng 2017;2017:1-2. |
| 8. | Lim HW, Collins SA, Resneck JS Jr., Bolognia JL, Hodge JA, Rohrer TA, et al. The burden of skin disease in the United States. JAAD 2017;76:958-72. |
| 9. | Yadav N, Narang VK, Shrivastava U. Skin disease detection models using image processing: A survey. IJCA 2016;137:34-9. |
| 10. | Yasir R, Rahman A, Ahmed N. Dermatological disease detection using image processing and artificial neural network. 8th International Conference on Electrical and Computer Engineering; Dhaka. 2014. IEEE. pp. 687-690. |
| 11. | ALEnezi NS. A method of skin disease detection using image processing and machine learning. Procedia Comput Sci 2109; 163:85-92. |
| 12. | Hegde PR, Shenoy MM, Shekar BH. Comparison of machine learning algorithms for skin disease classification using color and texture features. In: 7 th International Conference on Advances in Computing, Communications and Informatics. Bengaluru: 2018. p. 1825-8. |
| 13. | Bajaj L, Kumar H, Hasija Y. Automated system for prediction of skin disease using image processing and machine learning. IJCA 2018;180:9-12. |
| 14. | Wei LS, Gan Q, Ji T. Skin disease recognition method based on image color and texture features. Comput Math Method M 2018;2018:1-10. |
| 15. | Li Y, Shen L. Skin lesion analysis towards melanoma detection using deep learning network. Sensors (Basel) 2018;18:556. |
| 16. | Munia TK, Haque IR, Aymond A, MacKinnon N, Frkas D, Al-Hshim M. et al. Automatic clustering-based segmentation and plaque localization in psoriasis digital images. In: 2017 IEEE Healthcare Innovations and Point of Care Technologies (HI-POCT). IEEE: Bethedsa; 2017. p. 113-6. |
| 17. | Del Rosario F, Farahi JM, Drendel J, Buntinx-Krieg T, Caravaglio J, Domozych R, et al. Performance of a computer-aided digital dermoscopic image analyzer for melanoma detection in 1,076 pigmented skin lesion biopsies. J Am Acad Dermatol 2018;78:927-34.e6. |
| 18. | Hegde PR, Shenoy MM, Shekar BH. Noise removal in dermoscopic images using a novel software. Indian Dermatol Online J 2017;8:513-4. [ PUBMED] [Full text] |
| 19. | Bouabbache S, Pouradier F, Panhard S, Chaffiotte C, Loussouarn G. Exploring some characteristics (density, anagen ratio, growth rate) of human body hairs. Variations with skin sites, gender and ethnics. Int J Cosmet Sci 2019;41:46-54. |
| 20. | Oren O, Gersh BJ, Bhatt DL. Artificial intelligence in medical imaging: Switching from radiographic pathological data to clinically meaningful endpoints. Lancet Digit Health 2020;2:e486-8. |
| 21. | Pesapane F, Codari M, Sardanelli F. Artificial intelligence in medical imaging: Thrreat or opportunity? Radiologists again at the forefront of innovation in medicine. Eur Radiol Exp 2018;2:35. |
| 22. | Kourou K, Exarchos TP, Exarchos KP, Karamouzis MV, Fotiadis DI. Machine learning applications in cancer prognosis and prediction. Comput Struct Biotechnol J 2015;13:8-17. |
| 23. | Siar M, Teshnehlab M. Brain tumor detection using deep neural network and machine learning algorithm. In: 9 th International Conference on Computer and Knowledge Engineering (ICCKE 2019). Ferdowsi University of Mashhad. DOI: 10.1109/ICCKE48569.2019.8964846. |
| 24. | Lamy JB, Sekar B, Guezennec G, Bouaud J, Séroussi B. Explainable artificial intelligence for breast cancer: A visual case-based reasoning approach. Artif Intell Med 2019;94:42-53. |
| 25. | Shen L, Margolies LR, Rothstein JH, Fluder E, McBride R, Sieh W. Deep learning to improve breast cancer detection on screening mammography. Sci Rep 2019;9:12495. |
| 26. | Lin W, Chen JS, Chiang MF, Hribar MR. Application of artificial intelligence to electronic record data in ophthalmology. Transl Vis Sci Technol 2020;9:13. |
| 27. | Juhn Y, Liu H. Artificial intelligence approaches using natural language processing to advance EDR-based clinical research. J Allergy Clin Immunol 2020;145:463-9. |
| 28. | Willyard C. Can AI fix medical records? Nature 2019;576:S59-62. |
| 29. | Woldaregay AZ, Årsand E, Walderhaug S, Albers D, Mamykina L, Botsis T, Hartvigsen G. Data-driven modeling and prediction of blood glucose dynamics: Machine learning applications in type 1 diabetes. Artif Intell Med 2019;98:109-34. |
| 30. | Singh SN, Thakral S. Using data mining tools for breast cancer prediction and analysis. In: 2018 4th International Conference on Computing Communication and Automation. Greater Noida: IEEE; 2018. p. 1-4. |
| 31. | Alhaj MA, Maghari AY. Cancer survivability prediction using random forest and rule induction algorithms. In: 2017 8th International Conference on Information Technology. Jordan: IEEE; 2017. p. 388-91 . |
| 32. | Hashimoto DA, Witkowski E, Gao L, Meireles O, Rosman G. Artificial intelligence in anesthesiology: Current techniques, clinical applications, and limitations. Anesthesiology 2020;132:379-94. |
| 33. | Graham S, Depp C, Lee EE, Nebeker C, Tu X, Kim HC, et al. Artificial Intelligence for Mental Health and Mental Illnesses: An Overview. Curr Psychiatry Rep 2019;21:116. |
| 34. | Amin J, Sharif M, Yasmin M. A review on recent developments for detection of diabetic retinopathy. Scientifica (Cairo) 2016;2016:1-20. |
| 35. | Ratanapakorn T, Daengphoonphol A, Eua-Anant N, Yospaiboon Y. Digital image processing software for diagnosing diabetic retinopathy from fundus photograph. Clin Ophthalmol 2019;13:641-8. |
| 36. | Abdellatif MM, Mohamed W. Telemedicine: An IoT based remote healthcare system. iJOE 2020;16:72-81. |
| 37. | Waring J, Lindvall C, Umeton R. Automated machine learning: Review of the state-of-the-art and opportunities for healthcare. Artif Intell Med 2020;104:101822. |
| 38. | |
| 39. | Esmaeilzadeh P. Use of AI-based tools for healthcare purposes: A survey study from consumers' perspectives. BMC Med Inform Decis Mak 2020;20:170. |
| 40. | Gerke S, Minssen T, Cohen G. Ethical and legal challenges of artificial intelligence-driven healthcare. Artif Intell Healthc 2020:295-336. |
| 41. | |
| 42. | Buch VH, Ahmed I, Maruthappu M. Artificial intelligence in medicine: Current trends and future possibilities. Br J Gen Pract 2018;68:143-4. |
| 43. | |
| 44. | Bhattacharya S, Pradhan KB, Bashar MA, Tripathi S, Semwal J, Marzo RR, et al. Artificial intelligence enabled healthcare: A hype, hope or harm. J Family Med Prim Care 2019;8:3461-4. [ PUBMED] [Full text] |
[Figure 1], [Figure 2], [Figure 3]
[Table 1]
|
Search Pubmed for