Precision medicine is a medical model, which proposes customization of the healthcare to a subgroup of patients, based on a genetics, lifestyle and environment. This technique allows doctors and researchers to prognosis treatment and prevention strategies for a specific disease which can work on a group of people. It is opposed to a one-size-fits-all approach, in which disease treatment and prevention techniques are advanced for the average individual with much less attention for the variations among individuals.
There is an overlap between the terms “precision medication” and “personalized medicine.” As per the National Research Council, “personalized medicine” is a traditional word with a meaning close to “precision medication.” However the word “personalized” may be misinterpreted to suggest that treatments and preventions are being evolved uniquely for every person. In precision medicine, the focus relies on figuring out the methods which can be effective for group of patients.
Precision medicine approach leverages a patient’s genetic history, location, environmental factors, lifestyle and habits to determine a plan of action for treatment. Artificial Intelligence has been successfully able to classify problems using different algorithms and solve precision medicine problems e.g. accurate disease diagnosis, disease detection and prediction, treatment optimization. The analysis of multidimensional datasets to capture variations can be learnt (trained) by using AI algorithms and identify cryptic phenotypic orgenotypic structures. It can further be used to predict the risk of a disease, identification of the disease response and outcomes on the individual patients based on their own characteristics. Recently, prediction algorithms utilizing artificial intelligence approaches for cancer and cardiovascular disease have shown promising results, predicting disease risk with a higher degree of precision (Uddin et al., 2019). The article reviews applications of Artificial Intelligence and various algorithms which can help in better healthcare for humankind.
Precision medicine is an important and powerful technique for the diagnosis of diseases and patients precare. It involves analysis of patient’s personal data, genetic information, circumstances to diagnose and cure the disease. It allows researchers to design and develop the medication for prevention of specific viruses. It has the potential to improvise the traditional symptom driven retrospective practice of medicine, by allowing earlier interventions with advanced diagnostics, which can further be used for tailoring personalized treatments. Identification of the pathway for developing a personalized medicine involve analyzing comprehensive patient information along with broader aspects to monitor and distinguish between healthy and sick people, which will lead to a better understanding of biological indicators that can signal shifts in health. In order to positively impact the patient’s health and to provide real time decision support, it is vital to leverage the power of electronic health records by integrating disparate data sources and discover patient-specific patterns of disease progression. The goal of is to use multiple types of data and classify patients into precise groups that will benefit from a given treatment approach. Artificial Intelligence techniques can be used for clinical data extraction, aggregation, management and analysis to support clinicians by efficiently stratifying subjects to understand specific scenarios and optimize decision-making. Artificial Intelligence in healthcare has the potential to achieve the goals of providing real-time, better personalized and population medicine at lower costs (Ahmed et al., 2020). In this article, we will focus on various machine learning, deep learning models, and applications of AI which can pave the way for a new data-centric era of discovery in healthcare.
Traditional Medicine versus Precision Medicine
In traditional medicine, the doctor uses their expertise and trial and based method. Based on assumptions from the symptoms given by patients, doctor suggests same medicine with equal dosage. This type of treatment may not work every time. Techniques that benefit any victims are weak for others and the corresponding injection may additionally produce floor consequences in only some cases. (Gandikota et al., 2020)
Precision medicine is the orienting of clinical strategy to the particular characteristics of the individual patient. The technique depends on correct findings in our understanding of how a patient‘s unique outline gives them sensitive to some situations. The corresponding analysis is enhancing our information to divine which treatments be reliable and valid for any case. As shown in below figure, the doctor suggests medicine and dosage based on person DNA and personal health information. (Gandikota et al., 2020)
Precision medicine approach can be considered as an extension of traditional approaches to treat the disease with greater precision. A profile of a patient’s gene variations can guide the choice of drugs or remedy protocols that reduce side effects or ensure greater success outcomes.
Role of Artificial Intelligence in Precision Medicine
Artificial intelligence (AI) has been used for years in the field of healthcare and continue to grow tremendously each year with its ability to advance medicine and research. Even precision medicine is not completely possible without the addition of machine learning algorithms to assist in the process. Machine Learning (ML) is an application of artificial intelligence (AI) that can learn and upgrade from experiences and without being explicitly coded by programmer. ML specializes in the development of computer programs which can retrieve data and use it to learn for themselves. Most commonly used ML algorithms in medicine includes SVM, deep learning, logistic regression, DA, decision tree, random forest, linear regression, Naïve Bayes, K-nearest neighbor (KNN) and hidden Markov model (HMM). (Ahmed et al., 2020)
Above figure shows that ML algorithms are applied for clinical, genomics, metabolomics, imaging, claims, labs, nutrients and life style data fusion, integration and analysis. So, ML algorithms integrates multiple data sources to produce more consistent, accurate, and useful information than that provided by any individual data source.
Over the last few years, AI approaches has been used in neurodevelopmental disorders specifically in autism spectrum disorder, epileptic encephalopathy, intellectual disability, attention deficit hyperactivity disorder (ADHD) and rare genetic disorders. (Mohammed et al., 2019). AI algorithms can create an impact in 4 complex unresolved problems in neurodevelopmental disorders as shown in below figure.
1. Identifying Causal Genes
AI methods are crucial for identifying causal genes and locus. Bioinformatics prediction still not able to exactly classify the more common missense mutations as per pathogenicity. Even identifying causal genes from those ‘variations of uncertain importance’ (VUS) stays a major unresolved hassle that does lend itself to an AI solution.
AI models have recently shown reasonable success for improving genetic diagnostics in Neurodevelopmental Disorders (NDDs). Two AI algorithms named Human Splicing Code, and DeepSEA showed very promising results in the challenging task of correct classification of missense variants.
2. Phenotypic and genetic heterogeneity
Despite the fact that NDDs are ordinarily genetic in etiology, environment will nonetheless effect on genetically driven brain patterning, and consequently have the capacity to influence disease severity. Multiple independent reports have shown an association among postzygotic mosaic mutations and autism spectrum disorders, intellectual disability, epilepsy and other NDDs. In the last few decades, digitization of medical health record added a large amount of data related to healthcare. The application of AI algorithms might be significantly benefitted from those digitization efforts that can help establish genotype phenotype relation for genetic diseases and have the capacity to conclude numerous phenotypic correlations and associations.
3. Polygenic Risk Score and Gene-Gene Interactions
Gene – Gene interaction is a major contributor to the phenotypic variance of NDDs but there is currently no credible AI algorithm able to cope with data on this scale. There exist major complexities concerning deep phenotypic and large scale omics data. Unsupervised AI approaches may be applied to perceive previously unknown sub-structures within NDD cases based on environmental factors, dosage balance etc. Even though none of the techniques have been carried out in a quantitative context. So the future work can be done in quantifying polygenetic score and gene – gene interactions using AI/ML algorithms.
4. Drug Discovery
AI models are at the frontier for therapeutic intervention and drug design. Currently, there are 51 food and drug administration (FDA) approved targeted gene unique drugs for neurology and psychiatric situations. The advent of sequencing technology has basically been targeted on facilitating the implementation of early precision diagnostics. Recently the appearance of genome modifying technologies (i.e., CRISPR/cas9), and antisense oligonucleotide remedy has allowed scientists to mimic cellular phenotype, and help become aware of precise molecular objectives. Such drug design would require a primary push on AI algorithm implementation. Recently the idea of repurposing drug is turning into a prime area of research using AI algorithms.
Applications of Artificial Intelligence in precision medicine
An example is biomarker development in precision medicine for early-stage lung cancer. Biomarkers are characteristics of the body that we can measure. e.g. blood pressure or heart rate can be considered as a biomarker. Biomarkers are integral to drug development, because we need to measure the effects of investigational drugs on people during the clinical trials. It has shown use of precision medicine (biomarkers) to classify patients with early-stage lung cancer into subclasses to provide appropriate treatment. Below figure classifies early stage (IA and IB) lung cancers by biomarkers that predicts risk of recurrence generated using a precision medicine research strategy into low risk for recurrence and high risk for recurrence. (Ashley et al., 2016)
Artificial Intelligence takes precision medicine to the next level and increases the accuracy and prediction of outcome for patients. It can also make treatments more affordable and accessible to those who may not be able to receive those treatments due to cost and health insurance at this time. There are many challenges ahead for precision medicine to be perfect, but artificial intelligence can help drive us closer to those goals.
- Zeeshan Ahmed, Khalid Mohamed, Saman Zeeshan, XinQi Dong, Artificial intelligence with multi-functional machine learning platform development for better healthcare and precision medicine, Database, Volume 2020, 2020, baaa010, https://doi.org/10.1093/database/baaa010
- Sushen Zhang, S. M. Hosseini Bamakan, Qiang Qu, Sha Li, “Learning for Personalised Medicine: A Comprehensive Review from Deep Learning Perspective”, Ieee Reviews In Biomedical Engineering, Vol. XX, No. XX, March 2018.
- Gandikota Ramu, P. Dileep Kumar Reddy, and Appawala Jayanthi, “A Survey of Precision Medicine Strategy Using Cognitive Computing”, International Journal of Machine Learning and Computing, Vol. 8, No. 6, December 2018.
- Jamilu Awwalu, Ali Garba Garba, Anahita Ghazvini, and Rose Atuah, “Artificial Intelligence in Personalized Medicine, Application of AI Algorithms in Solving Personalized Medicine Problems”, International Journal of Computer Theory and Engineering, Vol. 7, No. 6, December 2015.
- Ashley J. Vargas, Curtis C. Harris, “Biomarker development in the precision medicine era: lung cancer as a case study”, Nat Rev Cancer, 2016
- Mohammed Uddin, Yujiang Wang, Marc Woodbury-Smith, “Artificial intelligence for precision medicine in neurodevelopmental disorders”, npj Digital Medicine (2019) 2:112 ; https://doi.org/10.1038/s41746-019-0191-0
Kamal Jain is a dynamic, result oriented digital technocrat with over 19 years of experience and proven track record of building futuristic products using digital technologies – Artificial Intelligence (AI), neural networks and deep learning, Machine Learning (ML), Natural Language Processing (NLP), Virtual Desktop Infrastructure (VDI), ‘Bring Your Own Device’ (BYOD), Cloud based Software as a Service (SaaS) solution stack. He has rich experience in building complex technical product solutions in matrix organizations with international experience across different countries – US, the UK, Spain, South Korea, UAE and India. As a part of giving back to the society and tech sector, he has been voluntary mentoring and providing knowledge sessions at various universities, tech start-ups in the areas of Artificial Intelligence, Machine Learning, Deep Learning & Neural Networks, Cloud based technologies etc.
Vinita Shah is working as an Assistant Professor in the Department of Information Technology, G H Patel College of Engineering & Technology, V. V. Nagar, Anand, Gujarat. She has more than 9 years of teaching experience. She obtained Master Degree in Information Technology from Gujarat Technological University. She has published more than 18 Research papers in various National/International Conferences and Journals. Her research interests include Artificial Intelligence, Machine Learning, Computer Vision and Data Science.