When Artificial Intelligence meets Personalized Medicine

By: Ikram Issarti, Project Manager

Artificial intelligence (AI) has witnessed evolutionary growth and became an integrated part of various sectors, including medicine. In parallel, the rapid and exponential emergence of new medical technologies has accelerated and raised the availability of digital health data, creating niche opportunities for the application of AI in healthcare. In the last couple of years, AI has transformed medicine with great vigor; with a demonstrated effectiveness in medical practices, in decision making, early diagnosis, predicting surgical outcomes, assessing future risk progression of diseases, and improving health care management systems. Today, we highlight how a successful collaboration between medicine and AI can lead to an informed treatment decision for a specific patient.

AI refers to a demonstrated intelligence by machines, appearing in the ability to mimic certain human characteristics, such as the capabilities to learn, adapt, sense, intercept, comprehend and generalize insights, etc. There is a non-uniform agreed definition for AI, however, John McCarthy, one of the founders of AI research, defines the field as ‘’AI is getting the computer to do things which, when done by people, are said to involve intelligence“.1 Nowadays, Artificial Intelligence has invaded every life walk, e.g. computer vision, handwriting recognition and audio recognition. These are frequently encountered examples where a machine (smartphone, computer, etc.) is given the capabilities to sense/learn/comprehend, facial, vocal, and written features. Another example is the Natural Language Processing (NLP), which refers to giving the computer the capability to understand text and spoken words. There are various subdomains of Artificial Intelligence; from which, Machine Learning (ML) and Deep Learning (DL) are the most widely used, and thus the terms AI/ML/DL are often used interchangeably. ML refers to a family of intelligent algorithms that are capable of identifying and extracting patterns from data. While DL is a subdomain of ML, capable of analyzing layers of data, frequently applied for image processing. The process of learning, however, is somehow equivalent to the human learning concept. In machine learning, the algorithm learns from previously collected data (experience/life scenarios) and generates further insights, which could be interpreted as classification, prediction, or forecasting tasks, etc. To summarize:


  • AI is a machine that exhibits certain features of intelligence.
  • The intelligent features are created by humans.
  • Once the machine is given intelligent features, it can perform tasks without being instructed to do so.
  • The machine gains the capability of generating new insights from the learning.


Medicine is the science of practice and caring for a patient, and we all agree that it is important to maintain our health and our life, but it is always a surprise seeing a drug work for some people and be less effective for others. Or why do some people develop certain diseases, while others do not? Or why would you feel better after administrating an Aspirin, while your partner may need a double dosage to feel the same? If the response to one pill of Aspirin can be different for partners with the same lifestyle, how could it be that patients are treated similarly for more life-threatening conditions, such as cancer? Concerns have been raised about the applicability of generalized guidelines to an individual patient, and whether individualized guidelines could result in better care and lower costs.2,3

To approach each patient individually with a medication tailored to the patient biology and history is known as personalized medicine. According to the definition of the Horizon 2020 advisory group, personalized medicine is a medical model using characterization of individual phenotypes and genotypes such as molecular profiling, medical imaging, lifestyle data, to identify the right therapeutic strategy for the right person at the right time. The primary goal of personalized medicine is to predict the best patient-specific treatment strategies and whether or not a patient will develop a certain disease. In standard practice today, doctors use evidence-based medicine, which is derived from the results of randomized controlled trials (RCTs) or expert consensus, to make informed treatment decisions. However, this approach does not cover all clinical conditions and it is limited in the number of subjects recruited. The question was raised whether combining information from clinical treatment guidelines and historical data could lead to an improved decision making and more informed treatment strategies.4

Nowadays, technologies and healthcare facilities enable us to generate massive amounts of data at various scales, such as individual clinical parameters, therapeutic protocols, patient outcomes, electronic health records, imaging scans, DNA sequences, genetic information, and wireless health data.5 This enables a better monitoring for the patients but also results in massive overwhelming amounts of digital data that are difficult to interpret for practitioners. Therefore, clinical specialists tend to focus on a small number of diagnostic/treatment criteria in addition to their subjective expertise, increasing the risk that certain indicators for a focused treatment could escape their attention. A research study that included 20 years of data for nearly 2.5 million patients from a large database of the Atrius Health care provider in eastern Massachusetts, found surprising results: “More often there are multiple treatment options, some of which are efficient and yet may go unconsidered in the treatment guideline planning”. Therefore, the same research group implemented an AI software that was able to determine the best treatment option, and in addition, suggests informed decision for each patient.

Despite personalized medicine being effective, it is still difficult to implement in practice. With the ability of AI to analyze complex data and extract meaningful insights in an objective and automated way, AI could add a valuable outcome. If a close collaboration between healthcare specialist and AI technologies integration could be maintained in clinical practice, this would enable to enhance medical service delivery to make tailored data driven treatment decision, additionally to reducing the burden on the healthcare system and optimizing treatment costs.


At Modis we have a multidisciplinary team with full stack of expertise to help your organization implement technology centered projects, as such applied Artificial intelligence tailored to your specific challenges. For more details, please do not hesitate to get in touch!

– Ikram Issarti


Talk to an expert!


Interested in the topics of AI and personalized medicine? Modis is attending Life on Chip next week, an event entirely dedicated to cross-over innovations and health technologies. See you there?



  1. Shubhendu, S. & Vijay, J. Applicability of Artificial Intelligence in Different Fields of Life. IJSER, (2013).
  2. Kent, D. M. & Hayward, R. A. Limitations of applying summary results of clinical trials to individual patients: the need for risk stratification. JAMA 298, 1209–1212 (2007).
  3. Tang, P. C. et al. Precision population analytics: population management at the point-of-care. Journal of the American Medical Informatics Association 28, 588–595 (2021).
  4. Ng, K., Kartoun, U., Stavropoulos, H., Zambrano, J. A. & Tang, P. C. Personalized treatment options for chronic diseases using precision cohort analytics. Sci Rep 11, 1139 (2021).
  5. Schork, N. J. ARTIFICIAL INTELLIGENCE AND PERSONALIZED MEDICINE. Cancer Treat Res 178, 265–283 (2019).