In the last few years, the field of informatics has been applied in different areas.The Artificial Intelligence (AI) technology has been a common approach in devices and computer algorithms that are ubiquitously present in modern technology, such as smartphones. A rising number of data aggregation is complicating the extraction of evidence from it, thus companies are now searching for new methods (1). These AI approaches are one of the possible answers to this issue and in several other aspects they are extremely beneficial to the health sector. The growth of health applications and digitalization of paper-based registries have helped to set a promising direction in the future of AI in health. In the following, the major impacts that AI technology might have in the rare diseases community are underlined.
One possible way to leverage the collected data is to apply innovative computer methods such as AI on more specifically a branch of it called Machine Learning (ML) which consists in tackling this challenge through computational algorithms and mechanizing processes that usually are done by doctors (2). Reaching a diagnosis normally involves both the person with a specific disease and the doctor, but innovations in this area make it possible to simplify this process through algorithms that might identify potential CDG bearers or facilitate a drug approval. The use of data from multiple parameters such as symptoms, age groups, current medication, and other variables will help to create exclusion and inclusion criteria in the algorithm. Likewise, continual monitoring of the health status of a potential CDG infirm via an AI device will help to improve our understanding of disease patterns and help us to standardise diagnoses afterwards for these diseases which allows a better treatment and a customized response to their needs. Examples of such devices are those used for monitoring people's health status at the first appearance of common symptoms and to inform the stakeholders. This permits an early diagnosis and better response to the problem. The recent framework issued by the Food Drug and Administration (FDA) (3) allows the application of Real-World Data (RWD) in several situations which provide unique opportunities to create new AI machines to integrate in Research and Development (R&D) (4).
The results of a survey made to life science companies (5) show that 42% of the respondents use web-based services, such as apps to monitor the individuals with certain diseases and 21% of the companies utilizes wearables, devices, and sensors as mentioned before.
According to another survey from 2018 by Deloitte (1) targeting biopharmaceutical companies’ approach to Real-World Evidence (RWE), 60% of the companies are already applying Machine Learning algorithms to analyse RWD, while 95% anticipate an increase in the use of ML in the coming years.
This type of approach has already been used for rare diseases such as cardiac amyloidosis (6) where the variables applied in their algorithm were: age diagnosis of cardiac arrest, chest pain, congestive heart failure, hypertension, prim open angle glaucoma, and shoulder arthritis. The accuracy of this process had used 73 positives for cardiac amyloidosis and 197 negatives. Although the cross-validation score was high it was stated that further studies must be made to validate the accuracy of this algorithm.
Alternatively, GNS healthcare has a partnership with Manton Center for Orphan Disease Research at Children’s Hospital in Boston where it is shown to be a good application of AI approach (7). This collaboration is based on the collection of data in a platform called “20 Rare-Disease Questions (20RDQ)”. This method focuses on using AI to create a prioritized list of suspect variants of a rare disease.
For a better understanding of these concepts watch the 3 videos shown bellow that focus on analytics driven by data, machine learning for treating rare diseases and identifying patients for rare diseases respectively.
For a better understanding, watch below the following videos are available:
- Created by Axtria Inc and named Data Driven Pharmaceuticals & Industrialization of Analytics
- Created by Webmedy Entitled Machine Learning for treating Rare Diseases
- Named Identifying patients for rare diseases by IQVIA
- https://www2.deloitte.com/content/dam/Deloitte/us/Documents/life-sciences-health-care/us-ls-2017-real-world-evidence-survey-031617.pdf Accessed March 2021
- https://www.ibm.com/cloud/learn/machine-learning Acessed April 2021
- https://www.fda.gov/media/120060/download Acessed March 2021
- Wang, F., & Preininger, A. (2019). AI in Health: State of the Art, Challenges, and Future Directions. Yearbook of medical informatics, 28(1), 16–26. https://doi.org/10.1055/s-0039-1677908
- https://assets.kpmg/content/dam/kpmg/xx/pdf/2018/01/digitalization-in-life-sciences.pdf Acessed March 2021
- Garg R., Dong S., Shah S. et al. (2016) A Bootstrap Machine Learning Approach to Identify Rare Disease Patients from Electronic Health Records
- https://blog.gnshealthcare.com/how-ai-is-aiding-discoveries-in-rare-disease-research Accessed March 2021
- https://insights.axtria.com/blog/three-rare-disease-diagnoses-opportunities-for-ai/ml-artificial-intelligence-and-machine-learning Accessed March 2021
Didn't find what you are looking for?
Your question may help others
Alexandre Gil and Pedro Granjo from Sci and Volunteer Program Nova School of Science and Technology 2021.
The Site cannot and does not contain medical or health advice. The information is provided
for general informational and educational purposes only and is not a substitute for
Accordingly, before taking any actions based upon such information, we encourage you to
consult with the appropriate professionals. We do not provide any kind of medical or health
advice. The use or reliance of any information contained on this site is solely at your own
Like the World CDG Organization Facebook Page. Share the page on your own timeline, and tell your friends to share it.
Follow us on Twitter and LinkedIn.
Subscribe to our Youtube channel and invite your friends to subscribe too.