The direction of precision medicineTechnology & Trends Article 5 Minute Read GE Healthcare Global
What is precision medicine?
In 2015, the U.S. government launched the “Precision Medicine Initiative.”1 Precision medicine is often interchangeable with the terms personalized medicine and precision health. Each of these ideologies are similar overall but with slightly different approaches. Precision medicine was actually referred to as personalized medicine until it became apparent that people were getting confused about whether treatments and methods were being developed on an individual level. Precision health comes from the idea that physicians should be focusing on preventing disease rather than treating it after the fact. Precision medicine aims to individualize healthcare. In other words, precision medicine is the idea of providing the right medicine to the right person at the right time. One of the most common examples of this has been happening even before this initiative was launched: blood transfusion. When preparing to give a patient blood transfusion, the physician checks a patients blood type and match it to a donor. Since the presidential initiative, precision medicine has begun to expand to include more of the healthcare spectrum.
The goal of precision medicine is to treat diseases and disorders based on each person’s genetic fingerprint, environment, and lifestyle.2 As an example, cancer is the most advanced area of care in terms of precision medicine. Oncologists use a variety of precision medicine techniques to provide better cancer care. To do that, the hospitals and clinics where patients receive treatment at needs to understand how each person will react to different medicines and why they will react that way. Clinicians have discovered that each cancer is unique and caused by mutations in a person’s cells. Once they identify the mutation, they could provide treatment specific to it. Similarly, other fields have the same sort of issue, especially in terms of genetic and hereditary disorders. Medical fields, like oncology, are striving towards discovery of the causes of diseases in order to treat and, if possible, prevent them. Magnetic resonance (MR) technology now uses imaging genomics and AI, among many other tools, to help other disciplines align to precision medicine.
MR imaging and genomics
One of the biggest emerging fields related to precision medicine is genomics. Genomics is the area of genetics focusing on sequencing and monitoring one’s entire DNA content, or genome.3 Genomics aims to understand how a particular gene is affected in relation to the entire genome. On the other hand, genetics aims to understand how one gene works. Genomics has been around for a number of years, but scientists are only now gaining access to previous studies and sequenced genomes. This is due to the amount of storage required for genomic data. Cloud storage has quickly expanded the possibilities for scientists around the world to store their own studies and potentially access others’.
Imaging genomics studies different imaging techniques and how the images relate to a person’s genome.4,5,6 Studies can be conducted based on diseases or disorders that have a genetic basis, based on twins or other related people, or based on cancer types. This enables precision medicine based on a person’s genomes and, in the future, based on the results of an MRI. Imaging genomics allows for an evaluation of the brain and how different genes affect diseases, as well as actions and intelligence.
Two collaborative projects, The Cancer Genome Atlas (TCGA) and The Cancer Imaging Archive (TCIA,) allow scientists across oncology to study the different tissues and liquid present within a tumor (TCGA) in relation to how the tumor appears on an MRI (TCIA.)4,5 One study, conducted by Pascal O. Zinn, et at., showed that fluid-attenuated inversion recovery (FLAIR) MRI is able to detect cancer genomic components related to cellular invasion in Glioblastoma Multriforme, allowing for screening for different types of cancer and more precise treatments for each patient.
Functional MRI (fMRI) shows the way brains work in relation to different sort of tasks among a wide pool of people.6 This allows doctors to compare the connection and ability of the brain to process different things in relation to cognitive disorders. Likewise, diffusion tensor (DT) MRI allows radiologists to create virtual maps of the brain and integrity of the tissue. Studies completed on twins, comparing identical to fraternal, using DTI have proven that there is a significant correlation between genetics and the structure and function of the brain. This is possible due to the fact that identical twins have identical DNA.This information could lead to a better understanding of certain diseases and more precise screening for neurological and psychological illnesses among certain patients.
Strides in artificial intelligence and MRI
Artificial Intelligence (AI) is becoming more and more important to citizens of the world. One of the best examples of this is the technology used to open phones based on a face or iris identification. People don’t look at their phone from the exact same angle in the exact same light every time, but the phone still identifies them individually. This is an example of machine learning. Every time someone goes to open their phone, the processor adds an image to its recognition software. MR technology is beginning to gain similar capabilities through machine learning.
As this technology is developed, physicians look for ways to create a deeper understanding from the machine of how abnormalities appear.7,8 Researchers at Mayo Clinic are using MR images created by AI to train deep learning models of AI to recognize abnormalities in MRIs. This process uses generative adversarial networks (GANs) to create new images from samples of existing images. The new images are similar to the old images but have no potential privacy violations since they are not images of real people. For example, for brain imaging, it can create a substantial number of images that have differently sized tumors in different locations, as well as create an image of a healthy brain. The computer will learn to identify abnormalities based upon the new images in each patient, allowing for potentially faster and more useful exams.
Precision medicine has made significant advances, through imaging genomics and AI, since the Precision Medicine Initiative was announced in 2015. These fields had already begun making progress but are now doing so at exponential rates. Imaging genomics will allow for physicians to screen for genetic diseases based on their imaging and the patient’s genome. This will allow them to treat the patient before the disease progresses too far and learn the best treatments. AI is learning to identify abnormalities in the anatomy through deep learning, thus making it possible for MR results to come back sooner and be more precise. These techniques will infinitely help the medical industry, both through causative and preventative care.
1. Francis S. Collins & Harold Varmus. “A New Initiative on Precision Medicine.” N Engl J Med. 26 February 2015; 372(9): 793-795. Web. 31 October 2018. <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5101938/>.
2. Subha Madhavan, et al. “Art and Challenges of Precision Medicine: Interpreting and Integrating Genomic Data Into Clinical Practice.” American Society of Clinical Oncology Educational Book. 23 May 2018; 5: 546-553. Web. 6 November 2018. <http://ascopubs.org/doi/full/10.1200/EDBK_200759>.
3. Ananya Mandal. “What is Genomics?” News-Medical.net. 23 August 2018. Web. 31 October 2018. <https://www.news-medical.net/life-sciences/What-is-Genomics.aspx>.
4. C. Carl Jaffe. “Imaging and Genomics: Is There a Synergy?” Radiology. 1 August 2012; 264(2): 329-331. Web. 31 October 2018. <https://pubs.rsna.org/doi/full/10.1148/radiol.12120871>.
5. Pascal O. Zinn, et al. “Radiogenomic Mapping of Edema/Cellular Invasion MRI-Phenotypes in Glioblastoma Multiforme.” PLOS ONE. 5 October 2011; 7(2): 10.1371. Web. 31 October 2018. <https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0025451>.
6. Paul M. Thompson, Nicholas G. Martin, & Margaret J. Wright. “Imaging genomics.” Curr Opin Neurol. 23 August 2010; 23(4): 368-373. Web. 31 October 2018. <https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2927195/>.
7. Jennifer Bresnick. “MRI Images Created by AI Could Help Train Deep Learning Models.” healthitanalytics.com. 24 September 2018. Web. 31 October 2018. <https://healthitanalytics.com/news/mri-images-created-by-ai-could-help-train-deep-learning-models>.
8. Melissa Rohman. “AI, imaging project to improve precision medicine for treating concussions.” HealthImaging.com. 3 July 2018. Web. 31 October 2018. <https://www.healthimaging.com/topics/artificial-intelligence/ai-imaging-precision-medicine-concussions>.