"Live a good life, and in the end, it’s not the years in the life, it’s the life in the years."

How AI will make better medicine

By on July 23, 2018 in Columnist with 0 Comments

Jim BrownBy Jim Brown, M.D.

For generations in our classic medical teaching, training and clinical practice, we physicians were taught which medical treatments had been successful for populations of “average” patients with certain diseases, symptoms and physical findings including their lab tests and radiologic studies.

We were taught the average patient with a certain diagnosis was best treated in a way that had been successful in other patients like them.

This “one size fits all” approach worked for many patients but not for others.

Precision medicine, also referred to as “personalized medicine,” is changing medicine in a dramatic fashion now and in the future.

Precision medicine is a relatively new approach to disease treatment taking into account the individual variations in the genes, environment and life style.

This approach will allow physicians and researchers to predict more accurately which treatment or medication for a particular disease or condition will be the best or safest ones for a particular patient.

Precision medicine is only possible due to the development of AI or artificial intelligence.

AI is a branch of computer science dealing with the simulation of intelligence by computers. AI does not “think” like human intelligence. Artificial intelligence is capable of solving some of the most complex computational problems we have ever grappled with.

What AI can do is process millions of data points automatically and constantly in a way human intelligence cannot do. Without AI there would be no precision medicine.

Genes are the basic unit of heredity. Humans have between 20,000-30,000 genes.

Each person has two copies of each gene, one inherited from each parent. Genes are made up of DNA, which acts as an instructor to make proteins. DNA is the hereditary material in humans and almost all organisms.

The information in human DNA consists of four nitrogenous bases: adenine, guanine, cytosine and thymine arranged in more than 3 billion sequences.

About 99 percent of these sequences are the same in all people, but one percent of the total are slightly different in each of us. These relatively small differences contribute to each person’s unique physical differences.

(It seems ironic to me these seemingly small differences have lead to many of our society’s prejudices and our difficulty accepting each other even though we are 99 percent genetically identical in every way.)

An important property of DNA is its ability to replicate, making copies of itself. When our cells divide, each new cell is an exact copy of the DNA in the old cells. A defect in a gene can continually be replicated and can lead to an abnormality, a physical defect or a medical condition.

Without AI it would be difficult or nearly impossible to discover many of these genetic defects. It would be like finding a needle in a proverbial haystack.

A practical real life example of AI is the company Atomwise that launched a virtual search for safe existing medicines that could be redesigned to treat the Ebola virus.

They found two drugs predicted by AI technology that might reduce Ebola infectivity. This analysis with AI was completed in one day versus what typically might have taken months to years.

Another example has to do with diabetic retinopathy, a leading cause of vision loss affecting 100 million people globally. Diabetic retinopathy affects diabetic patients older than age 40, and if found through early detection, is treatable.

If diabetics worldwide were to be screened, it would take 32 million retinal images annually, a daunting or impossible task to be reviewed, but not so for AI.

Another recent example was published by researchers at Northwestern University Medical school regarding rheumatoid arthritis, a crippling arthritic condition.

The study’s author Dr. Perlman said, “We have so many different biologic drugs and there is no rhyme or reason to give one drug versus the other. We waste $2.5 billion a year on ineffective therapy and patients go through 12 weeks of therapy and often don’t respond and get upset.”

These scientists took biopsies of joints in 41 affected patients. They identified two particular groups who shared certain aspects of the genetic profiles of particular cells of the immune system that are overactive in rheumatoid arthritis. They then identified which of these populations had joints getting better and what biologic therapies they were taking. They found a particular gene sequence that they could treat earlier and more effectively. This is hope of precision medicine.

“Deep learning” is a branch of AI with the potential to identify when patients are likely to develop a particular disease or condition so they might be treated earlier and more aggressively. For these patients this is precision medicine.

Google, Facebook and Apple have been using deep learning data analysis for years to predict how we individuals search the internet, where we travel, what foods we like, and even who our potential friends are.

Hopefully, soon we will be using deep learning to identify which patients are likely to develop a particular disease. The hope is early intervention might change the future course of that disease.

AI’s ability to access tons of data and learn from it definitely applies to the health care industry. Half of all hospitals in the U.S. reportedly plan to adapt AI within the next five years.

Then there is China. China is becoming a major technology super power.

China’s goal is to rival the U.S. and become the world leader in AI by 2030.

In health care they are already number one in AI. China currently has 130 companies specifically working in the health care field and health care research. For China, improving access to quality heath care is a priority and not an option.

I am afraid our government’s priorities are currently not even close to China in this area.

On Jan. 30, 2015, President Obama announced his Precision Medicine Initiative. When the Trump administration came into power, they proposed cutting funding by 22 percent or $7.6 billion to the National Institutes of Health’s budget for the coming fiscal year. This cast doubts about the future of precision medicine research.

Fortunately, six months later the House and Senate appropriation committees both approved bipartisan bills that not only restored the proposed cuts, but gave NIH a $2 billion raise in their budget.

Thank goodness, as I don’t think we can afford to fall behind in this important field.

Jim Brown, M.D., is a retired gastroenterologist who has practiced for 38 years in the Wenatchee area. He is a former CEO of the Wenatchee Valley Medical Center.

About the Author

About the Author: .


If you enjoyed this article, subscribe now to receive more just like it.

Post a Comment

Your email address will not be published. Required fields are marked *