Digital Health in Action Add to reading list

Jeopardy! Champion Squares Off Against Cancer

IBM’s Watson supercomputer first became famous in 2011 when it appeared as a contestant on the TV quiz show Jeopardy! IBM developed the powerful machine specifically to play on the game show as a way to demonstrate the capabilities of a computer system that uses advanced natural language processing and information retrieval to answer questions posed to it verbally. Watson defeated the two most successful players in Jeopardy! history—Ken Jennings, who has won the most games, and Brad Rutter, who has won the most money—by listening to host Alex Trebek’s questions, pressing the buzzer, giving answers and making a wager for Final Jeopardy, just like a human player.

Watson’s victory on Jeopardy! served as proof of concept for the system’s usefulness in more important areas, such as health care. Two years later, IBM struck a deal with health insurer WellPoint (now known as Anthem) to allow health care providers to tap Watson’s vast storehouse of knowledge for advice and support when making treatment decisions.

The system’s AI can analyze immense amounts of data from genomes, images and tissue samples to find signs of disease, sometimes detecting patterns that escape even the brightest, most capable diagnosticians. For example, when oncologists at the University of North Carolina used Watson for Genomics to perform a retrospective analysis of 1,018 cancer cases, the system identified therapeutic options that the clinicians hadn’t considered in 323 cases, according to a November 2017 research article in The Oncologist.

This powerful tool is still in its infancy, but it could herald a future where humans and machines work side by side to decide which treatment is best for the right patient at the right time—not just for cancer, but practically any ailment.

For example, the neurodegenerative disease ALS has no effective treatments, but it’s known to be linked to alterations in certain RNA-binding proteins (RBPs). In 2017, a team of neurological researchers used Watson to discover five more RBPs linked to ALS that had not previously been identified, bringing medical science one step closer to a treatment for ALS.

Using AI to Unlock the Power of Genomics

The project made its genome sequence freely available worldwide.

In 2003 the scientific community celebrated a historic achievement: The Human Genome Project announced that it had successfully completed its 13-year effort to map a human genome—all the genetic material in a human being. The project made its genome sequence freely available worldwide so scientists could begin using it to inform biological and medical research.

Many researchers are using genomics to try to develop “personalized medicine".

Many researchers are using genomics to try to develop “personalized medicine,” an approach that would allow clinicians to design and deliver treatments uniquely targeted to each patient. For instance, genomics might make it possible for a scientist to trace the origin of a patient’s breast cancer to a specific genetic mutation that caused cellular dysfunction in her breast tissue. Then a physician could use a targeted gene therapy to attack the cancer at its true point of origin—the genetic level—where it could be permanently cured.

One of the most promising technologies in this area, called CRISPR (pronounced “crisper”), is a suite of nano-sized tools that scientists are already using to cut and rearrange DNA. They hope to one day use CRISPR to rewrite a person’s genetic code to erase the risk of specific diseases. However, many genomic regions appear very similar, so sometimes CRISPR works on the wrong gene. This can lead to unintended consequences called “off-target effects.”

That’s where algorithms can make an impact. Elevation, a cloud-based search engine released by Microsoft, enables scientists to input the name of a gene they plan to modify using CRISPR. The search engine uses a type of AI called machine learning to provide a list of guides that researchers can use to predict—and prevent—off-target effects during the gene-editing process.

Solving Medical Mysteries Through Genetic Sequencing

Researchers at UNC launched the NCGENES project to study whether a new genomic test called whole exome sequencing could be a useful diagnostic tool. The exome is a tiny subset of the human genome, only constituting about 1% of all human genetic material; but mutations in the exome are believed to account for 85% of all disease-related genetic mutations, making it fertile ground for diagnostic research.

The NCGENES scientists conduct their research by performing exome sequencing on patients who have certain genetic disorders or who have health problems that have gone otherwise undiagnosed. The project has sequenced the genomes of 750 people who meet these criteria.

One NCGENES patient had developed a mysterious condition at the age of 6 that caused her to begin walking on her toes. As the condition persisted and grew worse, other kids made fun of her, and by the time she was in sixth grade she had to use crutches to walk. She underwent a total of 10 foot surgeries—five on each foot—saw doctor after doctor and took medications, but nothing helped. The condition lasted into adulthood, when she was referred to NCGENES for help.

Six weeks later she was walking again—without the aid of crutches, which she hasn’t needed since.

The researchers sequenced her exome, and something caught their attention: a variant in a gene associated with a condition called dopa-responsive dystonia. Doctors gave her L-dopa, a drug commonly used to treat Parkinson’s disease, and within three days her toes began relaxing. Six weeks later she was walking again—without the aid of crutches, which she hasn’t needed since.

Full-Spectrum Medical Innovation Through Digital Health

An onrushing avalanche of developments in information technology is revamping health care from the ground up. Starting with basic medical research, progressing through the art and science of diagnosis and extending into front-line treatment, digital health is changing—and improving—how we use medicine to support health. And we’re just getting started. Digital health researchers and clinicians will continue providing us with new tests, therapies and even cures to help people live longer, healthier lives.